Economics and Information Theory

ECONOMICS
AND
INFORMATION THEORY
Studies in Mathematical and Managerial Economics

NUNC COCNOSCO EX PARTE
TRENT UNIVERSITY
LIBRARY
Digitized by the Internet Archive
in 2019 with funding from
Kahle/Austin Foundation
https://archive.org/details/economicsinformaOOOOthei
ECONOMICS AND INFORMATION THEORY
STUDIES
IN MATHEMATICAL AND
MANAGERIAL ECONOMICS
Editor
HENRI THEIL
VOLUME 7
1967
NORTH-HOLLAND PUBLISHING COMPANY - AMSTERDAM
RAND McNALLY & COMPANY - CHICAGO

ECONOMICS
AND
INFORMATION THEORY
by
HENRI THEIL
Center for Mathematical Studies in Business and Economics
The University of Chicago
1967
NORTH-HOLLAND PUBLISHING COMPANY - AMSTERDAM

V\biu-
No part of this book may be reproduced

in any form by print, microfilm or any
other means without written permission
from the publisher
Sole distributors for U.S.A. and Canada:

PRINTED IN THE NETHERLANDS

INTRODUCTION TO THE SERIES
This is a series of books concerned with the quantitative approach to

problems in the behavioral science field. The studies are in particular in the
overlapping areas of mathematical economics, econometrics, operational
research, and management science. Also, the mathematical and statistical
techniques which belong to the apparatus of modern behavioral science
have their place in this series. A well-balanced mixture of pure theory and
practical applications is envisaged, which ought to be useful for universities
and for research workers in business and government.
The Editor hopes that the volumes of this series all of which relate to such
a young and vigorous field of research activity, will contribute to the ex¬
change of scientific information at a truly international level.
The Editor
88690
PREFACE
The use of probability theory in modern economics is at present substantial.

It is convenient to describe its impact on the development of the economic
discipline in terms of the following three stages:
(1) Probability theory is basic to statistics. Statistics is useful for eco¬
nomics (regression analysis, sample surveys, etc.). In this way probability
theory affected applied economics indirectly, viz., via the statistical methods
used.
(2) More recently we have seen the development of the theory of decision¬
making under uncertainty. The simplest way to handle the uncertainty aspect
is by means of probability concepts. This is how probability theory entered
directly into theoretical economics.
(3) I think that there is a third and much wider area in which probabilities
can play a useful role. It can be described conveniently by recalling a well-
known definition of economics: that of the discipline which is concerned
with the allocation of scarce resources to alternative ends. Consider any such
scarce resource and define the quantity unit such that the total available
quantity is 1. Allocation to n alternative ends then amounts to a specification
of n numbers px,..., pn for the quantities (fractions of the total available
quantity) which are to be used for the various ends. It is clear that these p’s
satisfy all requirements which are necessary in order that they can qualify
as the probabilities of some distribution, since they are nonnegative and add
up to 1. Indeed, probability theory in the usual sense of the word can be
regarded as a special allocation problem. The scarce resource with which it
is concerned is the confidence that we have in the things of the world. When
we are sure that some event E will happen, we allocate our full confidence to
E, which is formalized by the statement: The probability of £ is 1. When
there are n events Eu..., En, one of which will certainly take place (but we
do not know which), we allocate only partial confidence to each of them.
This is measured by their probabilities pl,...,pn, which are allocation pro¬
portions based on our confidence in their realization.
This book is concerned with allocation proportions in economics. The
method of analysis is that of information theory, which is part of probability
theory. Information theory was developed by communication engineers
PREFACE VII
(mainly after World War II); its basic concept is the logarithm of a proba¬
bility. It seems plausible that the liberal use of logarithms in economics for
many decades is at least partly responsible for the applicability of the
logarithmic probability concept as an important measure in economic
problems. The applications presented in this book cover a wide range.
They include the measurement of the inequality of income, of industrial
concentration and of concentration in international trade. They also include
the evaluation of surveys, the measurement of the fit of behavioral relations,
the aggregation analysis of input-output models, the price and quantity
index number problem, and the formulation of systems of demand relations.
The mathematical treatment is rather elementary (with the exception of
Chapter 6, parts of Chapter 7, and Chapter 9, where matrix algebra is used).
The book may therefore be useful as a text.
I am indebted to the Editors of Econometrica, the International Economic
Review, the Journal of Political Economy, the Review of Economic Studies
and the Review of Economics and Statistics for their permission to use in this
book material which was originally published in their journals. I am par¬
ticularly indebted to my colleagues A. P. Barten and T. Kloek of the
Econometric Institute, with whom I worked for several years in the area of
demand equations and index numbers. Finally, there are those who assisted
me in the preparation of this book: G. A. C. Beerens, J. Boas, A. Kunstman
and Myron Scholes, who programmed the computations; Myron Scholes
again, who read the manuscript and advised me on its intelligibility; Mrs.
Irene Grad, my secretary in Chicago, who typed the greater part of the
manuscript during an extremely busy period, improved its style, and also
read the second proof; Miss C. A. Berger, my secretary in Rotterdam, who
typed the rest and spent evenings and weekends reading the first proof; and
A. P. J. Abrahamse and C. Dubbelman, who also read the first proof while
working on their examination in the course Economics and Information
Theory. My gratitude goes to all of them.
Summer 1966 Henri Theil

ABBREVIATED CONTENTS
Introduction to the Series. v
Preface.VI
Abbreviated Contents.VIU
Table of Contents.IX
List of Tables.xrv
List of Figures.xvm
Outline of the Book.xix
1. The Information Concept. 1

2. Expected Information.24
3. Economic Relations Involving Conditional Probabilities.49
4. The Measurement of Income Inequality.91
5. A Statistical Approach to the Problem of Price and Quantity Com¬
parisons .135
6. The Consumer’s Allocation Problem.182
7. Empirical Implications of the Allocation Model of the Consumer . . 227
8. Industrial Concentration and the Allocation Problem of the Firm . . 290

9. Input-Output Analysis and Its Aggregation Problems.319
10. Information Measures in the Analysis of International Trade . . . 357
11. Continuous Information Theory and a Multiplicative Decomposition of
Prediction Error Variances.390
Bibliography.423
Tables.428
Index.483
TABLE OF CONTENTS
Introduction to the Series. v
Preface. VI
Abbreviated Contents. VIII
Table of Contents..
List of Tables.xiv
List of Figures.xvhi
Outline of the Book.xix
Chapter 1. The Information Concept
1. The Dog and the Chess-board. 1

2. The Information Content of a Definite Message. 3
3. An Axiomatic Approach to the Information Concept. 5
4. The Information Gain of Weather Forecasts; A Generalization of the
Information Concept. 8
5. The Information Gain of Survey Forecasts; The Relative Gain. . . 10
6. The Information Gain of Survey Forecasts, Given Last Period’s Pre¬
diction and Realization Experience.14
7. The Sampling Variability of the Information Value.18
8. Historical Notes.19
Appendix
Chapter 2. Expected Information
1. The Expected Information in a Determinative Message; The Entropy

Concept.24
2. The Expected Information of an Indirect Message.27
3. The Information Content of Survey Forecasts. 29
4. Bivariate Information Theory.33
5. The Information Inaccuracy of Decomposition Forecasts .... 35
6. The Information Inaccuracy of Input Structure Piedictions and the
Information Improvement of Forecast Revisions.39
7. The Aggregation Theory of the Information Inaccuracy.43
8. Summary of Concepts.46
X TABLE OF CONTENTS
Chapter 3. Economic Relations Involving Conditional Probabilities
1. Conditional Distributions and Their Entropies.49

2. The Determinants of Production Plan Revisions.52
3. Multivariate Information Theory.55
4. A Log-linear Model for Conditional Odds; The Logit ..... 59
5. Statistical Estimation of the Log-linear Model.65
6. A Partial Log-linear Model.69
7. Logit Regressions.71
8. Simulation of Logit Regressions.76
Appendix
Chapter 4. The Measurement of Income Inequality
1. Inequality at the Individual Level; The Aggregation Problem ... 91

2. Some Simple Continuous Income Distributions; White and Nonwhite
Lamilies in the United States.96
3. Two-level Aggregation: States and Sets of States in the United States 101
4. Income Inequality among Countries: 1949, 1957, 1976 . 107
5. Inequality Changes in the Case of Unchanged Population Shares . . 110
6. Migration and Its Effect on Per Capita Income; Maxwell’s Demon on
Ellis Island.114
Appendix
A. Notes on Some Alternative Measures of Income Inequality. . . . 121
B. Details on White and Nonwhite Lamily Incomes.128
C. Limits on Inequality when Income Brackets are Used.128
Chapter 5. A Statistical Approach to the Problem of Price and
Quantity Comparisons
1. Random Selection of Commodities.135

2. Coal Miners in the European Economic Community; The Lactor
Reversal Test and the Allocation Discrepancy.138
3. One-dimensional Price and Quantity Scales.146
4. Variation over Time: Consumption in the Netherlands, 1921-1963 . . 150
5. Variances and Covariances of Logarithmic Price and Quantity Differ¬
ences . 154
6. Partial Index Numbers and the Decomposition of Price and Quantity

Variances and Covariances.158
7. The Decomposition of the Dispersion of Value Share Changes: An In¬
formational Approach.163
TABLE OF CONTENTS XI
8. Concluding Remarks. 166

Appendix
A. Details on the Cross-section Data.169
B. Details on the Time Series Data.172
C. The Decomposition of the Variance of Value Share Changes ... 180
Chapter 6. The Consumer’s Allocation Problem
1. The Consumer’s Optimality Conditions and the General Form of

Demand Equations. 182
2. The Fundamental Matrix Equation of Consumer Demand Theory . 188
3. The Income Effect and the Substitution Effects of Price Changes. . . 191
4. Demand Equations in Infinitesimal Changes; Preference Independence
and Block-Independence.194
5. Demand Equations in Finite Changes; Real Income and Two Kinds of
Relative Prices.200
6. The Marginal Utility of Income as a Function of Income and Prices . 206
7. The Indirect Utility Function, Its Income Solution, and the True Cost
of Living Price Index.208
8. A Comparison of Cost of Living and Marginal Price Index Numbers . 212
9. Monotonic Transformations of the Utility Function.219
Appendix
A. Proof of the Lemma on Local Quadratic Approximation .... 222
B. The True Index of Real Income.223
Chapter 7. Empirical Implications of the Allocation Model of the
Consumer
1. A Reformulation of the Demand Equations; A Marginal Utility Shock

Model.227
2. Aggregation and Disaggregation of Demand Equations; Partitioning of
Demand Analysis.233
3. A Numerical Illustration; The Information Value of Demand Equations 237
4. The Information Value of Demand Predictions; The Impact of the
Random Variability of Coefficients and Disturbances.245
5. The Marginal Price Index and the Marginal Utility of Income; Standard
Errors of Their Log-changes.251
6. The Covariance of Price and Quantity Log-changes; Statistical Testing
of the Income Flexibility.. • 253
7. Moments of Log-changes Based on Marginal Value Shares; The
Luxury-Necessity Index.258
XII TABLE OF CONTENTS
8. Equivalent Income Changes.263

9. The Coal Miner Families Revisited: Cost of Living Comparisons at
Different Real Income Levels.268
Appendix
A. Breese and Chead: The Principal Components of Consumer Prefer¬
ences .276
B. Sampling Variances and Covariances of Coefficient Estimates . . . 280
C. Summary of the Main Symbols Used in Chapters 5, 6 and 7 . 282
Chapter 8. Industrial Concentration and the Allocation Problem of
the Firm
1. The Measurement of Industrial Concentration.290

2. Concentration in Different Geographic Areas: New Car Registrations
in the United States, 1959-1964 295
3. Further Details on the New Car Registrations.300
4. The Demand for Factors of Production: The Fundamental Matrix
Equation of Production Theory.302
5. Solving the Fundamental Matrix Equation in the Case of Constant
Returns to Scale; The Substitution Flexibility.306
6. The Special C.E.S. Case; The Problem of the Specification of the
Factor Demand Equations.311
Appendix
Chapter 9. Input-Output Analysis and Its Aggregation Problems
1. The Input-Output Technique: A Brief Summary.319

2. The Aggregation Bias of Intermediate Demand Predictions.... 322
3. The Aggregation Bias of Primary Demand Predictions.328
4. The Information Content of an Input-Output Table.331
5. The Information Decomposition Equation of Input-Output Aggre¬
gation .335
6. An Application to the Dutch Input-Output Tables, 1949-1960 . . . 341
7. Partial Disaggregation of the Food Industry Set.345
8. Input-Output Prediction Before and After Disaggregation; A Mini¬
max Interpretation of the Input Heterogeneity Criterion.348
Appendix
Chapter 10. Information Measures in the Analysis of International

Trade
1. Predicting Trade Flows from Total Exports and Total Imports . . 357
TABLE OF CONTENTS XIII
2. An Application to the Trade of Eight Regions.360

3. Concentration with Respect to Origin, Destination and Composition for
Separate Countries and Commodity Groups.365
4. Aggregative Measures of Concentration.372
5. Two Country Sets: The Common Market and the Rest.377
Appendix
A. The RAS Method for Adjusting Bivariate Frequency Tables . . . 388
B. Some Details on the Import-Export Data of Eight Countries ... 389
Chapter 11. Continuous Information Theory and a Multiplicative
Decomposition of Prediction Error Variances
1. The Entropy of a Continuous Distribution; The Case of the Normal

Distribution.390
2. Predicting the Future and Estimating the Past.392
3. The Distribution of the Prediction and Estimation Errors; A Multi¬
plicative Decomposition of Error Variances.397
4. The Adjustment of the Parameters of the Error Variances; Tests for
Normality.399
5. The Information Gain of the Next Stage.406
6. A Mean-square-error Evaluation of Individual Variables and Years. . 409
7. An Informational Appraisal of Stage 1.413
Appendix
Bibliography.423
Tables
A. Logarithms to the Base 2 and Natural Logarithms of Integers from 1

to 10,000 . 429
B. Logarithms to the Base 2 and Natural Logarithms of Reciprocals of
Probabilities.448
C. Logits to the Base 2 and Natural Logits.458
D. Transformation of Relative Changes to Log-changes and Vice Versa . 463
Index.483
LIST OF TABLES
Chapter 1
1.1. Information Gains and Relative Gains for Survey Forecasts of Ten
Variables.13
1.2. Information Gains for Eight Variables, Given the Prediction-Reali¬
zation Experience of the Month Before.17
1.3. Prediction-Realization Tables for the German Spinning Industry. . 22
1.4. Prediction-Realization Tables for the German Weaving Industry. . 22
1.5. Prediction-Realization Tables for the German Clothing Industry. . 23
1.6. Prediction-Realization Tables for the Danish Textile and Clothing
Industry.23
Chapter 2
2.1. The Information Content of Survey Forecasts of Ten Variables . 31

2.2. Predicted and Realized Allocations According to the Dutch Construc¬
tion Programme, 1950-1951 . 36
2.3. Input Structure of Agriculture, Forestry and Fishing in the Nether¬
lands, 1948-1950 40
2.4. Information Tableau of the Input Structure of Agriculture, Forestry
and Fishing, 1948-1957 41
2.5. Average Information Inaccuracy for Different Time Spans of Pre¬
diction .42
2.6. Total Sector and Primary Input Ratios and Conditional Input Ratios
of the Agricultural Sector.45
Chapter 3
3.1. Production Plan Revisions Determined by Surprises on Orders

Received and the Appraisal of Inventories.53
3.2. Conditional Frequencies of Ship Ticket Preference, Given Alter¬
native Combinations of Motives.70
3.3. Means of Point Estimates for the Case of One Explanatory Variable 79
3.4. Mean Square Sampling Errors and Mean Square Standard Errors for
the Case of One Explanatory Variable.81
3.5. Means of Point Estimates for the Case of Two Explanatory Variables 83
3.6. Mean Square Sampling Errors and Mean Square Standard Errors for
the Case of Two Explanatory Variables.84-85
3.7. Average Percentage of Missing Groups.86
3.8. Mean Square Standardized Residuals.87
LIST OF TABLES XV
Chapter 4
4.1. Income Distributions of White Families and of Nonwhite Families in

the United States, 1963 . 99
4.2. Average Income and Income Inequality for White and Nonwhite
Families, 1947-1963 . 100
4.3. Per Capita Income Inequality among the States of the United States
1929-1961 . 103
4.4. Population Shares and Income Shares of 54 Countries 1949 1957
1976 .’ 108
4.5. Per Capita Income Inequality among 54 Countries, 1949, 1957, 1976 109
4.6. Per Capita Income Inequality among 54 Countries, Given Fixed
Population Shares. 112
4.7. Per Capita Income Inequality among the States of the United States,
Given Fixed Population Shares. 113
4.8. Some Inequality Measures in the Two-Country Case. 118
4.9. Income Distributions of White and Nonwhite Families in the United
States, 1947-1951 . 129
States, 1952-1957 . 130
States, 1958-1962 . 131
4.12. Approximate Lower and Upper Limits to the Inequality Figures of
Table 4.2. 134
Chapter 5
5.1. Sub-regions and Their Weights within Their Region.139

5.2. Commodity Groups and Value Shares.140
5.3. Binary Comparisons of Price and Quantity Levels.142
5.4. One-dimensional Scales for Prices and Quantities.148
5.5. Discrepancies of the Binary Comparisons from the One-dimensional
Scales.149
5.6. Commodity Groups and Value Share Averages, The Netherlands
1921-1939, 1948-1963 . 152
5.7. Logarithmic Price and Quantity Changes and Allocation Dis¬
crepancies for Consumption Per Capita in the Netherlands, 1921-1963 153
5.8. Logarithmic Variances of Value Share, Price and Quantity Changes
and Logarithmic Covariances of Price and Quantity Changes for
Consumption Per Capita in the Netherlands, 1921-1963 .... 156
5.9. Same as Table 5.8 for the Cross-section Case . . . . . . 157
5.10. Log-changes in Partial Price Index Numbers for Food, Vice, Durables
and Remainder in the Netherlands, 1921-1963 . 160
5.11. Same as Table 5.10 for Partial Quantity Index Numbers .... 161
5.12. Decomposition of Variances and Covariances of Log-changes in
Prices and Quantities for Consumption Per Capita in the Netherlands,
1921-1963 164
XVI LIST OF TABLES
5.13. Logarithmic Variances of Value Share Changes and Their Information

Decomposition for Consumption Per Capita in the Netherlands,
1921-1963 167
5.14. Logarithmic Price Differences for 21 Commodity Groups . . . 170
5.15. Logarithmic Quantity Differences for 21 Commodity Groups . . 171
5.16. Value Shares of Sixteen Commodity Groups and Four Sets of Groups,
The Netherlands 1921-1963 . 173-175
5.17. Price Log-changes for Sixteen Commodity Groups.176-177
5.18. Quantity Log-changes for Sixteen Commodity Groups . . . 178-179
Chapter 7
7.1. Dependent Variables and Disturbances of the Demand Equations for

Four Commodity Groups.240
7.2. Information Inaccuracies of Alternative Demand Predictions. . . 243
7.3. Decomposition of the Expected Value of Average Information In¬
accuracies .249
7.4. Log-changes of Price Indices and of the Marginal Utility of Income . 252
7.5. Decomposition of the Covariance of Price and Quantity Log-changes 255
7.6. Variances and Covariances of Price and Quantity Log-changes
Weighted by Value Shares and by Marginal Value Shares . . . 259
7.7. The Luxury-Necessity Index and Its Decomposition.262
7.8. The Variance of the Equivalent Income Changes and Its Decompo¬
sition .267
7.9. Prior Means of the Marginal Value Shares of 21 Commodity Groups 271
7.10. Prior Means and Standard Deviations of Logarithmic Purchasing
Power Parities.273
Chapter 8
8.1. The Entropy of Passenger Car Production in the United States, by

Makes and Sets of Makes, 1936-1964 . 294
8.2. The Entropy and Its Geographic Decomposition of New Car Regis¬
trations in the United States, 1959-1964. 297
8.3. The Entropy of New Car Registrations in the United States, by Makes
and Sets of Makes, 1959-1964. 300
8.4. Percentage Shares of S4, Rambler, and Foreign Cars.302
8.5. Within-Set Entropies of New Car Registrations in the United States,
1959-1964 303
Chapter 9
9.1. The Information Decomposition of Twelve Input-Output Tables . 343

9.2. Contributions of Industry Sets to Input and Output Fleterogeneity . 344
9.3. The Gain in Information Content Obtained by Partial Disaggregation
of the Food Industry Set.348
9.4. Mean Square Errors of Intermediate and Primary Demand Predictions
LIST OF TABLES XVII
Before and After Partial Disaggregation of the Food Industry Set 350
9.5. Logarithmic Errors Before and After Partial Disaggregation of the
Food Industry Set..
Chapter 10
10.1. Import and Export Shares: 1938, 1948, 1951-52, 1959-60 ... 361
10.2. Information Inaccuracy Values of Two-stage Information Forecasts 362
10.3. Information Inaccuracy Values for Export Share and Import Share
Predictions.354
10.4. Eight Countries and the Percentage of Total Exports Covered... 365
10.5. Description of Ten Commodity Sets.366
10.6. Entropy Values for Individual Countries.369
10.7. Entropy Values for Ten Commodity Sets.373
10.8. Conditional and Unconditional Entropies.376
10.9. Shares of Total Trade of E.E.C. and the Rest.378
10.10. Mutual Information Values: E.E.C. versus the Rest.379
10.11. Mutual Information Values: E.E.C. versus the Rest for Individual
Countries.381
10.12. Mutual Information Values: E.E.C. versus the Rest for Individual
Countries by Commodity Sets. 383-386
Chapter 11
11.1. Seven Successive Forecasts and Estimates of the Log-changes in Five

Variables, 1953-1962 . 396
11.2. Adjusted Parameter Values of Error Variances and Some Moment
Measures of Standardized Errors. 402
11.3. Information Gains of Successive Stages. 407
11.4. Information Gains Per Month of Successive Stages under Alternative
Conditions of Imperfect Stage 7 Estimates. 409
11.5. The Coefficients Bt, B't, Bt of Ten Years for Three Mean-square Speci¬
fications . 413
11.6. The Coefficients A{, A\, A’l of Nineteen Variables for Three Mean-
square Specifications. 414
11.7. Two Alternative Forms of the Information Gain of Stage 1 Forecasts,
by Variables and Years.418-419
11.8. Description of Variables and Estimated Log-changes in 1961 and 1962 422
A. Logarithms to the Base 2 and Natural Logarithms of Integers from I

to 10,000. 429
B. Logarithms to the Base 2 and Natural Logarithms of Reciprocals of
Probabilities.44^
C. Logits to the Base 2 and Natural Logits.458
D. Transformation of Relative Changes to Log-changes and Vice Versa 463
LIST OF FIGURES
Chapter 1
1.1. The Information Content h{ ). 5
Chapter 2
2.1. The Expected Information H() for n=2.26

2.2. Contour Lines of Constant Expected Information of Indirect Messages
for n=2. 29
Chapter 3
3.1. The Logit as a Function of the Probability.64
Chapter 4
4.1. The Relationship between Net Migration Per State and State Per
Capita Income, 193(MD and 1940-50 115
4.2. A Lorenz Curve.122
Chapter 6
6.1. Illustration of the True Cost of Living Price Index and the True
Marginal Price Index.212
6.2. Illustration of the Intermediate Utility Level Ut.218
Chapter 11
11.1. Seven Successive Forecasts and Estimates.393

11.2. Frequency Distribution of 1063 Standardized Prediction and Esti¬
mation Errors. 494
11.3. Frequency Distribution of 190 Constant-change Extrapolation Errors 416
OUTLINE OF THE BOOK
This book can be divided more or less naturally into four parts. The first
consists of three chapters:
1. The Information Concept

2. Expected Information
3. Economic Relations Involving Conditional Probabilities
The main objective of these chapters is to introduce the basic technical in¬
formation concepts, illustrated with economic examples. This holds particu¬
larly for Chapters 1 and 2. Since information is defined in terms of logarithms
of probabilities, an immediate extension of the subject is the analysis of
relations which describe probabilities in terms of certain determining factors.
For example, the probability of car ownership as a log-linear function of
family income, family size and similar variables. This is pursued in Chap¬
ter 3.
The remainder of the book is devoted to specific classes of economic
problems. The second part consists of four chapters:
4. The Measurement of Income Inequality

5. A Statistical Approach to the Problem of Price and Quantity Compari¬
sons
6. The Consumer’s Allocation Problem
7. Empirical Implications of the Allocation Model of the Consumer
The family household is in the center of the discussions of these chapters.

How is income distributed among the families of a nation? How is personal
income distributed among the states of a nation? These are the questions
considered in Chapter 4, where it is argued that information concepts
provide an appropriate measure. How should we compute the change in the
cost of living and in real income, over time and between different regions?
This question is considered in Chapters 5 and 6. The approach of Chapter 5
is largely statistical in nature: How large are the relative price and quantity
differences on the average when the commodities are appropriately weighted?
XX OUTLINE OF THE BOOK
Chapter 6 is based on classical economic theory. It defines the cost of living

index in terms of the amount of (money) income which is required for a given
standard of living, and the real income index in terms of the amount of in¬
come which the consumer can afford to spend in a given price situation. In
spite of the difference of these approaches it appears to be possible to obtain
an elegant synthesis.
The usefulness of information concepts in this type of analysis is due to the
prominent role of value shares: the amount spent on a particular commodity
divided by total expenditure. Such value shares are nonnegative and add up
to 1, so that they are mathematically equivalent to probabilities - and
probabilities are the basic ingredients of information concepts. Value shares
play also an important role in the analysis of consumer demand. Classical
demand theory takes income and price changes as given and uses these to
explain the changes in the quantities bought. The demand equations which
are introduced in Chapter 6 have as dependent variable the quantity com¬
ponent of the change in the commodity’s value share. This is particularly
convenient when the statistical approach of Chapter 5 is extended to higher
moments (such as the covariance of price and quantity changes) which are
expressed in terms of income and price changes.
It should be mentioned that Chapter 6 is exceptional among all chapters of
this book to the extent that it is exclusively theoretical. The empirical
illustrations are given in Chapter 7. The only honest advice to the reader is to
re-read the appropriate parts of Chapter 6 if his memory appears to be in¬
sufficient.
The third part of the book, consisting of three chapters:
8. Industrial Concentration and the Allocation Problem of the Firm

9. Input-Output Analysis and Its Aggregation Problems
10. Information Measures in the Analysis of International Trade
contains an extension of the analysis to problems of the firm and of inter¬

national trade. Chapter 8 does for the firm, in abbreviated form, what
Chapters 4 and 6 do for the consumer. Its first half is devoted to the measure¬
ment of industrial concentration, which is a subject similar to that of income
inequality; the second half deals with the formulation of demand equations
for production factors. In Chapter 9 firms are combined to industries in the
form of an input-output table. Industries are then combined to industry
groups, so that a smaller input-output table is obtained, and the question
asked is: If the familiar input-output prediction technique is applied at this
OUTLINE OF THE BOOK XXI
aggregative level, what is the error committed and how is it possible to

control this error? It appears that, within limits, an informational criterion
of the input heterogeneity of industries is suitable for this purpose.
Chapter 10 deals partly with concentration problems in international
trade: concentration with respect to composition, to origin (exporting
countries) and to destination (importing countries). It deals also with the
prediction of the trade between pairs of countries when total exports and
total imports of each country are given. The forecast technique is of the in¬
formational type and turns out to be closely related to the RAS method
which was originally formulated for input-output forecasts.
The fourth and last part consists of one chapter:
11. Continuous Information Theory and a Multiplicative Decomposition

of Prediction Error Variances
Here information is defined in terms of continuous rather than discrete

distributions, probabilities being replaced by probability density functions.
The continuous information measure is then applied to the evaluation of
forecasts for a number of variables in a number of years. The basic idea is
quite simple: When the actual data (the “true” realizations) contain more
information, given the forecast made, the forecast is worse. When there is
little information in the truth, given what was predicted to happen, the
prediction is accurate. This idea is applied to successive forecasts made by
the Netherlands Central Planning Bureau as well as to estimates and revisions
of estimates made by the Central Bureau of Statistics.
As stated earlier, information is defined in terms of logarithms of proba¬

bilities (and density functions). The base of the logarithms determines the
informational unit. Two different bases are commonly used: 2 and e, both of
which are also applied in this book. However, in order to avoid confusion we
shall never use both logarithms to the base 2 and natural logarithms in the
same chapter. Logarithms to the base 2 will be used in the first part (Chap¬
ters 1 through 3) and in the third (Chapters 8 through 10). Natural logarithms
will be used in the second part (Chapters 4 through 7) and in the fourth
(Chapter 11).
At the end of the book there are four tables which should facilitate the use
of the logarithmic measures:
A. A table of logarithms to the base 2 and natural logarithms for integers
^10,000.
XXII OUTLINE OF THE BOOK
B. A table of log (1 /p), where p is a probability. This table contains both

logarithms to the base 2 and natural logarithms.
C. A table of the so-called logit, log (p/q) where q=\—p, also both in
natural logs and logs to the base 2.
D. A table transforming relative changes (percentage changes divided by
100) into log-changes (changes in the natural logarithm) and vice versa.
Log-changes are used on a large scale in Chapters 5 through 7 and 11.
CHAPTER 1
THE INFORMATION CONCEPT
1.1. The Dog and the Chess-board
Your dog ran away. You have no idea where he is, but you do know that
he must be in a certain rectangular field, which looks like a chess-board
because it is divided into 64 squares. The problem is: In which square
is the dog?
1 9 17 25 33 41 49 57
2 10 18 26 34 42 50 58
3 11 19 27 35 43 51 59
4 12 20 28 36 44 52 60
5 13 21 29 37 45 53 61
Dog
6 14 22 30 38 46 54 62
7 15 23 31 39 47 55 63
8 16 24 32 40 48 56 64
The chess-board shows that the dog is in the 53rd square. You don’t know
this, however. Your task is to ask a number of consecutive questions, each of
which can only be answered by Yes or by No, such that you will know by the
end in which square the dog can be found. Each question asked will cost you
one dollar. You are therefore interested in formulating your questions in
such a way that their number is as small as possible.
The simplest question technique is in terms of separate squares:
Question: Is the dog in the first square?
1
2 THE INFORMATION CONCEPT
Answer: No.
Question: Is the dog in the second square?
Answer: No.
Etcetera.
If this procedure were adopted, you would have to pay $ 53. With a great deal
of luck it might be only $ 1, viz., if the dog had been in the first square, but
it is equally possible that you will have to pay as much as $ 64. Obviously,
it is in your interest to find out whether a cheaper method exists. It is im¬
mediately evident that there exists no competing method which is cheaper
under all circumstances. The dog may be in the first square, in which case the
method just described cannot be defeated. However, it is conceivable that
there is a method which is cheaper on the average.
To find such a method, we shall proceed in a more systematic, selective
manner. Consider the following series of six questions:
Question 1: Is the dog in one of the first four columns?
Answer: No.
This rules out one-half of the possibilities, so that 32 squares remain.
Question 2: Is the dog in the fifth or sixth column?
Answer: No.
Question 3: Is the dog in the seventh column?
Answer: Yes.
The only squares left are the eight of the seventh column.
Question 4: Is the dog in one of the squares 49-52?
Answer: No.
Question 5: Is the dog in one of the squares 53-54?
Answer: Yes.
Question 6: Is the dog in the 53rd square?
Answer: Yes.
We have thus succeeded in finding out the dog’s place after asking 6
questions. This is always possible when there are 26 = 64 possibilities. More
generally, whenever we have 2N possibilities, N questions - each of which is
answered by Yes or by No - are sufficient to establish beyond doubt which
of the 2N possibilities actually applies.
We know now that you have to pay $ 6 to find out in which square you
can find your dog. Is this on the average cheaper than the square-by-square
method? We shall answer this question on the assumption that each square
has the same probability -g1^ of containing the dog. If the dog is in the /th
square, /= 1, ..., 64, i questions and hence i dollars are needed to verify this.
Hence the expected amount spent on the square-by-square method is
1.2 A DEFINITE MESSAGE 3
6*4 ($ 1 + $24-1- $64) = $ 32.50
This is more than five times larger than the amount which is really needed.
We conclude that it pays off, on the average, to ask the questions in a
selective manner.
1.2. The Information Content of a Definite Message
It is worthwhile to stress that we assumed at the end of the previous section

that all 64 squares have the same probability of containing the dog. It remains
true, even if the 64 probabilities are different, that 6 questions are sufficient
to establish beyond doubt where we can find the dog, but it is conceivable
that a still smaller number of questions is sufficient (on the average) when we
know more. Suppose, for example, that it is known that sausages are stored
in the 53rd square. Then it might be advantageous to start with a question on
that square - you would have lost only $ 1 in that case! - and then proceed
systematically with the other squares. However, let us leave the dog where he
is and continue at a more abstract level.
Suppose it is known that some event E will occur with probability x,
0<x<l. Suppose that at some later stage you receive a definite and reliable
message which states that E has indeed occurred. When x=.99 you will not
be surprised at all, for it was practically sure that E would take place. In
other words, the message has very little “information content” when x is
close to 1. The situation changes in this respect when x takes smaller and
smaller values. When x is close to 0, for example, x=.01, we are greatly
surprised by the message, for it was practically sure that E would not take
place. The message has then a very large “information content.”
Our objective is to give a more precise meaning to “the information con¬
tent of a definite and reliable message.” It should be stressed that the infor¬
mation concept which we shall define has nothing to do with the emotional
content of the message. It will be defined as a function of only one argument,
viz., the probability x that the event would take place before the message
came in. Thus, if the chance of a multiple birth is .01 for every couple, the
doctor’s message which states that this occurred to you has the same infor¬
mation content as the message which states that this happened in Mr. Jones’s
family.
It seems obvious to require that the information content h(x) of the
message be a decreasing function of the probability x: The more wnlikely the
event before the message on its realization, the larger the information content
of this message. This is in accordance with the x—.99 and x = .01 examples
of the second paragraph of this section. In principle one is completely free in

the choice of this decreasing function, but it is generally agreed to take the
logarithm of the reciprocal of the probability x:
(2.1) h (x) = log - = — logx

x
The reason for the choice of this logarithmic function among all decreasing
functions is the additivity in the case of independent events. Thus, let Et and
E2 be events with probabilities xt and x2, respectively, and suppose that these
events are stochastically independent, so that XjX2 is the chance that both
events occur. The information content of the message which states that both
Ex and E2 did occur is then
(2.2) h (xjx2) = log-- = log — + log — = h (x}) + h (x2)

*1*2 xt x2
that is, equal to the sum of the information content of ttE1 occurred” and the
information content of “E2 occurred.”
We now return to the selective approach of Section 1.1, which showed that
6 questions were sufficient to find out which of the 26 squares contains the
dog. On the assumption that all squares have the same chance 2~6, we con¬
clude from (2.1) that the definite and reliable message which states in which
square the dog can be found has the following information content:
(2.3) /7(2~6) = log = log26 = 6
provided that 2 is chosen as the base of the logarithm. We conclude that the
number of questions to be answered in the selective approach is equal to the
information content of the message which the dog owner wishes to receive.
This holds more generally. When there are 2N possible events, all with the
same chance 2 N, the information content of the message which states re¬
liably that one particular event occurred is
(2.4) fc(2-*) = logTs = JV
provided again that the base of the logarithm is 2. The definition (2.1) is,
of course, still more general, because x need not be a power of When this
is not the case, the information content /z( ) will take non-integer values.
The use of logarithms to the base 2 is very common in information theory.
It has the convenient property that the information unit corresponds to x = -§-.
1.3 AN AXIOMATIC APPROACH 5
a fifty-fifty event: h(f) = 1. If we use 2 as a base, the information content

is said to be expressed in “binary digits,” or for short, bits. The information
computed in (2.3) and (2.4) are then 6 bits and N bits, respectively, and 6
bits is the amount of information which you need to find out where your dog
is (on the assumption that all 64 squares are equally likely). Besides loga¬
rithms to the base 2 there are also the natural logarithms which have a fairly
prominent place in information theory. That type is more useful when we
need derivatives or Taylor expansions of logarithms. We shall therefore use
both types in this book. Following D. K. C. MacDonald,1 we shall call the
information unit a nit when natural logarithms are used. We have
(2.5) 1 bit = .693 nit 1 nit = 1.443 bit
in view of
1
loge2 = .693
L443
1.3. An Axiomatic Approach to the Information Concept
The information definition (2.1) implies that h( ) decreases monotonically

from oo to 0 when x increases from 0 to 1. This is illustrated in Figure 1.1,
. “Information Theory and Its Application to Taxonomy," f«/ +P** ***>,

Vol. 23 (1952), pp. 529-531.
which contains two vertical axes. The axis on the left measures information
in bits, the axis on the right in nits. The function is tabulated at the end of
this book.
The objective of the present section is to show that the logarithmic in¬
formation definition is the direct result of some rather natural axioms, given
that it is decided to make information dependent on the probability a only
(which is the first axiom). The second axiom states
(3.1) h (x) is a continuous function of x, 0 < x ^ 1
The third expresses our infinite surprise when we are informed that something
happened which had zero probability, and our zero surprise when we are told
that something happened which had unit probability:
(3.2) h{ 0)=oo h( 1) = 0
The fourth axiom states that h( ) is a monotonically decreasing function:
(3.3) h(x1)>h(x2) if 0 < xx < x2 < 1
The fifth states that there is additivity in the case of independent events:
(3.4) h (xjX2) = h (xx) + h (x2) if 0^x1,x2^l
We shall now give an outline1 of the proof that these axioms necessarily
lead to the conclusion that /;( ) has to be specified in accordance with (2.1),
which is unique except for the choice of some positive number as the loga¬
rithmic base. Let us start with (3.4) and introduce
1
(3.5) yt = log i — 1,2
*i
so that Xi — e~yi on the assumption that log stands for natural logarithm.
Hence the two right-hand functions of (3.4) can be written as
(3.6) h (x,) = h (e~y‘) — h* (y,) say i = 1,2
J The reader who prefers rigor is referred to A. I. Khinchin, Mathematical Foundations

of Information Theory (New York: Dover Publications, Inc., 1957), pp. 9-13. The set of
axioms chosen by Khinchin is slightly different from (3.1)—(3.4); the proof given here is
largely based on B. O. Koopman and G. E. Kimball, “Information Theory,” Chapter 9
of Notes on Operations Research 1959, assembled by the Operations Research Center,
M.I.T. (Cambridge: The Technology Press, M.I.T., 1959).
1.3
AN AXIOMATIC APPROACH 7
or equivalently /?*(- log xt). In the same way we can write for the left-hand
side of (3.4):
h(x1x2) = h*(- logx1x2) = h*(- logxx - logx2) = h*(y1 + y2)
On combining this with (3.6) we conclude that (3.4) amounts to
(3-7) h*(yt + y2) = fc*(yl} + h*(y2) yuy2 > 0
Furthermore, since yl as defined in (3.5) is a decreasing function of jq, the

fourth axiom stated in (3.3) is equivalent to
(3-8) h*(yi)<h*(y2) if 0^yi<y2
We proceed to substitute yx=y2 = 0 in (3.7), which gives A*(0) = 2/z*(0)

and hence h*(0) = 0 in accordance with h{ 1) = 0, see (3.2). [The alternative
solution A*(0)=oo is contradicted by /?(1) = 0.] Next we write /z*(l) = c and
consider, using (3.7):
h* (2) = h*(l) + h*(l) = 2c
h*(3) = h*(2) + h*(l) = 3c
/i+ (n) = nc (n — integer)
Note that c must be positive in view of (3.8). Furthermore, assuming that m

is a nonnegative integer and n a positive integer, we have
me = h* (m) = h*
where the second equality sign is based on repeated application of (3.7). We

conclude:
m
(3.9)
n
which means that h*(y) = cy, c> 0, holds for any nonnegative rational y. A
limiting process based on the second axiom, (3.1), can be employed to show
that h*(y) = cy applies to irrational y as well. But this is equivalent, in view
of (3.6) and (3.5), to
(3.10) h (x) = c log -

X
This, in turn, is equivalent to (2.1), the specification of c being a matter of

appropriate choice of the logarithmic base.
1.4. The Information Gain of Weather Forecasts; A Generalization of the

Information Concept
A weather forecast can be regarded as a message from the Weather Bureau
to the public. In this section we shall make an attempt to evaluate such a
message in informational terms.1 This is not a trivial application of the ideas
of Section 1.2, since we are all aware of the fact that weather forecasts are
not necessarily reliable messages. In fact, the analysis will lead to a gener¬
alization of the information concept.
We shall consider sunshine predictions, in particular predictions of the
following event, which will be indicated by E: “There is sunshine on day t
during a period which is less than 30 per cent of the time from sunrise till
sunset.” That is, there are either only clouds, or there are so many clouds on
that day that the sun is not visible during a proportion of at least 70 per
cent of the total time when it could have been visible.
The first forecast of this nature is made by the Royal Netherlands Meteoro¬
logical Institute 30 hours prior to the beginning of day t. Before that time
no sunshine forecasts are available, so that the only thing which is known is
the so-called climatological chance of E, which is defined as the observed
frequency of E during a long period. Of course, the climatological chance is
not constant throughout the year. If our day t is in February with its many
clouds, the chance is .65 that the sunshine is less than 30 per cent. If we
pick out a day at random, however, giving all 365 days of the year the same
chance of being selected, the chance of E is only .50:
(4.1) x0 = .50
This is the climatological chance of E which we shall use from now on. Given
that chance, the information content of the message which states reliably
“E occurred” (“There was less than 30 per cent sunshine on day t”) is there¬
fore
(4.2) /i(x0)=lbit
We now consider the first forecast, which as stated above is made 30 hours
prior to the beginning of day t. Specifically, the forecast is “E will occur”
1 The numerical data of this section have been taken from a study by N. Klijn and
W. J. Zwezerijnen described by H. Theil in Applied Economic Forecasting (Amsterdam-
Chicago: North-Holland Publishing Company and Rand McNally & Company, 1966),
Chapter 9. The reader is referred to this book for further details.
1.4 WEATHER FORECASTS 9
(“There will be less than 30 per cent sunshine on day r”). Obviously, one
should not expect that such a forecast will be perfect in the sense that it will
never occur that there is more sunshine on day t than 30 per cent. But one
should expect that, if the forecast has some merit at all, E will occur more
frequently when that forecast has been made than when the forecast has not
been made. In other words, we should expect that the conditional probability
of E under the condition that E has been predicted to occur will exceed the
climatological chance a0, which is the unconditional probability. Using the
data of all days of the year 1961, we find that this is indeed the case. We have
(4.3) Xl = .68
for the probability of E, given that E is predicted to occur 30 hours prior

to the beginning of day t. Clearly, if we take xt as a starting point, the
information content h(xl) of the reliable message “E occurred” is less than
the value (4.2). This is the natural consequence of the fact that the forecast
has some “information value” of its own. If Eis predicted to occur, we should
be less surprised to hear later on that E did indeed occur than in the case
when no forecast had been made. The difference between the two h values,
(4.4) h (x0) — h (xj) = log 1 = .44 bit

x0
can be regarded as a measure of the value of the prediction.1 It will be called

the information gain of the weather forecast. Its numerical value is positive
when the conditional probability - given the forecast - exceeds the climato¬
logical chance, zero if the two probabilities are equal, and negative if the
latter exceeds the former. The third possibility is evidently the catastrophe
from the forecaster’s point of view (also from the viewpoint of those who
believe him!).
The first forecast is followed by three others at twelve-hour intervals. So
we have a probability x2 that E will occur, given that it was predicted to
occur 18 hours prior to the beginning of day t, and similarly x3 and x4.2 The
1 Note that this measure treats all days of the year uniformly by applying xo = .50 to each
day. A more refined procedure is that of computing (4.4) for separate months, so that the
climatological chance is (approximately) constant; but this requires more data than are
available.
2 It should be noted that a “day” in the sense in which the word is used by the Royal
Netherlands Meteorological Institute starts at 6 p.m. of the day before (“ordinary” time
scale) and ends at 6 p.m. of the day itself. Hence the fourth forecast, which is made 6 hours
after the “day” started, is a midnight prediction of sunshine from next sunrise till next sunset.
numerical values of these probabilities are as follows:
(4.5) x2 = -69 x3 = .80 x4 = .88
and the corresponding information gains are:
h(x1) — h (x2) = log — = .02 bit

xx
(4.6) fi(x2) — /i(x3) = log 3 = .21 bit

x2
/;(x3) — h(xf) = log — = .14 bit

*3
We conclude that the successive probabilities increase, so that the information

gain of each forecast over its predecessor is positive. Obviously, the numerical
outcomes presented here are subject to sampling fluctuations. We shall con¬
sider the sampling problem in Section 1.7.
The information gain formula can in principle also be used in much more
general situations, quite apart from its use in forecast evaluation. If we com¬
pare (4.4) with (2.1), we see that the only difference is that the numerator
1 after the first equality sign in (2.1) is replaced by the probability x3 of
“E occurred” after the message (the first forecast) is received. This suggests
the following generalization of the definition (2.1):
(4.7) Information received probability ex post

log-
with message probability ex ante
where
probability of the event after
probability ex post =
the message is received
(4.8)
probability of the event before
probability ex ante =
the message is received
Clearly, the definition (4.7)-(4.8) includes (2.1) as a special case, viz., the
case in which the message is completely reliable. In the weather example: the
forecast which is correct with probability 1. We have thus succeeded in gener¬
alizing the information concept to messages whose reliability is incomplete.
1.5. The Information Gain of Survey Forecasts; The Relative Gain
We shall now apply the information gain concept to economic data dealing
with forecasts of the directions of change in certain economic variables. These
1.5 SURVEY FORECASTS 11
forecasts and the corresponding realizations are obtained by written question¬

naires sent out by various institutions to a number of firms. The participants
are asked to predict “increase,” “no change,” or “decrease” for the develop¬
ment of certain variables in the next period, and also to report “increase,”
“no change,” or “decrease” for the observed development of these variables
in the period just past. Our objective is to evaluate such forecasts in the light
of the reports on realizations supplied by the forecasting participants them¬
selves. For that purpose we shall use data on a number of variables that were
collected by the Ifo-Institut in Munich, Germany, the Konjunkturinstitutet in
Stockholm, Sweden, and the Central Bureau of Statistics in The Hague, The
Netherlands.1 Most of such surveys are on a monthly basis, but the Swedes
have their survey once every quarter.
Let us indicate “increase” by the index 1, “no change” by 2, and “decrease”
by 3. Let us write for the relative frequency of all cases in which some
participant predicts a Type i change in a certain variable for the next period
and reports one period later that there was actually a Type j change, where
i and j take the three values just mentioned. Thus,/X1 is the frequency of
all correct “increase” forecasts,/12 is the frequency of those cases in which
“increase” was predicted for the next period but in which “no change” was
reported as the realization,and so on. As a whole, the/’s form a 3 x 3 array:
Realization Marginal
increase no change decrease

Prediction: increase fi i ft 2 /13 fu
no change fu f22 fi 3 fi.
decrease /31 f3 2 f3 3 h.
Marginal 1
/1 f. 2 f. 3
The f and f} in the fourth column and fourth row are the marginal fre¬
quencies:
3 3
(5.D ft. - y fu fj - z fu
7=1 i=1
They measure the total frequency of Type i forecasts and of Typey reali¬
zations, respectively.
1 The data used are taken from Chapter 12 of Applied Economic Forecasting, to which
we refer for further details.
As an example, let us consider such a “prediction-realization table” for

Purchase prices:

Prediction: increase .111 .043 .003 .157
no change .039 .523 .089 .651
decrease .004 .069 .119 .192
Marginal .154 .635 .211 1
We observe that in 15.7 per cent of all cases an increase was predicted, that
this forecast was correct in 11.1 per cent, that “no change” was reported in
spite of the increase forecast in 4.3 percent, and “decrease” in .3 per cent. It
is easily seen that the diagonal elements fu are the correct forecast frequen¬
cies: 11.1 per cent for increase, 52.3 per cent for no-change, and 11.9 per cent
for decrease.
Let us apply the information gain concept to the increase forecast. On the
basis of the evidence of the table we conclude that before the forecast is made an
“increase” realization has a probability.//= .154. [One may argue that this
figure corresponds to the climatological chance of the previous section.]
After the forecast is made this realization has a probability
fi i .ill
.707
A Tl57
because the chance of an increase forecast is .157 and a fraction of such

forecasts turns out to be realized as “increase.” Hence the information gain
of an increase forecast for Purchase prices is
7u fn .111
h(/.i) - M 7^ ) = log-jly- = log = 2.20 bits
VI. / Jl.J.l (.157)(,154)
We can proceed in the same way for no-change and decrease forecasts. The
general formula is
(5.2) h (/,•) - h i = 1,2,3
The results are shown in the first three columns of Table 1.1, both for Purchase
prices and for nine other variables. We conclude that, typically, the informa¬
tion gains of increase and decrease forecasts are of the order of 1 or 2 bits,
whereas the no-change gains are considerably smaller.
TABLE 1.1
INFORMATION GAINS (iN BITS) AND RELATIVE GAINS FOR SURVEY FORECASTS OF TEN VARIABLES
No No
Variable Increase Decrease Increase Decrease
Change Change
Information gain (bits) Relative information gain

Purchase prices 2.20 .34 1.55 .815 .518 .693
Sales prices 1.99 .22 1.75 .655 .416 .721
Work force 1.89 .17 2.06 .686 .349 .721
Production 1.40 .19 2.07 .606 .328 .707
Investment in machinery 2.02 .36 .75 .596 .376 .556
Rate of profit 2.13 .25 .93 .551 .320 .611
Sales .68 .62 .77 .537 .255 .584
Raw materials bought 1.12 .20 1.27 .504 .244 .593
Orders received 1.01 .20 1.15 .496 .212 .552
Inventories 1.18 .17 1.00 All .224 .457
One reason that the information gain of no-change forecasts tends to be on

the low side is that for most of the variables the no-change realization fre¬
quency/^ is so large that there are not many bits to be gained. To understand
this effect we return to the weather forecasts of the previous section, in partic¬
ular to equation (4.2), which specifies that a reliable message “There was less
than 30 per cent sunshine on day /” has 1 bit information content, given the
climatological chance xo = .50. This information content is reduced by the
first forecast, but the reduction can never be larger than 1 bit. The actual
reduction obtained was .44 bit, see (4.4), so that 1 — .44=.56 bit is what re¬
mains for the second forecast. Its information gain cannot be larger than
that. It goes on in this way for the third and the fourth forecasts. The upper
limit to the information gain becomes smaller and smaller, provided that the
previous gains are positive.
In the present case of survey forecasts the maximum value of the infor¬
mation gain (5.2) is /?(/;), which is obtained when/,; =/• and hence h(fn/f ) = 0.
This condition implies that all Type i forecasts are realized as Type i changes.
For i= 1: Whenever an increase is predicted, this is correct in the sense that
the participant reports one period later than there was indeed an increase.1
1 Note, however, that there may be realized increases that were predicted before as
no-change or decrease: fn,fzil> 0. Hence the condition implies that the increase forecasts
are perfect, not that there is perfect prediction of all increase realizations.
Now the attainable maximum h(ff) is, for most of the variables, much larger
for 7 = 1 and 3 (increase and decrease) than for i=2 (no change). The reason
is that h(fi) is a decreasing function offt and that/1,/3</2 holds for
almost all variables. This is illustrated by the prediction-realization table for
Purchase prices which was presented above: f x = .154,f2 = .635,/3 = .211.
This argument suggests that it may be worthwhile to measure the infor¬
mation gain as a fraction of its maximum value:
(5.3)
h(f.d
log -
.i
The numerical values of these ratios are presented in the last three columns of
Table 1.1. They indicate that the information gains of increase and decrease
forecasts, when measured as a fraction of their maximum values, vary between
about 50 and 80 per cent. The picture of the no-change forecasts is better in
these relative terms than in absolute bits. Still, it is not very good. The
no-change information gains vary between 20 and 50 per cent of the attainable
maximum, which is decidedly below the corresponding figures for increase
and decrease forecasts.
1.6. The Information Gain of Survey Forecasts, Given Last Period’s Prediction
and Realization Experience
There is no need to stress that the figures of Table 1.1 may change de¬
pending on the circumstances. It is conceivable, for example, that the increase
forecasts are better when an increase was predicted correctly in the month
before. This was analyzed recently by Anderson and Bauer for monthly
German data and by Munksgaard for monthly Danish data.1 Both sets of
data deal with the textile and clothing industry. We shall use them here for
an evaluation in informational terms.
As an example we take Production in the German clothing industry in the
1 O. Anderson and R. K. Bauer, “Uber eine KT-Untersuchung unternehmerischer ex

ante-Ansichten in Abhangigkeit von den ex ante- und ex post-Ansichten des Vormonats,”
mimeographed report of the Institute of Statistics of the Mannheim School of Economics
read at the CIRET Conference in Vienna, April 1963. H. Munksgaard, “Einige Resultate
des Danischen Konjunkturtests 1956-1959,” mimeographed report of the Odense School
of Commerce read at the CIRET Conference in Rome, April 1965. The data used in this
section are taken from these reports and are reproduced in the Appendix of this chapter.
1.6 SURVEY FORECASTS AND LAST PERIOD’S EXPERIENCE 15
period 1950-1957. First, we consider all cases in which the participant pre¬
dicted correctly that production would be raised in month t. The prediction-
lealization table (in absolute rather than relative frequencies) for the next
month t + l is then as follows:

Prediction: increase 501 194 21 716
no change 217 222 47 486
decrease 4 17 9 30
Marginal 722 433 77 1232
Next we take all cases in which an increase was predicted for t but in which
the realization turned out to be no-change. Hence there is a prediction error
in month t. For this set of data we have the following prediction-realization
table in t+ 1:

no change 115 301 35 451
decrease 7 15 21 43
Marginal 269 450 61 780
Third, we consider all cases in which no-change was predicted correctly in

month t, which leads to the following table for month t+1:

no change 456 2692 267 3415
decrease 20 108 135 263
Marginal 626 3006 423 4055
We can go on in this way, since there are 3 x 3 = 9 possible configurations

in month t and hence also nine different prediction-realization tables in month
t+ 1, given the experience in month /. However, the three examples will be
sufficient to illustrate our point. Consider an increase prediction for month
t+ 1. In the first case (given a correct increase prediction in t) this forecast is
correct in a fraction 4re = .70 of all cases. In the second case (given an
increase prediction in t which turned out to be in error) the correct forecast
fraction is .51, which is less. In the third case (given a correct no¬
change prediction in t) it is .40, which is still less. The conclusion
seems to be that the value of the increase forecasts decreases in these three
successive situations.
Let us consider the merits of the increase forecasts in informational terms,
using the gain concept (5.2). The result is that the information gain is .26
bit in the first case (given a correct increase prediction in t), .58 bit in the
second case (given the incorrect increase prediction in t), and 1.37 bit in the
third case (given the correct no-change prediction in t). This suggests that the
increase forecasts become successively more valuable! How should we inter¬
pret these entirely different results?
The answer is that the information gain criterion is only partly interested
in the correct forecast fraction/u//j.. It argues as follows. In the first case
(given a correct increase forecast in month t) the unconditional frequency
of an increase realization is fA = f — -59. This is known before any
forecast is made. When an increase is predicted, the .59 value is raised to
fill ft. — -70- This is a fairly high figure, which implies that the forecast is
correct in a substantial majority of all cases. On the other hand, the infor¬
mation gain argues that such an increase prediction is merely a message
which states that the .59 probability is to be raised to .70, which is a rather
small relative increase. The corresponding information value is therefore only
l°g '-59= -26 bit. In the second case (given the prediction error) the uncon¬
ditional frequency f A is -§-§-§■ = .34, which is raised by 50 per cent to/u//i. = .51
when an increase is predicted. Although the correct forecast fraction is
lower than in the first case, the information gain is larger. This is evidently
quite reasonable. The gain concept measures the value of a forecast given
what is already known, which in this case amounts to: given the prediction
experience in month t. We could have a situation in which 95 per cent of
the forecasts are correct (/u//j = .95) but which is nevertheless without any
value as far as these predictions are concerned because fA = .95. This is
correctly recognized by a zero value of the information gain, but not by the
overly optimistic fraction of correct forecasts. When somebody predicts that
some event E will occur, it is valuable to know that the chance is x that the
forecast will turn out to be correct, and it is gratifying (from the standpoint
of predictive accuracy) to know that x is close to 1. But it is at least as
important to know how much larger x is than the probability y that E will
take place anyhow, and when x is not larger than y the forecast has no merit
at all. This is, of course, completely analogous to the case of our weather
forecasts in Section 1.4.
1.6 SURVEY FORECASTS AND LAST PERIOD’S EXPERIENCE 17
TABLE 1.2
INFORMATION GAINS (IN BITS) FOR EIGHT VARIABLES, GIVEN THE PREDICTION-REALIZATION
EXPERIENCE OF THE MONTH BEFORE
Last month’s experience

Variable Correct Correct Incorrect All cases
same sign other sign
Production Increase
German spinning industry .28 2.19 1.11 2.14
German weaving industry .33 2.05 .96 1.63
German clothing industry .26 1.40 .70 1.01
Danish textile and clothing .32 1.71 .71 1.09
industry
Sales prices
German spinning industry .21 1.97 1.02 1.41
German weaving industry .34 2.71 1.18 1.89
German clothing industry .36 2.52 1.03 1.93
Danish textile and clothing .56 3.21 1.15 2.46
industry
Production No change
German spinning industry .02 .34 .11 .06
German weaving industry .05 .42 .13 .13
German clothing industry .09 .37 .22 .24
Danish textile and clothing .12 .32 .13 .22
industry
Sales prices
German spinning industry .05 .44 .16 .21
German weaving industry .06 .58 .18 .18
German clothing industry .04 .48 .18 .13
Danish textile and clothing .05 .49 .12 .10
industry
Decrease
Production
.23 2.94 2.33 2.92
German spinning industry
German weaving industry .44 2.29 1.72 2.10
German clothing industry .65 2.38 1.88 2.10
.49 2.14 1.16 1.59
Danish textile and clothing
industry
Sales prices
.14 1.78 1.18 1.60
German spinning industry
.34 3.17 1.73 2.58
German weaving industry
.30 3.23 2.15 2.88
German clothing industry
.69 3.70 1.65 3.05
Danish textile and clothing
industry
Table 1.2 contains the information gain for Production and Sales prices
for four sets of firms: the German spinning, weaving and clothing industries
as well as all firms of the Danish textile and clothing industry. Not all 3 x 3 = 9
prediction-realization combinations in month t are considered separately,
because this would reduce several cell frequencies below acceptable levels.
For each Type i forecast in month /+1 the nine possibilities are combined
in three groups. One group deals with the information gain of a Type i
forecast under the condition that there was a correct Type i forecast in the
month before (indicated by “Correct same sign” in the table). For /= 1: the
information gain of an increase forecast for /+ 1 when an increase is reported
for t which was predicted one month earlier. The second group concerns the
information gain of a Type i forecast under the condition that there was a
correct Type j forecast, jf i, in the month before (“Correct other sign”).
This combines two separate cases for each i. For /= 1: the information gain
of an increase forecast for t+ 1, either when there was a correct no-change
forecast in t (j = 2) or when there was a correct decrease forecast in that
month (y=3). Finally, the third group deals with the information gain when
there was a prediction error last month (“Incorrect”). Six separate cases fall
under this category.
The results collected in Table 1.2 show that, without exception, the in¬
formation gain takes the smallest value when the forecaster confirms his pre¬
diction and states that the variable will behave in the same way in the next
month. It takes the largest value (also without exception) when the forecaster
confirms his prediction and states that something else will happen next
month. Even if some of the values are quite small, particularly for no-change
predictions, they are nonetheless all positive, which means that they all have
some merits with respect to the realizations that will be reported one month
later.
7.7. The Sampling Variability of the Information Value
The empirical illustrations of the information concept were all based on

observed relative frequencies. Such frequencies vary from sample to sample,
so that the associated information value will then also be subject to sampling
variability. The derivation of complete sampling distributions of information
measures is not one of the objectives of this book. To show what is involved
we shall consider here the asymptotic sampling variance in the simplest case.
Suppose that the relative frequency /is obtained from a random sample
of size n. Let the population proportion be p and write q for 1 -p. Then/is
subject to a binomial distribution with mean p and variance pqjn. The differ-
1.8 HISTORICAL NOTES 19
ential of the information value/,(/) = - log/is dh= - df/f if we use natural

logarithms (i.e., if we measure information in nits). Let us interpret this
differential as a sample deviation from the population value h(p). On squaring
both sides and taking the expectation we obtain
or
(7-1) var h(f)= —

pn
This result has only large sample validity, because we confined ourselves to
the first-order differential and hence neglected higher-order terms.
We conclude from (7.1) that, given the sample size n, the variance in¬
creases when p becomes smaller. This is not really surprising, because the
information h(p) is also larger when p takes smaller values. The sampling
variance is inversely proportional to the odds p/q of the event, i.e., to the
ratio of the probability that the event will occur to the probability that it
will not occur. [“The odds are 3 to 2,” which is equivalent to “the probability
of the event is .6.“] We shall meet the odds again as an important concept in
the analysis of conditional probabilities in Chapter 3 (Section 3.4).
1.8. Historical Notes
The origin of information theory is in statistical thermodynamics, where

the entropy concept was developed which will be considered briefly in Chap¬
ter 2 (Section 2.1). Information theory itself was developed mainly in com¬
munication engineering. An important forerunner was R. V. L. Hartley [1],
whose contributions date back to 1928. [Numbers in brackets refer to the
references at the end of this section.] The largest stimulus is due to C. E.
Shannon [2]; his work was published shortly after the Second World War.
Although most of the results are still in the field of communication theory
in the narrow sense, there are several applications in other areas ranging
from statistics to psychology [3, 4], There are some isolated attempts to apply
some of the concepts to problems in economics [5, 6] but these are mainly
confined to the presentation of analogies. The reason information theory is
nevertheless important in economics is that it is more than a theory dealing
with information concepts. It is actually a general partitioning theory in the
sense that it presents measures for the way in which some set is divided into
subsets - a point stressed particularly by De Jongh [7], This may amount to
dividing certainty (probability 1) into various possibilities none of which is
certain, but it may also be an allocation problem in economics. To make the

discussion more lively we shall use terms which are derived directly from
communication theory, such as messages sent and received and prior and
posterior probabilities. Nevertheless, it is appropriate to realize that the
area of application is much wider than this terminology suggests.
[1] Hartley, R. V. L. “Transmission of Information,” Bell System Techni¬

cal Journal, Vol. 7 (1928), pp. 535-563.
[2] Shannon, C. E. “A Mathematical Theory of Communication,” Bell Sys¬
tem Technical Journal, Yol. 27 (1948), pp. 379-423, 623-656. [This article
has been reprinted in C. E. Shannon and W. Weaver, The Mathematical
Theory of Communication. Urbana, Illinois: The University of Illinois
Press, 1949.]
[3] Kullback, S. Information Theory and Statistics. New York: John Wiley
and Sons, Inc., 1959.
[4] Attneave,F. Applications of Information Theory to Psychology. New
York: Henry Holt and Company, 1959.
[5] Lisman, J. H. C. “Econometrics and Thermodynamics: A Remark on
Davis’ Theory of Budgets,” Econometrica, Vol. 17 (1949), pp. 59-62.
[6] Pikler, A. “Optimum Allocation in Econometrics and Physics,” Welt-
wirtschaftliches Archiv, Yol. 66 (1951), pp. 97-132.
[7] Jongh, B. H. de. Egalisation, Disparity and Entropy. Utrecht: A. W.
Bruna en Zoons - Uitgevers Maatschappij, 1952.
APPENDIX 21
APPENDIX TO CHAPTER 1
The data used in the analysis of Section 1.6 are presented below in Tables
1.3 through 1.6. Nine squares are given for each variable. Each square
contains 9 numbers which are the absolute frequencies corresponding to /b.
For example, the first table considered in Section 1.6 corresponds to the
leading square of Table 1.5. It should be noted that the German forecasts
in the first four years were supposed to refer to the next two months, not to
next month only. It turned out that the forecasters’ ability with respect to
the second month was very meagre, so that the questionnaires were confined
to predictions one month ahead in later years. The computations described
here are all based on the assumption that the forecasts refer to next month.
TABLE 1.3
PREDICTION-REALIZATION TABLES FOR THE GERMAN SPINNING INDUSTRY
TABLE 1.4
PREDICTION-REALIZATION TABLES FOR THE GERMAN WEAVING INDUSTRY
appendix 23
TABLE 1.5
PREDICTION-REALIZATION TABLES FOR THE GERMAN CLOTHING INDUSTRY
TABLE 1.6
PREDICTION-REALIZATION TABLES FOR THE DANISH TEXTILE AND CLOTHING INDUSTRY
Last month’s realization

\
Production
a>
IS! 185 85 8 70 76 6 13 16 8
4>
75 99 19 55 120 23 5 10 5
3 10 8 1 11 4 1 1 2
4> 54 45 3 74 81 10 35 21 10
O C
96 152 30 137 1060 110 21 86 67
J=, 5
o 14 8 6 72 76 7 32 40
_o IS) 5 6 0 26 21 7 26 14 8
4> 73 23
•5 1— 2 12 5 14 16 42 41
4>
<D
X) 1 4 3 2 28 21 5 25 43
o.
oo
J2
Sales prices
C 4)
O w 76 19 0 20 29 0 0 1 0
£ 
27 29 0 66 71 4 2 2 0
O c
c rt 41 71 4 83 2507 82 5 80 49
o 2 0 0 0 28 28 2 14 19
<D
IS)
1 0 0 0 2 0 0 0 0
4)
1 0 1 0 32 12 0 35 20
X3
«
0 0 0 1 4 8 0 5 17
CHAPTER 2
EXPECTED INFORMATION
2.1. The Expected Information in a Determinative Message; The Entropy

Concept
We consider a set of n events, Eu ..., En. They form a complete system in
the sense that it is certain that exactly one of them will occur. Thus, if their
probabilities are
(1.1) x2 ... x„
they should add up to 1:

n
(1.2) £xi = 1 xi >0 i = l, ■■■’"

i= 1
When we receive a definite and reliable message which states that Et has
occurred, the information content of the message is h(xf) = — log x;. Before
the message is received we do not know, of course, how large this information
content will be, since it may be any of the numbers h(xi),..., h(xn). However,
we can compute the average or expected information content before the actual
message comes in. Since Et has probability xh the message that E; occurred
will be received with the same probability, so that the information content
will be h(x{) with probability x;. The expected information content is there¬
fore:
n n J n
(1.3) H(x) = £ x,h(xi) = £ X; log _ = - £ X; logX;

1=1 i=l X; 1=1
where the x on the left stands for the array of the n probabilities x1? ..., x„.
We conclude that the expected information content is nothing else than
minus the expectation of the logarithms of the probabilities.
It is obvious that H( ) cannot be negative, since it is a weighted average
with nonnegative weights of individual information values /;( ) which are
also nonnegative. When x; vanishes, the product of x; and log x; is of the
form 0 x ( — oo), which is not defined. It is customary, however, to define
(1.4) XilogXj—O if x,. = 0
24
2.1 A DETERMINATIVE MESSAGE 25
which is in accordance with the limit of x; log Xl for x; approaching zero.

This is easily checked - the formal proof is simple too - when we take loga¬
rithms to the base 10 and write Xi = 10~\ Then the left-hand side of (1.4) be¬
comes k x 10~,,: apart from sign, which is
.1 for k = 1 (x—.l)
.02 for k = 2 (x, = .01)
.003 for ^ = 3 (x—.OOl)
Given that negative values of H( ) are impossible, we conclude that its

minimum value is zero and that it is attained when xt = 1 for some i, Xj = 0 for
ah jî. In words: No information is to be expected from a message on the
realization of one of the n events when it is known before that some Et has
probability 1. This is a natural result. To find the largest value which H( )
can take we maximize this function subject to Ix~\. So we consider the
Lagrangian expression
where X is the Lagrangian multiplier. We differentiate with respect to x; and

put the result equal to zero:
— 1 — log X; — X = 0 i = 1,..., n
where log is interpreted here as natural logarithm. We conclude that

log x. = —1—X holds for each /', so that the probabilities should be all the
same and hence equal to 1/n. In words: The expected information takes its
maximum value when all events have the same chance. This result is natural
too, for when all possibilities are equally likely, the message which states what
actually happened must be expected to contain more information than in any
situation in which we know before that some possibilities are more probable
than others. Or equivalently, there is a maximum of uncertainty when all n
probabilities are equal to 1/n; and the more uncertainty there is now, the
more information is to be expected when we hear what actually happened.
Indeed, uncertainty and expected information can be regarded as dual con¬
cepts. Uncertainty prevails prior to the message, information is supplied
when the message comes in. When the probabilities are (.99, .01) there is little
uncertainty; when they are (.6, .4) there is a considerable amount of un¬
certainty; the information to be expected from a reliable message is small in
the former case, large in the latter.
26 EXPECTED INFORMATION
The maximum value of H( ) is log n as is easily verified when we substitute

Xi=\/n into (1.3). We have thus established:
(1.5) 0 < H (x) < log n
The maximum thus increases with n, which is in accordance with the in¬
creasing uncertainty that results from a larger number of possibilities. For
the special case n = 2, xx=x, x2 = l-x the behavior of the expected in¬
formation is illustrated in Figure 2.1.
l/l
-*->
C
C
c
o
o
E
o
c
X)
<D
-t—’
u
(D
Cl
X
LlJ
Fig. 2.1. The expected information H( ) for n = 2
The expected information of a distribution is frequently called the entropy

of that distribution. As stated in Section 1.8, this term is of physical origin.
It is a measure of‘‘disorder.” The closer the n probabilities are to \\n, and
the larger n is, the less order there is in the system. The second law of thermo¬
dynamics states that there is an inherent tendency of the entropy to increase.
There is no need to dwell on physical analogies here; we shall merely use
“entropy” as a synonym for the longer expression “expected information.”
2.2 AN INDIRECT MESSAGE 27
2.2. The Expected Information of an Indirect Message
The expectation concept of the previous section is based on the elementary

information definition h(^x)— — log x of Section 1.2. We generalized that
definition in Section 1.4 by taking account of the possibility that the message
does not really guarantee that a particular event occurred. We shall now make
the corresponding generalization for the expected information.
Again, we assume that there are n events
E1 E2 ... En
and that their probabilities are
(2-1) Xl x2 ... xn
These will now be called prior probabilities for the following reason. A
message comes in. It does not state that one of the events took place. Rather,
it implies that the odds are changed such that some events have become more
probable and some less probable. That is, the message transforms the prior
probabilities (2.1) into the following posterior probabilities:
(2-2) ki y2 ••• yn
If one of the /s is 1 and all others 0, we have the situation described in the
previous section and the message is direct. This is now a special case of the
present one, since the /s may take any values subject to
n
(2.3) Efi=l yt> 0 i — 1,..., n

i— 1
Therefore, the general case (with an arbitrary number of positive y’s) is called
that of an indirect message.
The expected information of such a message can be defined straight¬
forwardly if we do so on the basis of the definition (4.7)-(4.8) of Section 1.4.
Suppose for a moment that the message confines itself to one event Eh for
which the prior probability is x{ and the posterior probability yt. The infor¬
mation received with the message is then equal to the logarithm of yjxt.
However, the indirect message does not confine itself to one Ep, it states
that each Et has its own (posterior) probability yh i= 1, ..., n. Taking the
expectation of the separate information values, we find that
(2.4) i(y-x) = Z k; log -■

1=1 A;
is the expected information of the indirect message which transforms the

prior probabilities x1, ..., x„ into the posterior probabilities jq, yn.
This expectation is nonnegative, which can be shown as follows. We assume
in the first instance ^>0, z'=l, and define el5 as follows:
n
(2.5) X; = yt( 1 + £;) which implies £ ytet = 0

i= 1
Then I (y:x) can be written as
I(y:x) = -fjyi log - = - £ yt log(l + e,)

i— 1 ki ‘=1
= Z [ef — log (1- + 0]

i= 1
where the last equality sign is based on Syisi = 0, see (2.5). To prove
I(y:x)^0 we shall show that the expression in square brackets,
Si - log(l + £;)
is nonnegative. We note, first of all, that it vanishes for £; = 0. Assuming that

the log is a natural logarithm,1 we find for its derivative with respect to £;:
1 £;
1-=->0 if £; > 0
1 + £j 1 + £j
<0 if £; < 0
Hence this function vanishes for £; = 0, it has a positive derivative and thus
takes a positive value for £;>0, and it has a negative derivative and thus
takes a positive value for ^<0. This shows that I(y:x)^0 and also that
I(y\x) = 0 if and only if all s’s vanish, i.e., if and only if Xi—yi for each i.
When one or more y’s is zero, we can apply a limiting process to obtain the
same result.
We thus have I (y:x) = 0 if and only if xt=yt holds for every value of the
index. That is, the expected information content of an indirect message
vanishes when it leaves all probabilities unchanged - which is a natural
result. Note that I(y:x) may be infinitely large, viz., when >»i>xi = 0 for
some i. The prior probabilities then specify that Et has zero chance, but this
is raised to a positive value by the message. Hence the probability is increased
1 When any other logarithmic base is used, I(y:x) is multiplied by a positive constant
which does not affect its sign.
by a factor “infinity, and we are also “infinitely surprised” by the message.

Note further that in the special case of complete initial ignorance, x—l/n,
i = l, ...,«, the expected information I(y:x) becomes
" 1
(2.6) log n £ yt log = log n - H (y)
;= i
In words: The expected information of the indirect message is then equal to

the difference between the maximum value log n of the entropy of the
posterior distribution [see (1.5)] and the actual value H (y).
For the special case n = 2, =x, x2 = 1 —x, yt —y, y2 = l—y contour lines
of equal expected information are drawn in Figure 2.2.
2.3. The Information Content of Survey Forecasts
We shall now apply the indirect message idea to the survey forecasts that
were considered in Section 1.5. That is, we shall interpret an increase forecast
(similarly for no-change and for decrease) as a message which transforms
Fig. 2.2. Contour lines of constant expected information (in bits) of the posterior message
(y, 1 —y), given the prior probabilities (x, 1 — x)
prior probabilities into posterior probabilities. We reproduce the general

form of the prediction-realization table for the sake of convenience:
Prediction: increase /n fl2 fx 3 fx.
no change fix fl 2 f23 f2.
decrease fix f22 f3 3 A
Marginal fx f.2 f.3 1
We assume that before any message comes in nothing is known about the
changes that will be realized except the marginal frequencies 3-
These will be regarded as the prior probabilities xl5 x2, x3. Suppose now that
an increase is predicted. This message changes the prior probabilities to
yi=flj/f1 ,j— 1, 2, 3. We apply (2.4) and find that
j= 1 j= 1
is the expected information of the increase forecast when this prediction is
considered as an indirect message. We shall call it the information content of
an increase forecast. Using the numerical prediction-realization table of
Section 1.5 for Purchase prices, we find that it takes the following value:
.Ul_ .111 .043 .043 .003 .003

log +-.log 1.156 bit
T57l0g (.157)(.154)4 .157 (. 157) (.635) .157 (,157)(.211)
The results for all three prediction types (increase, no change, decrease)
and for all ten variables of Section 1.5 are presented in the first three columns
of Table 2.1. The figures indicate that the information content of the no¬
change forecasts is in all cases very much less than that of the increase and
decrease forecasts. This is entirely in accordance with our findings on the
information gain in Section 1.5.
It is, of course, necessary to pay explicit attention to the conceptual
difference between the information gain and the information content of a
forecast. Let us consider a Type i prediction (z = l, increase; 2, no change;
3, decrease). Its information gain is
fa
(3.1) log
fj.i
TABLE 2.1
THE INFORMATION CONTENT (IN BITS) OF SURVEY FORECASTS OF TEN VARIABLES
Expectation
Variable Increase No change Decrease
(4.8)
Purchase prices 1.156 .105 .608 .367

Sales prices .632 .052 .712 .237
Work force .648 .030 .799 .187
Production .401 .031 .747 .169
Investment in machinery .460 .066 .212 .150
Rate of profit .470 .040 .315 .152
Sales .196 .056 .244 .163
Raw materials bought .236 .023 .363 .118
Orders received .215 .017 .298 .103
Inventories .206 .017 .173 .067
and its information content is

3
(3.2) A.
fJ.j
We can regard both as weighted averages of the n logarithmic ratios
(3.3)
u log
fin
fj.
(n — 3 in our case), but the weights are different. The information gain has a
unit weight for the ith ratio and thus confines itself completely to the correct-
forecast fraction fu. The information content uses the weights fj/f%,
7 = 1, •••> n.
The difference in weighting schemes has important implications. When
forecasts and realizations are stochastically independent, i.e., when fj =
f. f holds for all pairs (/', j), the information content of the forecast is
zero. This is the minimum, because the information content cannot be
negative — see the discussion below equation (2.4). Therefore, from the stand¬
point of that measure the worst possible situation is stochastic independence
of forecasts and realizations. That situation is surely also bad from the stand¬
point of the information gain, which is then zero as well, but the important
point is that the information gain may be negative, which is still worse. There
is no finite lower bound to the gain, because we may have/-.,/;>/;; = 0. That

worst possible situation is not necessarily bad from the viewpoint of the in¬
formation content, which then takes a positive value because forecasts and
realizations are not stochastically independent in that case.
This apparent contradiction is simply due to a difference in interpretation.
The information gain takes the forecast at its face value: If an increase was
predicted, one expects that there will be an increase, and if this is not the case
this is counted as an error. The information content argues that there is
always something that can be learned from a forecast even if it is misleading.
Consider the following prediction-realization table (with no-change deleted
for simplicity):
increase decrease
Prediction: increase .2 .4 .6
decrease .3 .1 .4
Marginal .5 .5 1
The information gain of an increase forecast is log §, which is negative. The

same sign applies to the information gain of a decrease forecast. The in¬
formation content of an increase forecast is
i logf + f logf = .082 bit
which is positive. Why? It is because the information content argues that

when an increase has been predicted, it is better to proceed under the
assumption that there will be a decrease, and vice versa. This amounts to
revising the forecasts, which leads to the following prediction-realization
table:
increase decrease
Revised prediction:
increase .3 .1 .4
decrease .2 .4 .6
Marginal .5 .5 1
The information content of the revised increase prediction is equal to the in¬
formation content of the original decrease prediction. The same holds for the
2.4 BIVARIATE INFORMATION THEORY 33
revised decrease prediction, mutatis mutandis. Hence we interchange two

numbers as far as the information content is concerned; that is all. The
changes in the information gains are more drastic. For increase the new value
is decrease log y, both of which are positive. Hence both information
measures (content and gain) ascribe positive merits to the forecasts when
these are interpreted in the revised manner.
2.4. Bivariate Information Theory
When discussing stochastic independence of forecasts and realizations we

implicitly introduced certain elements of bivariate information theory. Note
that we started in Section 2.2 in a completely wm'variate manner: We com¬
pared two sets of probabilities, xlt ..., xn and yt, ..., yn, each of which
corresponds to one univariate distribution. Now that we did consider
bivariate aspects, however, it is better to do this more explicitly.
Our starting point is a set of m + n messages. The first m, to be indicated by
Xt, ..., Xm, are the messages sent, and the last n, indicated by Yu ..., Yn, are
the messages received. For example, we may have m — n = 3 and the following
interpretation of the six messages:
X1 = increase predicted Yt = increase realization reported

X2 = no change predicted Y2 = no-change realization reported
X3 = decrease predicted Y3 — decrease realization reported
This corresponds to the case of the previous section. It is important to

realize, however, that there is not necessarily a pairwise correspondence
between the messages sent and the messages received. It is conceivable, for
example, that we have m = 4, n = 3, and the following “prediction” inter¬
pretation of the fourth message sent X4: “No idea about what is going to
happen next month.”
Let us write ptj for the probability that Xt is the message sent and Y} the
message received, where i = l, ..., m and j= 1, ..., n. We write
Pi. = I Pit i = 1, ...,m

j= 1
m
p.j = I Pa j = l,...,n
i= 1
for the probability that Xt is sent and that Yj is received, respectively. These
are the probabilities of the two one-dimensional marginal distributions,
which should be compared with and yt of the previous section. We thus

have several distributions and hence also several information expectations.
One is the expected information or entropy of the bivariate distribution:
m n
(4.2) H(X,Y)= £ £ ft, log-

z = i i= i Pij
where X, Y in parentheses stands for the set of all m + n messages sent and
received. Two other information expectations refer to the marginal distri¬
butions:
H(X)= £ pL log -
t=l Pi.
(4.3)
H(Y)= £ p.j log —
j= i P.j
We thus have a joint entropy H (X, Y), a message sent entropy H (X), and a
message received entropy H (X).
An obvious question is now: What is the relationship between these one-
and two-dimensional entropies? To answer this question we introduce
(4.4) hu = log -PlJ i = 1,..., m ; j = l,...,n

Pi.Pj
which is called the mutual information between the message sent X,- and the
message received Yj. It vanishes when there is stochastic independence
between the two messages; it is positive when Yj is more frequently the
message received if Xt is sent than the independence pattern implies, negative
in the opposite case. [The information gain of a forecast, (3.1), corresponds
to hu (the special case of equal subscripts) when the messages are interpreted
appropriately.] Next we consider the average of all mn values /iy with the p
as weights:
(4-5) IXY = X X PiJ log ^

i= i j= i Pi.Pj
which is the expected mutual information. It is nonnegative, because it can be

regarded as the expected information of an indirect message which deals with
mn events. The prior probabilities are of the form pimPj, the corresponding
posterior probabilities have the form pti. We have IXy — 0 if and only if the
set of m messages sent is independent of the set of n messages received.
2.5 THE INFORMATION INACCURACY 35
We can write the right-hand side of (4.5) as follows:

m n
Pu
E Z Pu ]°g
»'=1 J=1 Pi.P.j
m n | m n | m n i
= Z Z
i=ij=l
P„ log -- +
Pi. i=lj=l
Z Z
Pu log - -
P.j i=lj=l
Pu log - z Z
m 1 " J m n ^
= I Pi. log + I p.j log - V £ p.. log -----

i=1 Pi- J=1 P-j i=U= 1 Pij
Hence, in view of the definitions (4.2) and (4.3):
(4.6) IXY — H(X) + H(Y)-H(X, Y)
We conclude that the expected mutual information is equal to the excess of

the sum of the two one-dimensional marginal entropies over the joint entropy.
The nonnegativity of IXY implies that the joint entropy is at most equal to the
sum of the marginal entropies:
(4.7) H(X,Y)^H(X) + H(Y)
We have the equality sign in (4.7) if and only if there is stochastic independence
between the messages sent and the messages received.
There are several important extensions of this analysis, but these will be
postponed until Chapter 3 (Sections 3.1 and 3.3). Here we confine ourselves
to an application to the forecast evaluation of the previous section. Let us
take a weighted average of the three information contents (3.2) with weights
equal to the frequencies of the corresponding forecasts:
3 3 3
(4.8)
y=i i = 1 j= l
This is the expected information content of the set of all three types of fore¬
cast; its numerical values for our ten variables are given in the last column of
Table 2.1. Clearly, the expected information content is nothing else than the
expected mutual information between the forecasts (messages sent) and the
realizations (messages received).
2.5. The Information Inaccuracy of Decomposition Forecasts
We return to the area of forecast evaluation, but the things that are
predicted will have a different nature. The forecasts considered until now
were of the following type:
“There will be less than 30 per cent sunshine on March 30, 1967.”
“The prices of ray raw materials will be higher next month than they are
now.”
We measured the frequency of such forecasts, the frequency of the realized
outcomes, the frequency of correct forecasts and of various types of error,
and so on. In other words, the object of prediction need not necessarily have
a quantitative (numerical) form; the only numbers used in the analysis are
the frequencies or probabilities just mentioned.
The scene will now change to the extent that the object of prediction is
quantitative and that probabilities in the usual sense of the word do not play
a role. The reason some of the concepts introduced in the previous pages
remain relevant is that we predict several quantities simultaneously, each of
which is nonnegative and the sum of which is 1. Hence we can say, formally at
least, that we predict probabilities. An example is presented in Table 2.2. It is
concerned with the Construction Programme of the Netherlands in the years
1950-1951. This Programme is based on a system of licenses. Plans are drawn
up every year for the issue of licenses next year, which implies a set of
predictions. When the year has come to an end we know how the available
resources have actually been allocated to the alternative ends, so that we can
compare forecasts and realizations.1
TABLE 2.2
PREDICTED AND REALIZED ALLOCATIONS ACCORDING TO THE
DUTCH CONSTRUCTION PROGRAMME, 1950-1951
Realized Predicted Realized

Allocation categories in 1950 for 1951 in 1951
Residential construction .531 .538 .497

Factories .169 .177 .198
Commercial and traffic premises .084 .080 .079
Farms and barns .063 .061 .067
Schools .034 .040 .041
Hospitals, churches, etc. .035 .045 .046
Repairs and rebuilding .083 .060 .072
Total 1 1 1
1 For details see Applied Economic Forecasting, Chapter 9.

2.5 THE INFORMATION INACCURACY 37
More generally, let there be n allocation proportions to be predicted, and

let us write for the predictions:
(5-l) *1 *2 ••• *„
and for the corresponding realizations as observed afterwards:
(5-2) >’i y2 ■■■ yn
Both sets of values add up to 1 and are nonnegative:

n n
(5-3) !*<= E» = l o i = 1,..., n

i= l i = 1
Therefore, both the x’s and the y’s satisfy the conditions which are necessary
and sufficient in order that they can qualify as the probabilities of some
distribution.
The question to be considered is: Does information theory supply us with
a “natural” measure for the accuracy of the set of n forecasts (5.1)7 It does,
and the argument is as follows. First, we note that the forecasts (5.1) are
available before the realizations are available, so that we can reasonably
argue that xu xn are prior probabilities. They are probabilities in this
sense: If we pick out at random a guilder of the Programme prediction for
1951 (see Table 2.2), how large is the chance that this guilder falls under
Residential construction? [Answer: xq^.538.] Next, the message comes in
which states how the a’location actually took place, and this message trans¬
forms the prior probabilities xt, ..., xn into the posterior probabilities
y..., yn. [Note that the realizations form the message now, not the forecasts
as in Section 2.3!] When this message has zero expected information, we
have Xi=yh i= 1, ..., n, and hence perfect forecasts. When its expected in¬
formation is very small, the forecasts are accurate although not perfect.
When the expected information is large, so that at least some of the x’s differ
substantially from the corresponding /s. the forecasts as a whole are very
inaccurate. Therefore, we shall call
(5.4) I(y:x) = £ yt log -

1=1 A;
the information inaccuracy of the forecasts x1, ..., xn with respect to the
realizationsyu ..., y„. It measures the expected information of the indirect
message yu ..., yn, given the prior probabilities xt, ..., xn. It also measures
the extent to which the decomposition forecasts xt, ..., xn differ from the
corresponding realizations ylf ..., yn.
The application of the information inaccuracy concept to the 1951 data of

Table 2.2 is straightforward. We predict the third column by the second,
which leads to the following inaccuracy value:
.497 .198 .072

(5.5) .497 log-+ .198 log + ... + .072 log —- = .0063 bit
v .538 .177 .060
Quite obviously, the number of bits will vary from year to year. Since the
information concept is essentially additive, a natural measure for the in¬
accuracy of a number of decomposition forecasts is the average information
inaccuracy.
T
where yit is one of the n realizations of the tth observation, xit the corre¬
sponding forecast, and T the number of observations. If we apply this
measure to the seven-category allocation forecasts of the Dutch Construction
Programme in the eleven-year period 1951-1961, we obtain .0142 bit. To
judge whether this is large or small, we may want to compare it with the
information inaccuracy of no-change extrapolation forecasts. This amounts
to predicting the third column of Table 2.2 by the first; in terms of /, to a
replacement of xit in the right-hand side of (5.6) by yiyt-^. The resulting
average information inaccuracy in the period 1951-1961 is .0096 bit, which is
smaller. Our appraisal of the allocation forecasts of the Construction
Programme must therefore be rather negative.
The information inaccuracy (5.4) is the sum of n terms, some of which are
positive and some negative. They cannot be all positive, because that would
imply yi>xi for each i, which is contradicted by Zy~Zx{ = 1. Given this
element of cancellation, one may prefer an expression which contains each of
the prediction errors xt—y( explicitly. This is easily accomplished by ex¬
panding the logarithms according to powers of these errors. If we work with
natural logarithms, we have
log ^ = - log I' = - log ^1 + X‘ ^
-+++)'++)■++)'
The expansion converges if (x;->>,■)/+• is less than 1, i.e., if X;<2yt. On multi-
2.6 INPUT STRUCTURE PREDICTIONS 39
plying both sides ot the equation by yt and summing over i in accordance

with (5.4), we find that the linear term vanishes:
- ^ (x, - y,) = 0
i= 1 <=i
So we obtain
n
(■*.• - yd2 i V (*, - y_l l V (*, - y,T

i= 1
yt 3 L
i= 1
y! 4!_,
i=l
yf
The leading quadratic term has the form of a chi-square. We conclude that
when the errors xi—yi are close to zero, the information inaccuracy criterion
is approximately equivalent to the chi-square criterion. The proportionality
constant \ applies when information is measured in nits. If we prefer bits, it
should be replaced by .721. The chi-square corresponding to the 1951 pre¬
diction of Table 2.2 is .0082. If we multiply this by .721, we obtain .0059,
which is close to the value of (5.5).
2.6. The Information Inaccuracy of Input Structure Predictions and the

Information Improvement of Forecast Revisions
We turn from Construction Programmes to input-output analysis. This is

a subject in which we shall have to go fairly deeply in Chapter 9; for our
present purposes, however, the example contained in Table 2.3 is sufficiently
illustrative. It deals with the inputs bought by the sector called Agriculture,
forestry and fishing in the Netherlands in each of the three years 1948-1950.
The first column deals with the situation in 1948 and should be interpreted as
follows. Out of all inputs bought by firms of this sector 18.6 per cent was
supplied by other firms of the same sector. This includes, for example, the
seed bought by one farmer from another. Furthermore, 8.0 per cent was
supplied by the food manufacturing sector (non-animal products), 3.5 per
cent by the chemical sector, 1.5 per cent consists of services rendered by
wholesale traders, and 9.1 per cent is the share of all other sectors which are,
as far as their supplies to Agriculture, forestry and fishing are concerned, of
minor importance compared with the four listed above. The country’s
domestic enterprise system as a whole has been divided into 35 sectors, hence
this remainder group consists of 31 sectors.
The total input bought by Agriculture, forestry and fishing consists not
only of purchases from the 35 sectors, but also of so-called primary inputs.
TABLE 2.3
INPUT STRUCTURE OF AGRICULTURE, FORESTRY AND FISHING IN THE NETHERLANDS,
1948-1950
Input structure in year

Inputs supplied by
1948 1949 1950
Sector inputs
1. Agriculture, forestry and fishing .186 .175 .167

2. Food manufacturing industries* .080 .089 .098
3. Chemicals and petroleum refineries .035 .037 .033
4. Wholesale trade .015 .015 .015
5. All other sectors .091 .094 .083
Primary inputs
1. Imports .047 .042 .041

2. Wages .142 .126 .110
3. Gross profits .404 .422 .453
Total input
1 1 1
* Non-animal products.
These are, by definition, the flows of goods and services supplied by economic
agents outside the domestic enterprise system. One example is imports
(4.7 per cent in 1948). Another is wages, which are the amounts paid for
labor services supplied by hired workers (14.2 per cent). There is also
depreciation on fixed assets (services rendered by capital equipment), interest
on loans, net profits (payments for “entrepreneurial services”), etc. The input-
output system is designed such that total input equals total output for each
sector, which implies the necessity to be exhaustive with respect to the various
inputs and to include such items as depreciation and net profits. The re¬
mainder term of the primary inputs has been labeled “gross profits” and
accounts for 40.4 per cent of total input in 1948.1
An important issue in input-output analysis is to what extent the input
structure of the various sectors is stable over time. We shall therefore con¬
sider the following problem. Suppose that we wish to predict the input
structure of Agriculture, forestry and fishing in 1949 (the second column of
1 For further details see Applied Economic Forecasting, Chapter 8, from which the data
discussed in this and the next section have been taken.
2.6 INPUT STRUCTURE PREDICTIONS 41
Table 2.3) by means of the input structure in 1948 (the first column). How
accurate is that forecast? The answer in terms of the information inaccuracy
is straightforward:
iacl -175 .422

•175 Iog .186
i 07 + ••• + -422 l°g .404
7T = -00365 bit
Suppose next that 1950 is to be predicted on the basis of the 1948 data. The
information inaccuracy is then
-167 .453
.167 log —- + ... + .453 log-= .01519 bit
.186 .404
which is larger as could be expected, because the change in the input structure
will on the average be larger when we consider two years which are not
consecutive. This idea is corroborated by Table 2.4, which gives the complete
information tableau of all input structure forecasts for the agricultural sector
in the years 1948 through 1957. The first row contains the information in¬
accuracy values which are obtained when the 1948 input structure is used to
predict the structure of 1949, 1950, ..., 1957. The second row uses the 1949
data, which are used to predict all later years 1950,1951, ..., 1957; and so on.
It is clear that the figures increase systematically when we move from left to
right in each row.
TABLE 2.4
INFORMATION TABLEAU OF THE INPUT STRUCTURE OF AGRICULTURE, FORESTRY AND FISHING,
1948-1957
Base
Year to be predicted t + r
year
1949 1950 1951 1952 1953 1954 1955 1956 1957
(0
1948 365 1519 2304 3643 3523 4603 5862 7109 8391
1949 547 1269 2092 2113 2946 3899 4957 5914
1950 765 846 1293 1677 2572 3562 4205
1951 547 845 1382 1643 2885 3833
1952 369 431 795 1687 2197
1953 131 359 800 1404
1954 256 550 945
1955 252 599
1956 175
Note. All information values are expressed in 10 5 bit.

TABLE 2.5
AVERAGE INFORMATION INACCURACY FOR DIFFERENT TIME SPANS OF PREDICTION
Agriculture, forestry Median over

Time span t
and fishing 15 sectors
1 year 379 900

2 years 802 1606
3 years 1373 1701
4 years 2028 2143
5-6 years 3336 3164
7-10 years 6073 5451
Note. All information values are expressed in 10 5 bit.
The number of information inaccuracy values obtained in this way is

rather large and it becomes even larger when we carry out the same compu¬
tations for the input structures of other sectors besides Agriculture, forestry
and fishing. Therefore, some averaging procedure comparable to (5.6) is
appropriate. For this purpose we notice that the diagonal figures of the table
(365, 547, ..., 175, all in 10”5 bit) are all inaccuracy values corresponding to
forecasts one year ahead. It is not unreasonable to regard this set of numbers
as fairly homogeneous, so that we may decide to characterize this set of values
by the average inaccuracy. We shall proceed in the same way for the forecasts
two years ahead (with inaccuracy values 1519, ..., 599, in 10-5 bit), and so
on. The results are shown in Table 2.5, which contains both the average in¬
accuracy values of the agricultural sector and the medians of these averages
over fifteen sectors, the input structures of which are analyzed in the same
way as that of the agricultural sector.1 The agricultural values differ from the
median outcomes, but both columns agree as regards the increase with an
increasing time span t. [For large x values some aggregation is applied,
because the number of observations would otherwise be rather small; the
indication “5-6 years” implies that the average inaccuracy value is taken for
all cases for which t = 5 or 6, and similarly for 7-10 years.]
1 The other 14 sectors are: Food manufacturing (animal products), Food manufacturing
(all other products), Beverages and tobacco, Footwear and other wearing apparel, Wood
and furniture, Leather and rubber products (excluding footwear), Chemicals and petroleum
refineries, Basic metal industries, Metal products and machinery, Electrical machinery and
apparatus, Transport equipment, Construction, Electricity, gas and waterworks, Transport
other than ocean and air transport.
2.7 AGGREGATION THEORY OF THE INFORMATION INACCURACY 43
It is interesting to consider this result in the light of prediction revisions.

Let our goal be to predict the input structure of some sector in 1956. Suppose
that 1949 is the most recent year for which input data are available, so that
2 years. Suppose that one year later we have 1950 data and that we are
still interested in 1956. Then we are in a position to make a prediction
revision. The original information inaccuracy of the input structure forecast is
(6.1) I(y'x) = £ Ti log —

1 Af
where xu x2, ... represents the 1949 input structure and yu y2, ... that of
1956 (the year to be predicted). One year later we have the 1950 input
structure at our disposal. This is a revised forecast of the y’s, to be indi¬
cated by x[, x2, .... Their information inaccuracy is
(6.2) I(y:x’) = Yjyi \og-,

i Xt
We should obviously expect that the new inaccuracy is less than its prede¬
cessor, at least on the average, for otherwise the revision makes matters
worse rather than better. So we subtract the new value from the old one:
(6.3) I(y-x)-I(y:x') = '£iyilog-

i A;
This is the information improvement of the forecast revision. It is positive

when there is indeed an improvement in the sense that the information in¬
accuracy is reduced, zero when the inaccuracy remains unchanged (which
does not necessarily imply x~x\ for each /), and negative when the in¬
accuracy is larger than it was before the revision. The conceptual similarity of
the information improvement and the information gain defined in equation
(4.4) of Section 1.4 will be obvious. Positive values should be normal for
both, at least on the average, but negative values cannot be excluded. One
can derive average improvements from Table 2.5 by taking the successive
differences of the elements of each column.
2.7. The Aggregation Theory of the Information Inaccuracy
The input coefficients can be divided naturally into two sets: those of
sector input and those of primary input. The question that will be considered
in this section is: If we have certain decomposition forecasts (5.1) and the
corresponding realizations (5.2), and if the n categories are divided into a
number of sets, can we then use the information inaccuracy to measure the
within-set and between-set inaccuracies in any natural manner? In the input

case: Can we handle the sector inputs separately, and the primary inputs
separately, and also the dichotomy between total sector input and total
primary input? The answer is in the affirmative. As a matter of fact, the in¬
formation measures have very beautiful aggregation and disaggregation
properties as will become clear in this section and also in several other places
in this book.
Suppose that the n categories are divided into G sets, Su ..., SG, such that
each category belongs to exactly one Sg, g — 1, ..., G. Let us write Xg and Yg
for the forecast and the realization, respectively, of the gth set:
(7.1) xg= Yj Xi Yg= YJyi g = i,...,G

ie Sg ie Sg
Also, we write £;for the conditional forecast of the fraction of the ith category,
given that we are in the set that contains this category, and similarly rjt for the
corresponding conditional realization:
(7-2) &= Vi = ieSg,g = 1,...,G

^9 *9
These concepts are illustrated in Table 2.6 for the input structures of the
agricultural sector given in Table 2.3, where it is to be understood that the
set S1 consists of sector inputs and S2 of primary inputs. The symbols before
the numerical columns refer to the concepts introduced in (7.1) and (7.2).
For example, the top figure .410 in the 1949 column is Y1 when the 1949 input
structure is predicted by the 1948 data, and it is X1 when the 1949 input
structure is used to predict the structure of 1950.
Let us write the information inaccuracy (5.4) as follows:
G
(7.3) ((y;X)= V V^logtl

g=l ieSg
Clearly, the expression in brackets can be regarded as the part of the in¬
accuracy that corresponds to Sg. Let us consider that part in more detail:
VtlY,
log + log —
ieS
9 L xtl x„ X9-
, Vi , v . Yg
Vi log - + Yg log -9
2.7 AGGREGATION THEORY OF THE INFORMATION INACCURACY 45
TABLE 2.6
TOTAL SECTOR AND PRIMARY INPUT RATIOS AND CONDITIONAL INPUT RATIOS
OF THE AGRICULTURAL SECTOR
Input structure
Inputs supplied by Symbol in year
1948 1949 1950
Total sector and primary input

1. Total sector input xu yx .407 .410 .396
2. Total primary input +2, T2 .593 .590 .604
Individual sector inputs
1. Agriculture, forestry and fishing £1, t]i .457 .427 .421

2. Food manufacturing industries* £2, n2 .196 .216 .247
3. Chemicals and petroleum refineries £3, >73 .086 .090 .082
4. Wholesale trade £4,174 .037 .037 .039
5. All other sectors £5,175 .224 .230 .211
Individual primary inputs
6. Imports £e, 17 6 .080 .070 .067

7. Wages £7, 777 .238 .214 .182
8. Gross profits £s, ?78 .682 .716 .751
* Non-animal products.
On combining this with (7.3) we obtain
(7.4) I(y:x) = I0(y:x)+ £ YgIg(y:x)

9= 1
where
5
(7.5) Io(r-x)= E Yglog
xn
9=1 J *9
is the between-set information inaccuracy and

7;
(7.6) 4(.V:x) = E 1i '°g Z 0 = 1.G
ieSg Si
is the information inaccuracy within Sg.

We have thus succeeded in decomposing the total information inaccuracy
into (7+1 terms. The first is the between-set inaccuracy I0(y:x), which
vanishes when Xg= Yg, g = 1, ..., G. In the input case there is a zero between-
set inaccuracy when the total sector input proportion and the total primary
input proportion are predicted perfectly. All other terms are information in-
accuracies within separate sets, each of them weighted by the realized set
proportion Yg. For sector input prediction (g = 1) the term vanishes when the
forecast fractions xu ..., x5 are proportional to the corresponding realization
fractions yt, ..., y5. The weighted sum of the separate within-set inaccuracies
may be called the total within-set information inaccuracy.
It is a matter of straightforward arithmetic to apply this decomposition to
the data of Table 2.6. Take the 1948 input structure as the prediction tool and
the 1949 input structure as the object of prediction. The between-set in¬
accuracy is
.410 .590
.410 log-+ .590 log — = 3 x 10 5 bit
6 .407 .593
which is very small as could be expected, given that Xg and Yg, g = 1, 2, are so
close to each other. The sector input inaccuracy is
427 230
.427 log - — + ... + .230 log ' - = 321 x 10“5 bit
.457 .224
and, in the same way, the primary input inaccuracy is 390 x 10"5 bit. The
decomposition (7.4) takes the form
365 = 3 + (.410) (321)+ (,590)(390)
where the left-hand figure is the total information inaccuracy expressed in

10~5 bit.
In this example the between-set information inaccuracy accounts for a
minor proportion of the total inaccuracy. It appears that this is generally true
for the input structures of most of the fifteen sectors that were analyzed in
this way, the proportion being about one-sixth on the average. The contri¬
butions of sector input and primary input to the total within-set inaccuracy
vary widely; neither dominates the other on the average. For details we refer
to Chapter 8 of Applied Economic Forecasting.
2.8. Summary of Concepts
We conclude with a summary of the main concepts that were introduced in

Chapters 1 and 2:
(1) Let x be the probability of some event E. Then the information content
of the message which states definitely and reliably that E took place is
2.8 SUMMARY OF CONCEPTS 47
where the base of the logarithm is either 2 (in which case the information unit
is a bit) or e (in which case the unit is a nit).
(2) Let A' be the probability of E prior to the message, and y the prob¬
ability after the message. Then the information received with the message is
(8.2) h (a) - h (y) = log - 0 ^ x, y < 1

x
which includes (8.1) as a special case for j=l. In particular, if the message
has the nature of a forecast and if x stands for the probability of the predicted
event E prior to the forecast, while y is the probability of E given that it is
predicted to occur, then (8.2) is the information gain of this forecast. The
gain may be positive, zero, or negative. It is positive when the chance of E is
raised by the prediction, negative when E occurs less frequently when it is
predicted than when it is not predicted.
(3) Consider an arbitrary discrete distribution whose probabilities are

xl5 ..., x„. Then
(8.3) x, log -
i=l xt
is the entropy or expected information of that distribution. It is the expected

value of information values of the form (8.1), and it can be regarded as a
measure of disorder and also as a measure of uncertainty. The more possi¬
bilities there are (/? large) and the more they approach the situation of being
equally likely (x,- moving towards 1/n, i = l, ..., n), the more disorder there is
in the system. Also, when n becomes larger and the x; approach 1/n, the more
uncertain we are about the question which of the n events will actually take
place.
(4) Let Xj, ..., x„ be the prior probabilities of the n events, which are trans¬
formed by an (indirect) message to the posterior probabilitiesyu ...,yn. Then
(8.4) i{y-x)= £ J>« log

;=i xt
is the expected information of this indirect message. It is the expected value

of information values of the form (8.2) in the same way as (8.3) is the
expectation corresponding to (8.1). Both (8.3) and (8.4) are nonnegative; the
components (8.2) of (8.4) are not all nonnegative, however, contrary to the
components 18.1) of (8.3).
The information expectation I(y:x) has two different prediction inter¬

pretations. First, let xx, xn be the probabilities of Ex, ..., En before the
forecast is made, and let this forecast change these probabilities to yx, ...,yn;
then I(y:x) is the information content of that forecast. Second, let xu ..., xn
be predictions of the allocation fractions of Ex, ..., En and yx, ..., yn the
corresponding realized fractions; then I(y:x) is the information inaccuracy
of the forecasts. In the former case the forecast is considered more valuable
when the posterior probabilities yt differ more from the prior probabilities xt.
In the latter case the forecast is considered more valuable when the posterior
probabilities yt are closer to the prior probabilities x{.
Continuing the second interpretation, let xx, ..., xn be the original pre¬
dictions of the allocation fractions, x\, ..., x'„ revised predictions, and
yu ..., yn the realizations of these fractions. Then
(8.5) I{y-x)- I(y:x')= £ yt log *-

i=l xt
is the information improvement of the revisions x\ over the original fore¬

casts xh given the realizations yt. Its value may be positive, zero, or negative,
just as the information gain (8.2).
(5) The concepts of bivariate information theory, introduced in Section 2.4,

will be summarized in Section 3.1.
CHAPTER 3
ECONOMIC RELATIONS
INVOLVING CONDITIONAL PROBABILITIES
3.1. Conditional Distributions and Their Entropies

This chapter will be devoted to various uses of conditional probabilities in
economic problems. We start by considering two-dimensional distributions,
for which purpose we return to Section 2.4, particularly to its messages sent
Vj, ..., Xm and its messages received Yj, ..., Yn. We definedas the proba¬
bility that Xt is sent and Ys is received, pL as the probability that Xt is sent,
and p j as the probability that Yj is received. This led to the one-dimensional
(marginal) entropies:
m 1 "1
(1.1) H(X)= £ ft. log- H(Y)= X Pylog-
‘■=i Pi. J= i P.j
to the two-dimensional (joint) entropy:
m n i
(1-2) H(X,Y)=ZZPlJ log

i = 1 j= 1 Pij
and to the expected mutual information:
m " v-■
C1-3) hr = E Z Pu lo§
i=i j=i pLP.j
which are connected by
(1.4) IXY = H(X) + H(Y) - H(X, Y)
Let us now consider the ratio Pij/Pi., which is the conditional probability
that Yj is the message received when it is given that Xt is the message sent.
Take a fixed Xt and let Y, vary over the set of messages received. The entropy
of the resulting distribution is
n
(1.5) V —log—
L Pi. Pu
i = 1,..., m
j=i
49
50 RELATIONS INVOLVING CONDITIONAL PROBABILITIES
This is the conditional entropy of the message received distribution, given

that the message sent is Xt. Next we take a weighted average of the m entropies
(1.5) with the probabilities pu of the conditions as weights:
m n
(1.6) HX(Y)= I I Pij log
This is the average conditional entropy of the messages received, given the
messages sent. We have:
m n m n m n
1
Z Z Pij lo§ - = Z Z P„ log
Pij i=lj=l
Z Z Pij Io§ -Pi.
i=1j =1 Pij i = 1 j= 1
m «
m 1
= «=ZI JZ= 1 Pij lQg Pij
n Z
i— 1
pl lo§ Pi.
and hence:
(1.7) Hx(Y)=H(X, Y) — H (X)
which means that the average conditional entropy of the messages received,
given the messages sent, is equal to the excess of the joint entropy over the
univariate entropy of the messages sent. On comparing (1.7) with (1.4) we
conclude:
(1.8) IXY = H(Y) — HX(Y)
In words: The expected mutual information is equal to the excess of the un¬
conditional entropy of the messages received over the average conditional
entropy, given the messages sent. Note also that we have:
(1.9) HX(Y)^H(Y) H(X)^H(X,Y)
The first inequality follows from (1.8) and the nonnegativity of IXY. It means
that the uncertainty as to the message received when it is known which
message is sent is on the average at most equal to the uncertainty which
prevails when this knowledge is not available.1 The equality sign applies if
1 Note that this holds only on the average. As an example, consider the following 2x2
array:
Yi Y2
Xi Pn = i Piz = i Pu=i
X2 P 21=4 P22=0 P2.=\
P-1=1 P-2 = i
3.1 CONDITIONAL DISTRIBUTIONS 51
and only if IXY = 0, which corresponds to independence of messages sent and

received. It stands to reason that the conditional uncertainty as to Y is then
the same as the unconditional uncertainty. The second inequality of (1.9)
follows from (1.7) and from the nonnegativity of HX(Y). [This nonnegativity
follows from the definition (1.6), which implies that HX{Y) is a weighted
average with nonnegative weights of the nonnegative entropies (1.5).] The
equality sign applies here if and only if HX(Y) = 0, i.e., if and only if it is cer¬
tain what value Y will take when the X value is given, for whatever X value
which has positive probability.
We can also look from the standpoint of the messages received and thus
take Pij/p.j as our starting point, which is the conditional probability thatX;
is the message sent when Yj is the message received. This leads to the con¬
ditional entropy of the message sent distribution, given that Yj is
received:
;= i
On weighting these entropies with the probabilities pj of their conditions

we obtain:
m n „ .
(1.10) Hy(X)= X X ptJ log—'

i=lj= 1 Pij
which is the average conditional entropy of the messages sent, given the
messages received. We have
Hy(X) = H(X,Y)-H(Y)
(LU) Ixy = H(X)-Hy(X)
corresponding to (1.7) and (1.8), and also
(1.12) Hy(X)^H(X) H(Y)<H(X,Y)
which corresponds to (1.9).
The unconditional entropy of Y is
H(Y) log i +i log 4 = .81 bit
The conditional entropy of Ygiven X=X\ is 1 bit; the conditional entropy given X =X*_
is zero; hence the average conditional entropy of Y (weights pi. —pz. — j) is i bit. This
average is indeed less than H(Y), but the first conditional entropy (given X — Xx) is larger
than H(Y). See also footnote 1 on page 58.
3.2. The Determinants of Production Plan Revisions
The relevance of the concepts introduced in the previous section will now
be illustrated by means of the surveys of the increase-no change-decrease
type which we met at various places in Chapters 1 and 2. Our problem can be
described as follows. Suppose a participant plans a lower rate of production
in month t but reports one month later that he kept the rate unchanged or
even that he raised it. In such a case one may argue that the participant made
a positive plan revision, because the actual production rate turned out to be
above the level that was planned before. We have the same kind of revision
when the participant plans “no change” but reports one month later that
there was actually an increase. There may also be negative plan revisions,
which occur when a no-change plan is followed by a decrease realization or
when an increase plan is followed by a no-change or a decrease realization.
We shall disregard all cases in which the plan was carried out, so that there is
always a plan revision, either positive or negative. The problem to be con¬
sidered is: What can be said about the determinants of such revisions? Fol¬
lowing Anderson and his associates,1 we shall take the following two factors
into account:
(1) It may be that the participant plans to reduce the rate of production
because he anticipates a drop in orders. Suppose that this drop does not
materialize and that no-change in orders received is reported one month
later, or perhaps even an increase. Then we should not be surprised when the
participant refrains from lowering the rate of production, i.e., when he makes
a positive plan revision. Therefore, let us make a dichotomy for orders
received similar to that of production. On the one hand we have a positive
surprise on orders received, which is defined as the case in which a decrease
anticipation for this variable is followed by a no-change or by an increase
realization or in which a no-change anticipation is followed by an increase
realization. On the other hand there is a negative surprise, which includes all
those cases in which an increase anticipation is followed by a no-change or
by a decrease realization or in which a no-change anticipation is followed by
a decrease realization. [This is indeed a dichotomy, because the correct-
forecast cases in which the realization is the same as the anticipation are
1 O. Anderson Jr., R. K. Bauer, H. Fuhrer and J. P. Petersen, “On Short-Term

Entrepreneurial Reaction Patterns,” Weltwirtschaftliches Archiv, Vol. 81 (1958), pp.
243-264, particularly the data in the footnote on page 253. The analysis of this section
is largely based on Section 13.1 of Applied Economic Forecasting.
3.2 PRODUCTION PLAN REVISIONS 53
TABLE 3.1
PRODUCTION PLAN REVISIONS DETERMINED BY SURPRISES ON ORDERS RECEIVED

AND THE APPRAISAL OF INVENTORIES
Surprise on Appraisal of Inventories

orders received too small too large
Number of observations
Positive 89 83
Negative 43 164
Relative frequency of positive production plan revisions
68 50
Positive — = .76 — = .60
89 83
14 19
Negative — = .33 -= .12
43 164
supposed to be eliminated.] It stands to reason that when there is a positive

surprise on orders received, the probability of a positive production plan
revision is larger than in the case of a negative surprise.
(2) The participants are also asked to state whether their inventories of
finished goods are too small, or just about normal, or too large. The “normal”
cases have been eliminated before, so there are only two possibilities: the
inventories are too small or too large. We should expect that the probability
of a positive production plan revision is larger when stocks are considered
too small than when they are considered too large.
As a whole, therefore, there are 2x2 = 4 possibilities with respect to the
two determining factors. The number of observations for each of these com¬
binations is given in the upper half of Table 3.1, and the corresponding
relative frequencies of positive production plan revisions in the lower half.
These data, based on surveys among German tanners, confirm our expec¬
tations: Given the appraisal of the inventory level, the frequency of positive
production plan revisions decreases when the surprise on orders received is
negative instead of positive (from .76 to .33 and Irom .60 to .12); and given
the sign of the surprise on orders received, the frequency also decreases when
the appraisal of inventories is “too large” instead of “too small” (from .76
to .60 and from .33 to .12).
The problem posed in the first paragraph of this section has thus been
solved: Both factors affect the probability of a positive production plan

revision. The surprise on orders received appears to be more important as a
determining factor than the appraisal of the inventory level. This can be
concluded from the .60 frequency of a positive plan revision when the sur¬
prise on orders is favorable for such a revision but the inventory appraisal is
not, and from the .33 frequency when the inventory appraisal is favorable
and the other determinant is not, the former being larger than \ and the latter
smaller than The next question to be considered can be described as
follows. We did succeed in obtaining a substantial reduction of the un¬
certainty as to the sign of the production plan revision when both factors are
unfavorable. The probability of a positive revision is then as small as .12 and
hence the probability of a negative revision as high as .88. [Remember that
there is always some revision, either positive or negative, because of the
elimination of the cases in which the production plans were carried out.] We
also succeeded fairly well in reducing the uncertainty as to the sign of the plan
revision for the case in which both factors are favorable, the probabilities
being .76 and .24. We were less successful in the two other cases. The
question is: How successful have we been, on the average, in reducing the
uncertainty on the production behavior by taking these two factors-sur¬
prises on orders and inventory appraisal - into account? It will turn out that
the average conditional entropy defined in the previous section provides a
natural measure for the degree of success, and that the two factors just
mentioned play the role of conditions.
Let us start with the following condition: both factors are favorable for a
positive plan revision (upper left in Table 3.1). The expected information of
the message which conveys us the sign of the plan revision is then
(2.1) .76 log —- + .24 log = .80 bit

.76 .24
Next we take as given a positive surprise on orders received and an inventory

level which is too large. The expected information of our message is then
1 1
(2.2) .60 log-b .40 log — = .97 bit
.60 6 .40
When the condition is a negative surprise on orders received and an in¬

sufficient inventory level, the expected information is
1 1
(2.3) .33 log — + .67 log = .91 bit
.33 .67
3.3 MULTIVARIATE INFORMATION THEORY 55
and when both factors are unfavorable for a positive plan revision:
(2-4) .12 log — + .88 log 1 =.38 bit
We conclude from the upper half of Table 3.1 that there are 379 observations
as a whole, so that the relative frequencies of the four conditions underlying
the expected information values (2.1) through (2.4) are obtained by dividing
the upper-half absolute frequencies of the table by 379. The average value of
the four information expectations is therefore
89 , 83 43 164
(2.5) - (.80) + (.97)+ (.91) + —(.38) .67 bit
379 379V ’ 379V ’ 379V ;
This is the number of bits which is to be expected, on the average, from the
message on the sign of the production plan revision when the values of the
two determinants (orders received and inventory appraisal) are known.
3.3. Multivariate Information Theory
In this section we shall interpret the result just obtained in terms of con¬
ditional and unconditional entropies. Since we deal with three variables
(production plan revisions, surprises on orders received, inventory appraisal),
we shall need a multivariate generalization of the analysis of Section 3.1. This
generalization is straightforward, but the message sent-message received
interpretation is specifically two-dimensional and therefore no longer ap¬
propriate.
Let us indicate the production plan revision by X, the surprise on orders
received by Y, and the appraisal of the inventory level by Z. Each of these
variables takes the two values:
X1— positive production plan revision

X2 = negative revision
yj = positive surprise on orders received
Y2 = negative surprise
Zj = inventories are too small
Z2 = inventories are too large
We shall write piJk for the probability of X=Xh Y= Yp Z=Zk, where the
three indices are either 1 or 2. As in Section 3.1 we shall indicate probabilities
of marginal distributions by putting a dot on the place of the index over which
summation takes place. For example:
2 2 2
P.jk Pijk P..k Pijk

i= 1 i= 1 j= 1
The relative frequencies in the lower half of Table 3.1 correspond to the
conditional probabilitiesPijk/p,jk; that is, the probability of X=Xt (a positive
or negative plan revision), given Y=Yj and Z=Zk (given the sign of the
surprise on orders received and the appraisal of inventories). The expected
information values derived in (2.1) through (2.4) are those of the conditional
distributions of X, given four alternative conditions. These information
values are weighted in (2.5) by the probabilities p jk of these conditions.
Hence the algebraic variant of (2.5) is
/fP 111 p 11 P 211 P.ll

p.ll log + log
VP. 11 Pill P.ll P211
hn 2 P.12 P 212 P.12

+ P.12 ( log + log
VP. 12 Pi 12 P.12 P 212
jfP\2\ P.21 P 221 P.21

+ -P.2 1 1 log + log
VP.21 Pi 2 1 P.21 P 221
/^Pi 27. P.22 P 222 P.22

+ P.22 log + log
\V P.22 P122 P.22 P222
= L £ Iîog-i-- £ y Pjk ,og 1

i= lj=U=l pijk j=lk=l P.jk
= H(X, Y,Z)-H(Y,Z)
We conclude that the expected information (2.5) is nothing else than the
excess of the joint entropy of the three-dimensional distribution of all our
variables over the joint entropy of the two-dimensional distribution of the
two explanatory variables. It seems rather obvious to compare this excess
with the average conditional entropies HX(Y) and HY(X) of equations (1.7)
and (1.11). In fact, it has a similar interpretation. Let us consider the con¬
ditional probability Pijklp.jk of X=Xh given Y=Yj, Z—Zk. The entropy of
this conditional distribution is
V Puk log Pjk

UP.jk Pijk
3.3 MULTIVARIATE INFORMATION THEORY 57
which is analogous to the expression (1.5). We weight these conditional

entropies with the probabilities p jk of their conditions and obtain
(3.i) i j k Pijk
= H(X, Y,Z)-H(Y,Z)
see the result obtained at the end of the previous paragraph. HYZ(X) is the
average conditional entropy of X, given Y and Z, and it is therefore a
straightforward generalization of HY(X) as given in (1.10) and (1.11). We
conclude that the numerical result obtained at the end of the previous section
[.67 bit, see (2.5)] is nothing else than the average conditional entropy of the
production plan revisions, given the surprise on orders received and the
appraisal of the inventory level.
We found in (1.12) that the conditional entropy HY(X) is at most equal to
H(X). This means that, on the average, knowledge of Y will certainly not in¬
crease our uncertainty with respect to X; the uncertainty will be reduced when
X and Y are not stochastically independent. This suggests that when a
condition is added, as in the case of HYZ(X) compared with HY(X) and HZ(X),
the value will become still smaller, at least not larger:
(3.2) Hyz{X)^Hy{X),Hz{X)
This is indeed true, which is proved as follows. Consider the difference

between HY(X) and HYZ(X):
Pj. P.jk
Hy(X)-Hrz(X) = ZEp,J. log lo8
i j Pij. i j k Pijk
(3.3)
= Y,YY,Pijk Io§ Pi-JX-
Pij.P.jk
^ Q
i j k
P-j.
The inequality sign follows from the fact that the triple sum on the second
line can be regarded as the expected information of an indirect message
which transforms the prior probabilities Pij.P.jk/P.j. t° the posterior proba¬
bilities pijk. [Note that these prior probabilities do indeed add up to 1 when
summed over i,j and k.] The difference between HY(X) and HYZ(X} vanishes
when the logarithm on the second line is log 1 for each subscript combination,
i.e., when
Pijk = Pip for all triples (/,;', k)
(3.4)
P.jk P.j.
This amounts to the special case in which the conditional probability

PijklP.jk is independent of k. In words: Given 7= Yp the chance of X=Xt is
the same whatever value Z may take, for each pair (/, j). In other words:
When 7 is given, knowledge of Z does not contribute anything to a reduction
of the uncertainty of X. In that case we have HYZ(X) = HY(X), but in all other
cases Hyz(X)<Hy(X). The analogous assertion on HZ(X) in (3.2) is proved
in the same way; the interpretation of its excess over HYZ(X) is also similar.
It is quite interesting to compare the average conditional entropies with
correlation coefficients in regression analysis. Actually, the objective of the
analysis of this section and its predecessor is largely the same as that of
regression analysis, because in both cases we are interested in “explaining"
the behavior of some dependent variable (here: production plan revisions) in
terms of certain independent variables (surprises on orders received and the
appraisal of inventories). Consider then
(3-5) 1 — R-x.yz — (1 — rfr)(l — rxz.r)
where RXYZ is the multiple correlation coefficient corresponding to the

dependent variable X and the independent variables Yand Z, rXY the ordinary
product-moment correlation of X and Y, and rXZ Y the partial correlation
of X and Z, given fixed Y. The left-hand side of (3.5) can be regarded as a
measure for the uncertainty of X, given Y and Z, so that it is comparable with
our Hyz(X)} The first factor on the right, 1— rXY, measures the uncertainty
of X, given Y only, so that it corresponds to HY(X). That factor is equal to
the left-hand side if it is multiplied by 1, i.e., if the partial correlation rxz y
of X and Z given Y is zero. This condition implies that Z has nothing to say
about X after Y said what it had to say. Clearly, this corresponds to (3.4) and
hence to a zero value of the difference (3.3). Whenever Z does have something
to say about Zafter 7 had its turn, the squared partial correlation rxzyand
the entropy difference (3.3) are both positive. The analogy of r\Z Y and
Hy(X) — Hyz(X) is also corroborated by the fact that both are symmetric
in X and Z:
(3.6) HY(X) - HYZ(X) = Hy(Z) - Hyx{Z)
1 Note that the left-hand side of (3.5) is a measure for the degree of uncertainty of X,
given Y and Z, on the average (in the mean square sense). This expression is equal to the
ratio of the mean square deviation of the X values from the regression plane to the mean
square deviation of the X values from their average. It may well be that there is an indi¬
vidual observation for which the deviation of its X value from the regression plane exceeds
the deviation from the average of all X values. This is fully analogous to the example given
in footnote 1 on page 50 for the average conditional entropy as a measure of uncertainty.
3.4 A LOG-LINEAR MODEL FOR CONDITIONAL ODDS 59
This is easily verified by inspecting the triple sum on the second line of (3.3).
These results show that there is in general a successive reduction of the
uncertainty as to the dependent variable (measured by the average condi¬
tional entiopy) when more and more explanatory variables are introduced.
This is nicely illustrated by the example of production plan revision. When
no independent variable is used at all, the uncertainty is measured by the
entropy of the one-dimensional distribution of the plan revisions:
H(x)= Z Pi., log —

;=i Vi..
l l
= -398 log — + .602 log — = .97 bit
.jyo .602
[The frequency .398 is derived straightforwardly from Table 3.1 by computing

(.76)89/379+ (.60)83/379 + ....] When the surprise on orders received is intro¬
duced as an explanatory variable, but not the appraisal of inventories, we
measure the uncertainty by the average conditional entropy
2 2
H,(X)= £ I Pu. log — = .75 bit

i=lj=l Pij.
which is indeed less than H (X). When the inventory appraisal is used as the
only independent variable, we have
2 2
HZ(X)= £ Z Pi.k log — = .89 bit

i=lfc=l Pi.k
which is also below H (X). Finally, when both variables are introduced, we
obtain
Hyz(X) = Z Z Z Pijk log Pj- = .67 bit

i=ij=ik=i Pijk
which is the value (2.5) which we obtained before. It is below both HY(X)
and HZ(X).
3.4. A Log-linear Model for Conditional Odds; The Logit
We shall pursue the analogy with regression analysis a little further by

considering regression coefficients besides correlations. For this purpose it is
important to realize that the dependent variable in regression analysis
corresponds with a conditional probability in the present case. In our example

this conditional probability takes values of the form Pijk/p.jk, the probability
of a positive production plan revision given the values Yj and Zk taken by the
two determining factors. In precisely the same way, regression analysis
associates a value of the dependent variable with each combination of values
taken by the explanatory variables.
The second important observation to be made is that regression analysis
imposes a certain model on the relationship between the dependent and the
independent variables; for example, the linear model X— a + b Y+ cZ. We did
not impose any such model on the conditional probabilities. These were
estimated as relative frequencies of positive plan revisions for each combi¬
nation of values taken by the determining factors. In our example there are
only four such combinations, and hence only four conditional probabilities
to be estimated: Pijk/pjk for j, k= 1, 2. But this number is mn if the first
determining factor takes m values and the second/?; it is n1n2-..nk if we have k
explanatory variables, the /zth taking nh values, h= 1, ..., k. If we would try to
estimate so many conditional probabilities with a reasonable degree of
accuracy, we would need a very large number of observations. It stands to
reason that we prefer to impose a certain model, similar to X=a + bY+cZ,
so that the necessary number of observations is reduced drastically. This is
the objective of the present section.
To illustrate the problem we return to Table 3.1, particularly to the four
conditional probabilities which we reproduce here for the sake of con¬
venience:
Inventories
Surprise on orders: too small too large
positive .76 .60
negative .33 .12
We start in the upper left-hand corner (.76). Then we move to the right
(to .60) and conclude that there is an inventory effect in the sense that the
probability of a positive production plan revision is reduced when the
inventories are considered too large instead of too small. We return to the
upper left-hand corner and then move in downward direction to .33. We
conclude that there is an effect of orders received in the sense that the chance
of a positive production plan revision is reduced when the surprise on orders
received is negative rather than positive. Now we move to the right, so that
the condition implies a negative surprise on orders received and an inventory
level which is too large. We should expect that the corresponding probability
is less than .33, because moving to the right is characterized by the inventory
effect (from “too small” to “too large”) which turned out to be negative in
the first step. Indeed, the chance decreases (from .33 to .12). An increase
from .33 to a higher level is not a priori excluded, but it would mean that
there is something special to the combination of a negative surprise and
excessive inventories which is not present in either of these factors separately.
We shall be interested in the special case in which there is no interaction
between the two explanatory variables.1 That is, there is the effect of a
negative instead of a positive surprise on orders received, and there is the
effect of an inventory level which is too large instead of too small, and when
both events occur (the surprise is negative and the inventories are too large)
the total effect is found by combining the separate effects without any special
interaction effect. We have such a situation in the linear regression
X— a + b Y+ cZ. When Y increases by A Y, the effect on X is an increase bA Y.
When Z increases by AZ, the effect is cAZ. When both changes occur, the
effect on X is simply bAY+cAZ.
The problem is: How is no-interaction to be defined in the present context?
We could do this in terms of a linear pattern on the conditional probabilities.
For example, our 2x2 array of conditional probabilities may then take the
following form:
.8 .4
.5 .1
This is linear in the sense that the values of the second row are obtained from
those of the first by subtracting the same number (.3) and, similarly, the
values of the second column are derived from the first by subtracting another
fixed number (.4). However, negative probabilities are not excluded under
this definition:
.8 .4
.1 -.3
The linear pattern does have one good property, though, which should be
respected. If there is no interaction of two factors with respect to some event
E (a positive production plan revision in our example), it would be absurd if
these factors would interact with respect to the complementary event not E
(negative plan revision). This means that the no-interaction definition should
be symmetric in p and 1— p, where p is any of the conditional probabilities
that enter into the definition. The linear pattern satisfies this condition. Take
1 The term “interaction” is derived from the analysis of variance. In fact, the similarity
with variance analysis goes even farther as will become evident from equation (5.2) below.
the first example and subtract the four figures from 1 to obtain the conditional
probabilities of “not E”\
.2 .6
.5 .9
This pattern is also linear, so that there is indeed no interaction with respect
to “not E” under the linear definition.
The proposal to be made here will be formulated in terms of the odds in
favor of E. [We met the odds briefly in Section 1.7.] Let E be a positive
production plan revision and write p for the upper left conditional proba¬
bility of E in Table 3.1. Write q for 1 —p, so that the odds in favor of E when
there is a positive surprise on orders received and a below-normal inventory
level take the following value:
Writepl for the lower left probability in Table 3.1 and qt for 1 — px. Then the
decrease in the chance from p to px is equivalent to a decrease in the odds
from plq to pjq^. Similarly, write p2 for the upper right probability and
<72 = 1 — Pi- Then the decrease in the chance from p to p2 is equivalent to a
decrease in the odds fromp/q to p2lq2. The two new odds take the following
values:
Pi. _ -33 Pi .60
= 1.5
qx .67 q2 .40
Hence the shift from a positive to a negative surprise on orders received

(given that the inventories are considered too small) reduces the odds in favor
of E to a fraction (,5)/(3.2), and the shift from a below-normal to an above¬
normal level of inventories (given that there is a positive surprise on orders
received) reduces the odds to a fraction (1.5)/(3.2) of the original value.
The no-interaction pattern will now be defined in terms of these relative
changes in the odds. Consider the probability pl2 of E in the case of a negative
surprise on orders received and above-normal inventories. We shall say that
the two factors do not interact in their determination of the chance of a
positive production plan revision when the odds p12/q12 satisfy
^4 Piling _ PtlQi P2IQ2

plq plq plq
That is, when the relative reduction of the odds from upper left to lower right
in the conditional probability array is equal to the product of the relative
reductions in the two elementary steps (one step downward and one step
from left to right). It is easily seen that this definition implies no-interaction
with respect to not E if and only if there is no interaction with respect
to E. This is proved by interchanging the p's and q’s. It is also immediately
seen that when/?,/?1 and p2 are all positive and less than 1, the no-interaction
value of p ^2 has the same property. This value exceeds both p1 and p2 when
both are larger than p; it is below both px and p2 when both are below /?; it is
between pl and p2 when p is between pl and p2. It is also instructive to
compare the no-interaction definition (4.1) with the condition of stochastic
independence of two events. We write S for the set of all possible outcomes,
so that its probability P (S) is 1, and consider two subsets St and S2 as well as
their intersection S12. We have stochastic independence when
P(S12) = P(S0 P(S2)

P(S) P(S) P(S)
Clearly, this independence multiplication rule has a form similar to the no¬
interaction multiplication rule (4.1). The concepts used are different, of
course, but this is the natural consequence of the fact that we describe
different phenomena. It is nevertheless interesting to find this similarity,
because no-interaction can be regarded as a form of mutual independence of
the factors Y and Z as far as their influence on X is concerned.
For the analysis which follows it is convenient to work with logarithms of
odds, to be called logits for reasons that will be explained more fully in
Section 3.7 below. The logit of a positive production plan revision under the
condition of a positive surprise on orders received and a below-normal
inventory level is thus
P 11
(4.2) L = log-= log --log -
1 - p 1 - p P
where p is the corresponding probability. In informational terms we can
regard the logit as the excess of the information content of the message
“not E” over that of the message “is.” It is easily seen that L is a mono-
tonically increasing function of p and hence that it is as much a measure of
the degree of certainty as the probability p is. The probability measures this de¬
gree by a number between 0 and 1, the odds plq perform the same service by
a number between 0 and oo, and the logit L does it by a number between — co
and oo. The customary probability measure is very convenient when we need
an addition rule for two or more mutually exclusive events or a multipli¬
cation rule for two or more stochastically independent events. Here, how-
ever, we consider only one event E and its complement, for which the two
logit values have the convenient property of being equal apart from sign.
This is important when we need a definition which is symmetric in E and its
complement.1 An illustration of the logit as a function of the probability is
given in Figure 3.1; the function is tabulated at the end of this book.
1 The situation becomes more complicated when there are several events (rather than E
and its complement) whose conditional probabilities are described in terms of a number
of determining factors. For example: a positive plan revision, a negative revision, and no
revision at all. For an attempt to impose a reasonably realistic pattern in such a situation,
see Applied Economic Forecasting, Chapter 13.
3.5 ESTIMATING THE LOG-LINEAR MODEL 65
The no-interaction definition is very simple when formulated in conditional

logits. We define Lu L2, Ll2 in a way similar to (4.2); then the definition (4.1)
amounts to
(4-3) L12 — L — (Lx — L) + (L2 — L)
i.e., the conditional logits follow a linear pattern.
3.5. Statistical Estimation of the Log-linear Model
The figures of Table 3.1 are not precisely in accordance with the log-linear
model. It would be surprising if they were in agreement, because they are
relative frequencies which are subject to sampling errors around the true
probabilities. The objective of the present section is to test the validity of the
no-interaction model on the basis of such observed relative frequencies and
to estimate the parameters which determine this model.
We consider the case of two determining factors, Y and Z, the former
taking m and the latter n values. We write njk for the number of observations
for which Y= Yj and Z = Zk, and fjk for the observed relative frequency of E,
given Y= Yj and Z = Zk. In the example of Table 3.1: m = n = 2, the njk are the
numbers in the upper half of the table, and the fjk the fractions in the lower
half. The corresponding logits are
fhjk j = 1,
(5.1) Ljk = log
1 - hjk k = 1,..., n
which should satisfy the linear model
(5.2) LJk — a + + yk
for certain values of a, /?l5 ..., y1, ..., yn. It is easily seen that (5.2) is
implied by (4.3). The total number of coefficients is m + n +1, but there are
two degrees of freedom due to the fact that one can raise all /Ts by A and
all y’s by B and subtract A + B from a without affecting any of the Vs. Hence
there are m + n — 1 coefficients to be determined, which is a very substantial
reduction compared with the number mn of conditional probabilities unless
both m and n are very small.
Since the Ljk as defined in (5.1) are subject to sampling errors, they will not
fit the pattern (5.2) exactly even if there is no interaction in the population.
We shall estimate the coefficients of (5.2) under the assumption that the mn
relative frequencies are based on independent samples from binomial popu-
lations. Hence the L,k are uncorrelated, but they have different variances. We
have from (5.1)
dfj,
jk
dL =
fjtV-fjk.)
under the assumption that natural logarithms are used.1 We interpret dfjk
and dLjk as sampling errors around the parent values, square both sides of
the equation and take the expectation. This gives
var fjk
. i
(5.3) var L jk
m-ur Hjkfjki1 - fjk)
which is a large-sample approximation. This applies both to the procedure of

squaring the differential and also to the use offjk instead of Sfjk in the right-
hand side of (5.3).
The estimation procedure is weighted least-squares with weights that are
inversely proportional to the variances (5.3). So we minimize
X X njkfjk(\ fjk) (Ljk — “ ~ Pj - l/k)2

j=1k= 1
We differentiate with respect to a, and yk and equate the result to zero.

This gives
L..-cc- ^ Wjfij - ^ w_kyk = 0

j=1 k= 1
n
^ Wjk
(5.4) lj. -<x-pj- ) ~yk = 0 j = 1,..., m
wi
k= 1
w jk
Lk cc — Pj - = 0 k= !,...,«
W:
j= 1
where2
njkfjk (1 ~ fjk)
(5.5a) wilc
JK
— m n
X X niJth 0 - fa)
i=lh=l
1 Natural logarithms are more convenient in the analysis of this section than logarithms
to the base 2, but we shall continue to use bits in the numerical applications, because we
prefer not to shift from bits to nits within the same chapter.
2 The weights wjk are defined here such that they add up to 1, so that L.„ Lj. and L.k are
simple weighted averages of Ljk. It will turn out in the Appendix of this chapter, which
deals with standard errors of the estimates of a, /fi and Yk, that it is actually more con¬
venient to use weights Wjk which are equal to the numerator of Wjk.
3.5
ESTIMATING THE LOG-LINEAR MODEL 67
wJ = X
Jfc= 1
wjk W.k = I wjk
J=1
m n
(5.5b)
L - = X I VV^Life
j= 1 k= 1
w’jk
Ljk L.k = Ljk
W; w
k=l j=i
This procedure is easily illustrated with the data of Table 3.1, which
contains the njk in the upper half and the fjk in the lower half. These give:
w11 = .258 w12=:.320

w2i = .152 w22 = .270
The Ljk are obtained from thtfjk and take the following values (in bits):
68 50
Ln = log — = 1.70 Ln = log — = .60
j3
14 19
L2i= log — = - 1.05 L22 = log = - 2.93
29 145
Since there are two degrees of freedom, we may put
(5.6) /?i = 7i = 0
so that (5.4) is then a system of three linear equations in three unknowns

a> P2, Jz■ We need three averages of the Us:
L - (,258)(1.70) + ... +(.270)(— 2.93) = - .32
.152 x .270
L22. = --(- 1.05) +-(- 2.93)> = - 2.26
.422 .422
.320 .270
L 2 - — (.60) + 1-(- 2.93) = - 1.02
.590v ’ .590v ’
The equation system (5.4) is then specified with numerical coefficients. Its
solution is
(5.7) a = 1.87 P2= — 3.22 y2 = - 1.41
Given (5.6) we can interpret the a value as the logit of a positive plan revision
under the condition of a positive surprise on orders received and below-
normal inventories, as the effect of a surprise which is negative rather than
positive, and y2 as the effect of an inventory level which is too large instead of
too small.
Does the pattern (5.6)-(5.7) give a faithful reproduction of the observed
relative frequencies of Table 3.1? To answer this question we arrange the
logits in the familiar 2x2 form:
^12 a + Ai + 7i a + Pi + h ' 1.87 .46"
_L2i -^22_ _a + /? 2 + 7i a + P2 + ?2_ 1.35 - 2.76_
which implies the following conditional probabilities:
.79 .58
(5.8)
.28 .13
If we compare this result with the observed relative frequencies of Table 3.1:
'.76 (.04) .60 (.05)

(5.9)
.33 (.07) .12 (.03)
we can conclude that the agreement is rather close. The figures in parentheses
are standard errors under the null-hypothesis, computed as the square roots
of
j,k= 1,2
ljk
where p°k is the appropriate conditional probability of (5.8). Needless to say,

the pattern (5.8) is an estimated pattern; it is based on the estimates (5.7).
We refer to the Appendix of this chapter for the derivation of their sampling
variances and covariances.
The no-interaction specification is easily generalized for the case of three
or more determining factors. Each additional factor implies an additional
subscript for the logit on the left of (5.2) and an additional one-subscript
term on the right. When there are k determining factors and when the hth
takes nh values, the total number of unconstrained coefficients to be estimated
is Znh — k + 1. They are derived from a linear equation system similar to (5.4).
We conclude by noting that, if the observed conditional frequencies are
replaced by those which correspond to the pattern (5.2), this has its impli¬
cations for the average conditional entropy HYZ(X). In Section 3.2 we
derived this entropy as a weighted average of the four expressions (2.1)-(2.4).
Each of these expressions is now of the form
3.6 A PARTIAL LOG-LINEAR MODEL 69
where p% is one of the four values of (5.8). It is a matter of straightforward

computation to find the values of these four expressions and to weight them
in the manner of (2.5) so as to obtain the conditional entropy value HYZ(X)
that corresponds with the linear logit specification.
3.6. A Partial Log-linear Model
The conditional probability model described in the two previous sections

will in many cases give a reasonably accurate description of the observed
frequencies with a relatively small number of parameters. There will of course
be exceptions, mainly because of interaction. An example is provided by an
early postwar survey among ship and airplane passengers who were asked
about their motives for their choice between these two modes of transpor¬
tation.1 These motives are grouped in three categories:
Mx: time considerations

M2: safety considerations
M3: financial considerations
Each of these considerations plays a role or it does not; hence there are eight
possibilities as a whole, for each of which the relative frequency of the ship
ticket preference can be determined. These conditional frequencies, given
each combination of motives, are presented in the upper half of Table 3.2.
Considering the four figures upper left, we conclude that if time considerations
(Mt) play a role, the frequency of ship preference reduces considerably
(from .93 to .10 and from .98 to .59). When safety considerations (M2) are
mentioned, the frequency goes up (from .93 to .98 and from .10 to .59). We
find a similar pattern for the four figures upper right, where financial
considerations (M3) play a role. The picture becomes more subtle when we
compare Not M3 with M3. If we take the situation in which none of the
considerations is mentioned as a starting point and then shift to the case in
which only financial considerations play a role, the frequency of ship ticket
preference decreases from .93 to .88. This suggests that financial consider-
1 The data are taken from H. Emanuel, L. H. Klaassen and H. Theil, “On the Inter¬
action of Purchasing Motives and the Optimal Programming of Their Activation,”
Management Science, Vol. 7 (1960), pp. 62-79. The balanced interaction concept introduced
in this article serves a purpose similar to the present no-interaction concept, but it is more
complicated due to the fact that effects in different directions have to be handled separately.
It should be noted that the commodities (ship and airplane tickets) and the motives are
fictitious, although the frequencies used are “real.” The reader is therefore invited to take
our “explanation” of the various effects with a grain of salt.
TABLE 3.2
CONDITIONAL FREQUENCIES OF SHIP TICKET PREFERENCE, GIVEN ALTERNATIVE COMBINATIONS
OF MOTIVES
Not M3 M3
Not M2 m2 Not M2 Ma
Not Mi .93 .98 .88 .99

Mi .10 .59 .30 .69
Adjusted values
Not Mi .98 .99

Mi .15 .59 .26 .74
ations induce people to go by plane. However, if we compare any of the

other three figures upper left with the corresponding figure upper right, we
observe precisely the opposite: there are increases from .98 to .99, from .10
to .30, from .59 to .69. This means that financial considerations induce people
to go by boat when either time or safety considerations or both also play a
role. Assuming that the figures are sufficiently stable in the sense that the
effect is not due to sampling errors, we must conclude that this is a case of
interaction: Financial considerations stimulate airplane preference when no
other considerations play a role, but they stimulate ship preference when these
other considerations are mentioned.
Two questions arise: Can we explain this effect? Is it possible to retain any
of the advantages of the linear logit model in this case? The explanation is
fairly simple if we assume that there are actually two groups of passengers.
One group is not particularly interested in time and safety considerations
but more in having a pleasant trip. To have a pleasant trip on a ship can be
rather expensive; hence, if financial considerations are mentioned, they may
be expected to reduce the preference for a ship ticket. The other group which
mentions time or safety or both is more interested in reaching the place of
destination than in the trip itself; given that attitude, financial considerations
work in favor of ship tickets because these are cheaper.
Regarding the second question, we can indeed retain part of the advantages
of the simple linear logit model if we apply it to all conditional frequencies
except the outliers .93 and .88. This amounts to using the pattern only for the
second passenger group. The lower half of Table 3.2 shows the result of this
adjustment for the case in which the Mx effect is put at — 5 bits, the M2 effect
at 3 bits, and the M3 effect at 1 bit. One additional parameter is needed to
3.7 LOGIT REGRESSIONS 71
specify the pattern completely; for that purpose the (Mls Mz, Not M3)
combination is put equal to \ bit. [These convenient figures are preferred to a
more sophisticated statistical estimation, which would be difficult due to the
way in which the data have been made available.] We conclude that six
conditional probabilities are thus specified by four parameters, so that the
partial linear logit model succeeds in saving two coefficients.
3.7. Logit Regressions
Until this point the explanatory factors which determine the conditional
probabilities were qualitative in nature: the sign of the surprise on orders
received, the question whether inventories are considered to be above or
below normal, whether time considerations play a role when crossing the
Atlantic, etc. There are also numerous examples of conditional probabilities
whose conditions involve variables which take numerical values. The
chance p that a family owns a second car depends on family income, family
size and possibly other quantitative (and qualitative) determining factors.
The chancep' that a person will replace his car next year depends on the age
of his car, on his income, on the price of a new car, etc. In such cases a simple
method of estimating the relationship is to compute a logit regression, which
is based on the model
P k
(7.1) log -- = x + Yj Pl,xh
1 - p h=1
where a, f}u ..., pk are parameters to be estimated, while x1, ..., xk are
explanatory variables such as income and family size or transformations of
these variables such as their logarithms. If the model is applied to household
data, say, the equation for the ith household is thus
p *
(7.2) log —5— = a + X (\xhi
1 - Pi h= i
where xhi is the value taken by xh for the ith household and pt the probability
that this household owns a second car (or replaces its car, or anything else)
given that the explanatory variables take the values xu, ..., xki. Note that
there is no random disturbance in (7.2). This is due to the fact that this
equation does not state whether or not the family owns a second car, it only
specifies the probability that the family owns a second car. In fact, the
familiar linear regression model is similar in this respect, since it is based on
k
(7.3) S yt = cc + ^ Phxhi
h= 1
which means that yt (the fth value taken by the dependent variable) is a
random variable whose expectation is equal to a linear combination of the
values taken by the explanatory variables, and on
(7.4) var yi = cr.2
which implies that the variance of yt is a constant. On combining (7.3) and

(7.4) we conclude that the regression model does not specify one particular
value of the dependent variable. It gives a probabilistic statement on this
value, given the values taken by the explanatory variables. In that respect
the regression model is indeed similar to (7.2), which expresses the logit
corresponding to the probability of second-car ownership in terms of the jc’s.
The model (7.2) does not specify explicitly how the probability pt depends
on the determining factors xxi, ..., xki, but it is not difficult to obtain such
an explicit relation. Under the temporary assumption that log stands for
natural logarithm1 we have
(7.5) Pi =
where exp {z} stands for ez. The function (7.5) has the form of a logistic,
which explains the term “logit.”2
The logittransformation of the probability on the left of (7.1) suppliesus with
a measure which is not confined to a finite interval, like (0, 1) in the case of p
itself. This is of great convenience, since the right-hand variables are usually
not confined to a finite interval either. It should be stressed that the logit
transformation is not the only method which achieves this result. An alter¬
native which is even better known is probit analysis.3 This type of analysis
was developed in biology and it is based on the normal distribution, not on
the logistic. Thus, let F( ) be the cumulated distribution function of a normal
variate with zero mean and unit variance; write F(y)=p, where p is the
conditional probability which occurs on the left of (7.1). The transformation
1 Natural logarithms are actually more convenient in logit regressions than logarithms
to the base 2, but we shall nevertheless continue to use bits in the numerical applications.
The term is due to J. Berkson, “Application of the Logistic Function to Bio-assay,”
Journal of the American Statistical Association, Vol. 39 (1944), pp. 357-365.
3 For a detailed exposition see D. J. Finney, Probit Analysis (Second edition; New York:
Cambridge University Press, 1962).
used in probit analysis is y = F~i (/?), where F~l ( ) is the inverse of F( ).

This transformation is not the same as the logarithm of pfq, but its form is
quite similar; it is also symmetric apart from sign in p and q=\-p. The
main theoretical idea underlying the normal curve is that, if some drug is given
to a number of insects (say), they will not die at some critical dosage all at
the same time, but this critical level varies from insect to insect according to
a certain distribution. There are many different factors which determine the
critical level; the central limit theorem is applied to justify the use of the
normal distribution.
If these ideas are applicable in any specific case, one may decide to apply
the probit transformation. The logit transformation, on the other hand, has
the advantage of being more transparent: It is based on the odds p/q, which
have a direct intuitive appeal, followed by a logarithmic transformation
which is so common in econometric regressions. The estimation procedure
is also rather straightforward as we shall shortly see, the number of para¬
meters being comparatively small.1 From now on, therefore, we shall con¬
fine ourselves to logit regressions.
One complication with respect to the estimation of a, />j, ..., of (7.2) is
that the left-hand side of that equation is unknown and has to be estimated.
Take the simplest case k=1 and suppose that xu=\ holds for one-half of
the observations and xu=2 for the other half. These two groups are subsets
of the set of all observations, say 5, and S2. The right-hand side of (7.2) is
equal to « + /?, for each ieSk and to a + 2/?t for each ieS2. Hence p,- is inde¬
pendent of / for i&S], so that this common probability can be estimated by
the relative frequency (of second-car ownership, say) among all observations
of S1; similarly for S2. This situation is analogous to that of Section 3.5,
but it is actually a very special case of the more general situation in which
1 When there are two explanatory variables or more, the normal distribution of the
probit model implies the presence of several means, variances and covariances, which
raises the number of parameters substantially. This procedure was actually followed by
J. S. Cramer, A Statistical Model of the Ownership of Major Consumer Durables (New
York: Cambridge University Press, 1962). A simpler approach was suggested (both for
the logit and for the probit model) by A. Zellner and Tong Hun Lee, Joint Estimation
of Relationships Involving Discrete Random Variables, Econometrica, Vol. 33 11965),
pp. 382-394. Their logit procedure is similar to the one to be considered here, but the
authors are less explicit with respect to the variation of the probabilities pi within each
group Sff. Reference should also be made to “Tobit analysis ; see J. Tobin, Estimation
of Relationships for Limited Dependent Variables,” Econometrica, Vol. 26 (1958), pp.
24-36. This type of analysis deals with dependent variables which cannot be negative but
which need not be probabilities.
the xu values may be all different. We then have to combine observations

with unequal xi; values and compute relative frequencies for such heter¬
ogeneous groups. For example, in the case of one explanatory variable {k= 1)
we can rank the observations according to increasing values of that variable
and define the first group Sj as the set consisting of the first n{ observations,
S2 as the set consisting of the next n2 observations, and so on. When there
are two explanatory variables or more, we have to group in several dimen¬
sions; examples will be given in the next section.
Let us write G for the number of groups and
h = 1, ...,k
(7-6)
g = 1 ,.;G
for the gth group average of the hth explanatory variable, ng being the number
of observations in Sg. Equation (7.1) is then estimated in the form
(7.7) log ,4_

= a + 2 Ph*hg + vg g = 1, •••, G
i h= 1
where fg is the relative frequency of the event in Sg. It is the estimate of the
probability p, computed as n*/ng where «* is the number of cases in which
the event took place (the number of two-car families in Sg). This probability
is not really constant, however, because the x’s take different values within
Sg and thus lead to different p’s in view of (7.1). Also, the relative frequency
fg is subject to random variation. Both aspects are taken care of by the last
term of (7.7), the disturbance vg. To express this disturbance in the observed
fg and the probabilities pt we average both sides of (7.2) over ieSg:
k
(7-8) ^ Z
i eSg
l0§ I - P = a+ Z
h= 1
^h*hB
We subtract (7.8) from (7.7) and find
(7.9) vg = log -A - log -A-

\ 1 ig 1 Pi/
ieS,,
The expression in parentheses can be simplified when fg is not too far from
Pi, ieSg:
U Pi 1 - Pi . Pi
log - log log-log —
1 - fg 1 - Pi 1 - Jfg 6 Jfg
= log
4- pj\
- log
wj
4 P' + £l Pi Jj~Pi
1-4 4 4(W.)
the approximation on the last line being based on the expansion of the
natural logarithm. On combining this result with (7.9) we find
(7.10) f,-pi
4(i - 4)
where pg is the average conditional probability in Sg:
(7.11) Pg = — V Pi
n9 Lj
ieSg
We conclude that the disturbance vg is approximately equal to the dif¬

ference between the observed relative frequency fg of the group Sg and the
average probability pg of that group, divided by fg{\ — fg). The frequency fg
corresponds to ng different populations, each of which is assumed to be
binomial. The expectation of^ is pg, its variance is
- \2 P0O--Pg)-™gpi
6 \Jg ~ Pg) =-—
where vargp, stands for the variance of ph ieSg, around pg} This term will
be neglected; moreover, we shall approximatepgbyfg:
4(i-4)
(7.12) *(f.-P,)2
Hence, as far as the variance of the p’s around pg is concerned, the expression
(7.12) overestimates the true variance offg. We now combine (7.10) and (7.12):
1
(7.13) var vg fife ZA)'
41 2(i-4)2 »94(i -fg)
where the first « sign is based on the neglect of the random character of
fg in the denominator of (7.10).
1 See, e.g., M. G. Kendall and A. Stuart, The Advanced Theory of Statistics (Vol. 1;
London: Charles Griffin & Co., Ltd., 1958), pp. 126-127.
The result (7.13) suggests that the logit regression (7.1) is to be estimated
by weighted least-squares with weights proportional to ngfg(l—fg); that is,
by precisely the same procedure as that of Section 3.5. There is an important
difference, though. In the present case we have the problem of the variation
of the values taken by the determining factors for each conditional frequency
- a problem which we did not have in Section 3.5. To solve it as well as
possible we should select our groups such that the differences between the
p’s within each group are small. The weighted least-squares procedure itself
is obviously simple, but we do have the problem of appropriate grouping;
moreover, there is the fact that (7.13) is only an approximation. To show how
the method works and to appraise its quality we shall present a small simu¬
lation experiment in the next section.
3.8. Simulation of Logit Regressions1
We start with the simplest case of one explanatory variable:
(8.1) jog -----= a + j8xj

1 - Pi
Interpretation: p{ is the chance that the ilh family owns a TV set when its
income takes a value whose logarithm is x;. So the model implies that the TV
logit is a linear function of the logarithm of income. We imagine that in¬
come is lognormally distributed over the set of all families. Thus, by gener¬
ating a normal variate x; by means of a computer we find the chance p; of
a TV set, using (8.1); thereafter the computer determines whether the family
actually has a TV set (chance pt) or not (chance 1 —pi).
Next we divide the xt values into G groups:
— GO < Xf <
X1 < x; ^ X2
XG_ j < X; < CO
and determine the average of the x’s in each group:
ieSg
1 The author is indebted to J. Boas of the Econometric Institute in Rotterdam for pro¬
gramming the simulation computations.
3.8 SIMULATION OF LOGIT REGRESSIONS 77
where ng is the number of observations in Sg. Let n*, ^ng, be the num¬
ber of times in which the computer has assured that a family in Sg does have
a TV set; then
(8-4) fg = i g = 1,G
n9
is the relative TV frequency of Sg.

Various alternative grouping methods (8.2) will be considered. They are
all derived from 18 elementary intervals. The first nine are given below; the
figures on the right are the probabilities that a standardized normal variate
takes a value in the relevant intervals.
1. (-oo, -2) .023

2. (-2, -1.5) .044
3. (-1.5, -1.2) .048
4. (-1.2, -1) .044
5. (-1, -.8) .053
6. (-.8, -.6) .062
7. (-.6, - .4) .070
8. (-.4, - .2) .076
9. (-.2,0) .079
The other nine intervals are symmetric on the positive axis: 10. (0, .2), ..., 18.
(2, oo). Consider then the following five grouping methods:
A. (7=18. Each elementary interval forms a separate group.
B. (7 = 14. Eight elementary intervals are combined pairwise:
Si = (1,2) V2 = (3,4) 513 = (15,16) S14 = (17, 18)
The other groups consist of one elementary interval each: S3 = (5), ... ,
Si2=(14).
C. (7=10. All intervals are combined pairwise except the two middle
intervals:
51 =(1,2)
52 = (3, 4)
53 = ( 5,6)
V4 = (7, 8)
S5 = (9)
plus the five analogous groups on the positive axis.
D. (7 = 7 by the following combinations:
5, =0.2)
S2=(3, 4)
S3 = (5> 6, 7)
5'4 = (8,9, 10,11)
5*5 = (12, 13, 14)
5'6 = (15, 16)
5'7 = (17, 18)
E. G — 5 by the additional combination of the two outer groups of the

previous case:
■Si =(1,2, 3, 4)
5'2 = (5, 6, 7)
S3 = (8,9, 10,11)
£4. = (12, 13, 14)
S5 = (15, 16, 17,18)
The normal distribution of the xt has zero mean and unit variance. Three
alternative a specifications will be considered: —2, 0 , 2, and one for /?,
viz., P = l. Hence the logit in (8.1) varies between —4 and 0 (apart from a
few exceptions) when a= — 2. This range corresponds to apL between yT and
[Bits are used to measure the logits.] The approximate range of pt is
from t to y for a = 0 and from \ to -yf for a = 2. Given these specifications,
the application of weighted least squares is straightforward except that one
should take account of the fact that logarithms to the base 2 are used, which
implies that the expression for the variance of vg given in (7.13) should be
divided by the square of the natural logarithm of 2. Write y for the G-element
vector whose gth element is the logit corresponding to f (base 2), Z for
the G x 2 matrix whose gth row is [1 xg], and D for the GxG diagonal matrix
whose gth diagonal element is ngfg(l—fg) multiplied by the squared natural
logarithm of 2. The estimator of a and p of (8.1) is then
(8.5) =(ZDZ) “ZDy
and its covariance matrix is
(8.6)
IP.
The vector of estimated v’s is
y - Z(Z DZ)-1Z Dy = [I — Z(Z DZ)-1Z D]y
The elements of this vector correspond to “true” p’s which have different
variances according to (7.13). We may therefore decide to standardize by di¬

viding by the square roots of these variances, which amounts to premultiplica¬
tion by D*:
(8.7) D2[I - Z(Z DZ)_1Z D]y
where D2 is the diagonal matrix whose elements are the nonnegative square
roots of the corresponding elements of D. The mean square of the elements
of (8.7) should be approximately 1; we shall compute it by dividing the sum of
squares by G — A — 2, where A is the number of groups for which fg is either
0 or 1. It will be noted that the weighting procedure implies that these
observations are not used at all in the computations.
Table 3.3 contains the means of the point estimates for three sample sizes:
TABLE 3.3
MEANS OF POINT ESTIMATES FOR THE CASE OF ONE EXPLANATORY VARIABLE
« = 200 n = 500 n = 1500 n = 200 n = 500 n = 1500

(25 runs) (10 runs) (10 runs) (25 runs) (10 runs) (10 runs)
Average a estimate Average ft estimate

(Case a = —2) (Case a = —2)
18 -1.73 -1.81 -1.94 .77 .93 .96

14 -1.79 -1.86 -1.95 .85 .99 .97
10 -1.88 -1.90 -1.95 .89 1.01 .97
7 -1.94 -1.90 -1.96 .91 .99 .98
5 -1.98 -1.90 -1.95 .95 1.00 .98
Average a estimate Average /? estimate

(Case a = 0) (Case a = 0)
18 .02 -.04 -.04 .84 .95 .98

14 .01 -.04 -.04 .93 .98 .99
10 .00 -.04 -.04 .95 .98 .99
7 .00 -.04 -.04 .95 .99 .99
5 .01 -.04 -.04 1.00 1.00 .98
Average a estimate Average estimate

(Case a = 2) (Case a = 2)
18 1.72 1.85 1.99 .74 .90 .98

14 1.79 1.87 2.00 .78 .93 .99
10 1.88 1.91 2.01 .83 .95 .99
7 1.95 1.91 2.01 .87 .93 1.00
5 2.01 1.92 2.02 .94 .92 1.01
« = 200, 500, 1500. There are 25 runs for n = 200, 10 runs for « = 500 and
a =1500. The results suggest that there is little bias except when there are
few observations (n = 200) and many groups (G large). Actually, it is rather
implausible that one will decide to have many groups when there are rela¬
tively few observations. Insofar as there is any bias, it seems to be a small
sample bias in zero direction, for the averages are close to the true values
in the case n—1500, while for n = 200 and 500 the means are mostly smaller
in absolute value than the corresponding parameter values.
To judge whether there is indeed a bias we should compare means with
mean squares. This is pursued in Table 3.4, which contains the mean squares
of the sampling errors (based on 25 or 10 runs) and also the mean squares of
the standard errors of the estimates [derived from the diagonal elements of
the right-hand matrix in (8.6)]. The first column shows that for n = 200 and
a = + 2 the precision of the a estimates increases substantially when the number
of groups is reduced. The fourth column contains the mean squares of the corre¬
sponding standard errors. They exhibit the same feature, but to a much lesser
extent. Indeed, it appears to be generally true that the mean square standard
errors are rather insensitive to alternative groupings. Also, the results suggest
that the bias estimates are hardly significant, at least when we disregard the
small n — large G combinations. The variances of the a and ft estimates in the
case ?j = 200, a= ±2 are typically of the order of .1, so that the standard
error of the mean of 25 independent estimates is about .06. This rules out
the significance of most of the corresponding bias estimates. We can draw
similar conclusions for the other n, oc combinations. A comparison of cor¬
responding MS sampling and standard errors shows that the former are
sometimes above and sometimes below the latter. They are roughly speaking
of the same order of magnitude; a much larger number of runs would be
needed to make more precise statements.
We proceed to consider the case of two explanatory variables. Equation
(8.1) is now replaced by
Pi
(8.8) log = a + Pixu + fl2x2i
l- Pi
where x1(- and x2i are both normal variates with zero mean and unit variance.
We shall consider the case of zero correlation (p = 0) between the two variates
and two cases of nonzero correlation: p — .l and p = .9. Such correlated
normal variates are easily obtained from a univariate normal distribution.
Take
3.8
MEAN SQUARE SAMPLING ERRORS AND MEAN SQUARE STANDARD ERRORS FOR THE CASE OF ONE EXPLANATORY VARIABLE
SIMULATION OF LOGIT REGRESSIONS
81
where xu and yt are uncorrelated normal variates, both with zero mean and
unit variance; then x2i has the same mean and variance and its correlation
with .vi; is p.
The procedure is completely analogous to that of the previous case. The
only difference is that Z is now a G x 3 matrix whose gth row is
[1 xlg x2g]
and that the sum of squares of the standardized residual vector (8.7) is to be
divided by G — A — 3. The grouping procedure is affected more drastically,
however, because this should now take place in two dimensions. Two methods
will be used, which can be displayed conveniently in a simple diagram:
-1 0 1
x
2
We have G = 9 on the left, G= 16 on the right. In both cases the grouping

is based on lines parallel to either of the two axes. The upper right-hand
group in the case G = 9 consists of all observations for which xu, x2i>%, in
the case G = 16 of all those for which xi;, x2l>l, and so on.
The numerical specifications of the coefficients are a = — 2, 0, 2 and pi — 1,
P2 — 2- Again, the sample sizes are a = 200, 500, 1500 with 25 runs for « = 200
and 10 runs for n = 200 and 500. The means of the point estimates are shown
in Table 3.5. We see that if the sample size is small (« = 200) the average
a estimates corresponding to oc= + 2 are all less in absolute value than the
true value. This suggests again a small-sample bias. For the ^ estimates
this is also true apart from two exceptions; for fi2 there is no such evidence.
It will be clear that p = .9 represents a situation close to multicollinearity.
A dramatic example is the case « = 500, a= — 2, G = 9. The average of the
ten /?, estimates for p — .9 is as low as .45, that of the ten fi2 estimates as
high as 1.05. We should expect that such average outcomes will be rare when
the number of runs increases.
3.8
SIMULATION OF LOGIT REGRESSIONS 83
TABLE 3.5
means of point estimates for the case of two explanatory variables
p n =200
IF
on
o
n =500
o
n =200 n =500 /z == 1500
^4v. a estimate (a = — 2, G =9) Av. a estimate (a= -2, G = 16)

0 -1.85 -1.89 -1.95 -1.70 -1.87 -1.95
.7 -1.83 -1.90 -1.95 -1.72 -1.85 -1.95
.9 -1.96 -1.90 -1.97 -1.87 -1.88 -1.96
Av. y?i estimate (a = — 2, G = 9) Av. Pi estimate (a = -2, G= 16)
0 .94 .95 .96 .77 .92 .98
.7 .93 .99 .94 .89 .91 .91
.9 .71 .45 .98 .88 .86 .86
Av. Pi estimate (a. = — 2, (7 = 9) Av. P2 estimate (a = -2, G = 16)
0 .47 .55 .48 .40 .54 .49
.7 .48 .48 .48 .38 .49 .50
.9 .65 1.05 .49 .43 .62 .59
Av. a estimate (a =0 , (7 = 9) Av. a estimate (a = 0, G = 16)

0 .00 -.01 -.02 -.02 -.00 -.02
.7 -.02 -.04 -.01 -.02 -.03 -.01
.9 -.04 -.01 -.01 -.04 -.00 -.01
Av. /?] estimate (a ==0, (7 = 9) Av. Pi estimate (a = 0,(7=16)
0 1.03 .99 1.00 .86 .93 .99
.7 1.03 .98 1.03 .97 .87 1.03
.9 .64 .79 1.02 .84 .93 .90
Av. /?2 estimate (a =0 ,(7=9) Av. Pi estimate (a = 0, c7 = 16)
0 .50 .53 .52 .45 .51 .50
.7 .43 .44 .35 .35 .56 .39
.9 .76 .62 .39 .56 .49 .53
1
Av. a estimate (a ==2, (7=9) Av. a estimate (a= 2, G = 16)
0 1.83 1.93 2.00 1.72 1.90 2.01
.7 1.89 1.91 2.01 1.80 1.87 2.01
.9 1.94 1.97 1.91 | 1.86 1.95 1.93
A v. Pi estimate (a = 2, (7 =9) Av. Pi estimate (a = 2, G = 16)

0 .94 1.02 .96 j .84 .97 .96
.7 .67 1.19 1.11 .66 1.02 1.02
.9 .81 .96 .87 .68 .81 .87
Av. P^ estimate (a ==2, (7=9) Av. Pi estimate (a =-2, G = 16)

0 .53 .41 .50 1 .51 .37 .50
.7 .66 .26 .43 .54 .38 .53
.9 .65 .52 .51 .67 .62 .55
o
o
O 10 00 00 NOr-
no y—i Cl 0 —H Cl 0 o o o
8 q q q q q *”2
(M
q q T—i 8 o o o
li
V. NO NO O'
si
o
v,
X.
Q)
V
O
O VO
x. Ss. V
11
MEAN SQUARE SAMPLING ERRORS AND MEAN SQUARE STANDARD ERRORS FOR THE CASE OF TWO EXPLANATORY VARIABLES
7 II
o
o es N5 00 CC e> NO NO On O
no
"ts
X. 73 co c- m
>s „ CO i'- CO d d
cc Cl 0 0 CO o o o
II -2 d 0 0 0 Cl 0 0 C3
x 1
I
si
r* s: 1 s: 1
2 g
II III 11
CO co
O Jo 8 J3y to
o
to
Cl Cl
C<J
NO d (N ^ ^
in to
d 00 On ON § 0 5 ON Cl
q q q Cl as q Cl as o o o
li
si
O
o
00 Cl __ c~ ON NO o
0 T—< . 0 Cl 0 CJ O r-^ 00 o o —<
8 q q q =5. q q ==5. q q q o o o
II
NO
s: <3
V
O
xO 0.
0
g nO
V J*.
o
11 X. 7
0
II ||
o 0 N3 »c
NO ON Cl CO NO CO
|p r „ Cl ON 5p Cl CO NO —< <N Tj-
co c~
II
■§, o' q q q
O 0 0 Cl q q d Cl q q CO
s; | 1 1 1
1 1
$ I 11
III
8
Q
CO
II Q II § «
O 'w' 8 ^3
o
d §
Cl On 0
§
NO co cc
0
ON
NO
0
*—<
d d d
co co
NO co On On r- § NO
c-
II
q q q C-
q q
Si
o
O r-, Cl On 00 NO h h
1— y—i 1-^ Cl Ci <M
—-i Cl CO o o o
II
8 q q q
«ol
q q
CQ.
q q o o o
s .__
s; ON V. ON ON
<3 O
||
5<.
|| X.
X.
||
O <o
O O
o ^2 0 d CO r- NO NO
“a
NO of 0
X. <3 o\ o o
co co co co 00 OO ro 00 0 —' d d
(N cf cT
II ■§
q q q q q m q q
I o o o
| ■S 1 | s:
s:
2
11 g II
s:
Q 11
a
CO Co Co
8 ^8
o
o S2 to to
0 co 00 00 ON ^ ^ d
Cl 00 ON ON 0 NO NO § ON NO CO >n >n in
ii 0 0 0 1 Cl ) 0 Cl 1—1 o o o
s; T“<
©
ON NO co m 00 d 0 no co on
O *—i y—H 0 *—i r—1
CQ 0 1—1 d O O O
||
8 q q q =1. q q cl q q d O O O
s; ON X. ON ON
|| V. || X. ||
O 7j
o <0 co
O
Sp ON On Cl 0 00 (N >n m
II
cf
1
co
O
co
O s f cf
1
0
0 O
NO £p
cf
1
CO
0
NO
O NO
H
— ci ^
o o o
s: 1
1
$3 ll a 11 5 II I l
^8 8 Co 8
© 50 to to to W
O ON NO
O OO r-
T—I
c-
r-
Cl
NO
O
c-
ON
c-
d
On
O
d
co
O ON
5 < d
ii T_-' O 0 0 co O 0 d On o S O
"
r- ON d ON d On r- on
<5. O 0 0 o
3.8 SIMULATION OF LOGIT REGRESSIONS
.008 85
.018
.077
oo 0 CN un nf- un
MS standard error: Pi
fN O OQ (N O
o o o q q q q q T-H q q
V.
(a = 0,G = 16)
V JN. /-S
3 3 X.
S' 0) NO ©
4. <3
1“H
3
7 7 3 II II
.025
.055
.238
nh o “3 rn NO un nf
<2) on nj- "3 ’"Q (N (N
*s <N rn m m rn 00 <23
o o 04 q O q q q rn q q m
cT oi
ss •8 Csf ■§ ri
§ s: !3
3 II 3 II 3 II II
05 03 0!
.074
.632
on
.163
o m 0 00 r-~ NO ON 0 m
NO NO m § ON 00 ON O 04 un ON NO
o NO q q q y—H 04 00 q CN OO
.004
.053
on r-
.021
m 00 0 (N rn
O <N i—i 0 <N O 0
MS sampling error: Pi
CQ. q o o q q q q q q q q
=4
N.
(a = 0, G = 16)
e> SO NO <3 NO O NO
3 4.
3 ii II 3 II ||
.066
.340
.028
<N ON NO 0 m nf 0 O ON 00
<2) <N ? (23 m <23 «N r- nf <2> rn NO un
o o 04 q q q q q m q q m
© ri cs (N
1 1
1
s: II II II s: II
a 3
02 ,q
O} s s
.680
.079
t-H
.117
ON ON 0 04 un m (N
<N 04 o r- O § ON NO 00 § 00 m <N
o r- q 04 un q T—( vn
.019
04 00 nf- <N r- wn
.008
.093
OO ON rn O O
o — On i-H *—1 i—i 1—< oi m ea i-H (N m
ea
MS standard error: Pi
q o o « q q q q q q q
=Q.
X.
V.
(a = 0, G = 9)
3 ON X. ON ON ON
11 V. || V. ||
3 II 3 3 NU
o- ON un m
.276
m NO O
.059
.025
on o~~ ON *3 (5 <23
T3 <2> <N on o- V. m rn m "a m 00 ON 33 m 00 ON
o o (N q q q q q m q q m
o' -§ oi oi rf
•8 ■8
s; II || || x- ||
'(3 <3
03 S 03 «
03 02
§ un 00 <N 0 ON
§ un ON (N
.175
.735
S ON
.071
5 NO 00 00 ON *5 O »/> NO ON M't NO
O i—i o- O 0 O (N r“? O (N T—H
1
OO un 00 m 00 1—1 OO r-
.020
un m
.005
.048
O
<N O m r- O r— 1 0 (N T-H CQ O i-H On
MS sampling error: Pi
q q q q q q q q q q q
s <3 7 3 57 3
(a = 0, G = 9)
X. S7 V
!<
S 5-.
G\
II II II II
On m 00 ON ON 0 00 m
.016
.099
00
04
1
.429
Sp
.s: 00 $ m m rn
q
.1 m
q
ON
0
0
<N •1
m
O
un
q q q
q n r7 cs ri
1
3
0
II s
a
II
Q
03
II s
3
<n
8
II
03 ^^
1 § §
.666
.208
%
.763
.063
.102
.251
.095
m
.588
.153
00
•205
•697
•063
(N 00
q q
on
The multicollinearity problem becomes much more evident when we con¬

sider mean square sampling and mean square standard errors. These are
given in Table 3.6. As in the one-variable case, most of the corresponding
sampling and standard errors are roughly of the same order of magnitude.
In all cases it is evident that the uncertainty with respect to the multiplicative
coefficients increases substantially when p increases. Their standard errors
are approximately the same for n = 200, p = 0 as for « = 1500, p = .9, for most
of the G, a combinations. As to the effect of G, we find that the mean square
standard errors of the a estimates are practically the same for G=9 and 16.
There is a clearer effect with respect to the multiplicative coefficients, par¬
ticularly when p is close to 1. Evidently, G=16 is preferable to G = 9, even
when the sample size is only 200. This is in contrast with our findings for the
one-variable case. It must be due to the fact that, when the explanatory vari¬
ables are highly positively correlated, a small number of groups suppresses
the evidence of the effect on TV ownership when the two variables have
opposite signs.
TABLE 3.7
AVERAGE PERCENTAGE OF MISSING GROUPS
a=—2 a =0 a—2 a= —2 a=0 a=2 a= —2 a=0 a=2
n = 200 n = 500 n = 1500

One explanatory variable
G=
18 17 9 19 4 1 4 i 1 2
14 9 2 11 0 0 1 0 0 0
10 6 2 8 0 0 0 0 0 0
7 8 2 10 0 0 0 0 0 0
5 2 0 2 0 0 0 0 0 0
Two explanatory variables, G = 9

P=
0 4 1 6 1 0 0 0 0 0
.7 17 10 16 3 2 4 0 0 0
.9 27 22 26 21 22 23 22 21 21
Two explanatory variables, G == 16

P=
0 20 13 22 4 1 4 1 0 1
.7 38 26 37 21 14 22 10 11 12
.9 48 40 46 39 37 41 36 37 36
TABLE 3.8
MEAN SQUARE STANDARDIZED RESIDUALS
a= —2 a=0 a=2 a= —2 a=0 a=2 a= —2 a=0 a = 2
n = 200 n = 500
T-H
o
o
s:
II
One explanatory variable
18 .77 .74 .68 1.09 .94 .98 .87 1.23 .79

14 .83 .84 .71 1.16 1.04 1.04 .91 1.39 .78
10 .89 .94 .82 1.11 1.09 1.04 .86 1.32 .83
7 .96 .92 .77 1.38 1.31 1.17 .81 1.55 .91
5 1.04 1.01 .70 1.69 1.42 1.51 .98 1.88 .89
Two explanatory variables, G = 9

P=
0 .88 1.19 .88 .99 .97 .88 .90 .70 .83
.7 .64 1.03 .85 .64 1.11 .95 .84 .97 .88
.9 .83 .73 .88 1.06 .74 1.23 .92 1.41 1.13
Two explanatory variables, G == 16

P=
0 .87 .81 .80 .80 .80 .82 .96 .80 .75
.7 .76 .75 .79 .81 .98 .77 .83 1.06 .87
.9 .84 .76 .87 .91 .91 .71 1.34 1.01 .93
A large number of groups raises the number of those groups Sg whose data
are not used in the computations because /^ is either zero or one. In addition,
there is a substantial increase in the number of empty groups, particularly
when p is large. The average percentage of all such missing groups is pre¬
sented in Table 3.7, which shows that it can be as high as almost 50.
Finally, we have the residuals vg. Their mean squares after standardization,
averaged over 25 or 10 runs, are presented in Table 3.8. They should be com¬
pared with their theoretical unit value. If we take square roots, so that we
obtain standard errors instead of mean squares, we find that almost two thirds
of the 99 figures are in the interval (.9, 1.1). There are about twice as many
below .9 as above 1.1, which again suggests that there is a small bias in zero
direction.
Consider the m + n + 1 estimation equations (5.4). We multiply the (j+ l)st

equation by wh,j = l, m, and the (k + m + l)st equation by w,k, k = 1,
..., n. The coefficient matrix on the left then takes the following symmetric
form:
1 W2. • •• *.2 •• W
rv .n
W1. *m. w.l
wl. Wj 0 . .. 0 *11 *12 •'.. WlB
*2. 0 w2. • .. 0 *21 *22 •' *2n
*m. 0 0 . •• *m. *»1 *m2 •'•• *m»
Wu 0 .,.. 0
* 1 *21 ' •• w»i *1
*2 WX2 W22 •• W„2 0 *2 .. 0
W.n Win W 2n • •• wmn 0 0 • ■ *.„.
This matrix is postmultiplied by the column of coefficients to be estimated,

and the result is another column in logits. These two vectors (written in row
form) are as follows:
[« Pi ••• Pm 7i yn ]
lL.. *iA. ... wmLm waLa ... w„L„]
In our special case where only three coefficients are to be estimated (m — n— 2,

P i =71—0) the equation system becomes
_ 1 * . *. 2 2 a L
*. *.
2 2 *22 P2 = W 2.L2'
_*.2 *22 *.2 _ -W.2L.2_
The further derivations will be simplified still more if we do not use the
“relative” weights wJk of (5.5) which have been obtained by normalization
such that they add up to 1, but the “absolute” weights
(A.2)
appendix 89
and their partial and total sums:
n m m n
(A.3) Wj. = £ WJk W„=Y, WJk
w.. = Y Z Wjk
* =1 j=1 j= 1 k= 1
Hence (A.l) is now written as
w w2 w2 ~ S WL ~
(A.4) w2. w2_ w22 =
h w2l2.
_W2 W22 w2 _ j2_ _w.2l.2_
where the hats on a, p2, y2 indicate that these are the estimates whose sam¬
pling variances and covariances will now be derived.
The underlying model is (5.2), which will be written in the more explicit
form
(A-5) LJk = a + pj + yk + sJk
where ejk is a random disturbance which is caused by the fact that Ljk is
based on a sample frequency rather than on a theoretical probability. We
have
1
(A. 6) var eJk «
wjk
because of (5.3) and (A.2). We proceed to consider the right-hand side of
(A.4), using (5.5) and (A.5):
WL, = H WJkL jk = W, a + W2JS2 + W. +H

j k j k
W2.L2. — Y ^2k^2k ~ W(2.a + ^2.^2 + ^22^2 + Y ^2fce2*

k k
^2-^.2 = Y Wj2Lj2 = W2a + W22p2 + W_2y2 + Y ^j2sj2

J j
On comparing this with the left-hand side of (A.4) we conclude:
~w w2. w2 ~ S — oc zzwjksjk
w2, w2, w22 =
ft2 ~ P2 ^^2ke2k
_w_2 w22 w2_ _y2 — y2_ _ zwji8j2_
Now the variance of Wjk£jk is approximately W?kIWjk = Wjk in view of (A.6).

Hence the variance of the top element of the right-hand vector in (A.7)
becomes ZZWjk = W , because the a’s are uncorrelated. In the same way we
find W2. for the variance of the second element of that vector, W2 for that of
the third, W2 for the covariance of the first and second element, etc. After
working this out we find that the covariance matrix is equal to the 3x3
matrix on the left of (A.7), which in our case takes the following value:
62.1 26.2 36.7'

(A.8) 26.2 26.2 16.8
36.7 16.8 36.7
The covariance matrix of the estimates a, /?2, y2 is then found by solving

(A.7) with respect to the sampling errors, postmultiplying the solution vector
by its own transpose, and taking the expectation. It is easily seen that the
result is of the form P_1PP_1 = P_1, where P is the matrix (A.8). In our
case P-1 is
4.84 - 2.46 - 3.71'
(A. 9) 1(T2 - 2.46 6.64 - .58
- 3.71 - .58 6.70
The standard errors of the estimates (5.7) are then found by taking the square
roots of the diagonal elements and dividing by the natural logarithm of 2
[because the estimates are expressed in bits whereas (A.6) is based on natural
logarithms]. This gives:
a = 1.87 (.32)
(A.10) = - 3.22(.37)
72 — ~ 1.41 (-37)
CHAPTER 4
THE MEASUREMENT OF INCOME INEQUALITY
4.1. Inequality at the Individual Level; The Aggregation Problem
We consider an arbitrary group of income recipients; for example, all

taxpayers in the United States in 1960. We assume that none of the incomes
are negative (losses) and that at least some of them are positive. Thus, when
there are N individuals, there are N nonnegative amounts of individual in¬
come which add up to a positive amount of total income. Equivalently, each
individual earns a nonnegative fraction yh i= 1,..N, of total income, and the
sum of the j’s is 1:
(1.1) Z Ei = 1 yt>0 * = 1»—»N

i= 1
The question to be considered is: Does information theory supply us with a
“natural” measure of income inequality among the N individuals which is
based on these y’s?
It does. We have complete equality when all individuals earn the same
income. This amounts to y; = l/A, /=1, ..., N. We have complete inequality
when one individual’s income is equal to total income, all others earning
nothing at all. Then yt= 1 for some yj = 0 for each7^/. The former case
corresponds with the maximum, log N, of
* 1
(1.2) h00 = X yt lQg -
i= 1 Ti
and the latter with the minimum of H(y), zero. Clearly, H(y) is nothing
else than the entropy of the income shares yl9 yN.
These results suggest that H\y) can be regarded as a measure of income
equality. It will appear below that is it preferable to work with a measure of
income inequality. This is easily obtained by subtracting H(y) from its own
maximum value:
N
(1.3) log N — H(y)= X U log Nyt

i= 1
It is immediately obvious that this inequality measure varies between zero
91
92 THE MEASUREMENT OF INCOME INEQUALITY
(complete equality) and log N (complete inequality).1 * * * * To verify whether it

is an acceptable measure for income inequality we shall apply the following
three tests:
(1) An obvious requirement is that the measure should not change when
all incomes change proportionally. This condition is satisfied, because such
income changes do not affect the income shares.
(2) It may seem objectionable that the upper limit log N of (1.3) increases
when the number of individuals increases. However, it is easily seen that
such an increasing limit is quite natural. When a society consists of two
individuals, we have a maximum of inequality when one receives everything
and the other gets nothing, in which case the value of (1.3) is log 2. When a
society consists of two million individuals, inequality is at its peak when as
many as 1,999,999 persons earn nothing at all. This is evidently much more
unequal, which is corroborated by the fact that the value of (1.3) is now the
logarithm of 2xl06. In our first society one half of the individuals has
everything and the other half has nothing. One might therefore guess that
the second society would be as unequal as the first when half of the individ¬
uals earn the total income and when each of these has the same income. Does
(1.3) satisfy this condition? The answer is in the affirmative.
To show this we assume that a set S consisting of Mindividuals, 0 < M< N,
earns all the income, and that their income shares are all equal to 1/M, so
that everyone in S has indeed the same income. Then we have:
N N 1
Z yt lo§ Myt = Z log Nyt = log — = log -
i= 1 ieS M 9
where 9 = M/N is the fraction of those individuals who jointly account for
total income. We conclude that, when total income is equally distributed
among some subset S of the population, all individuals outside S earning
nothing at all, the inequality measure (1.3) is uniquely determined by the
fraction 9 of the individuals who are part of S. In our example we have 0 = Jr,
so that the value of (1.3) is log 2 = 1 bit.
1 The excess of the maximum entropy value over the actual value is frequently called
“redundancy” in communication theory. [Sometimes the relative redundancy is used,
obtained by dividing by log N, but such a division is undesirable for decomposition
purposes - a subject considered later in this section.] See, e.g., L. S. Schwartz, Principles
of Coding, Filtering, and Information Theory (Baltimore-London: Spartan Books, Inc., and
Cleaver-Hume Press, Ltd., 1963), pp. 19-25.
4.1 INEQUALITY AT THE INDIVIDUAL LEVEL 93
(3) An even more obvious test can be described as follows. Suppose that
individual i is richer than individual j, yi>yj, and suppose that the income
of poor j increases at the expense of rich i in such a way that the total income
of both remains unchanged. The incomes of all other individuals do not
change; hence the constraint just mentioned amounts to y.+yj = constant.
The question is: Does the inequality measure (1.3) show a decrease in ine¬
quality up to the point where y—yj, and an increase thereafter?
The remainder of this section is devoted to this question. It will be answered
in such a way that the aggregation problem is handled at the same time. So
we suppose, as we did in Section 2.7, that there are G sets S1? ..., SG. Each
individual belongs to exactly one such set. We write Ng for the number of
individuals in Sg, g — 1, ..., G, so that
(1.4) ZA, = n
9= 1
We return to the entropy (1.2), which we write as follows:

G
(1.5) H(y)= £
9=1 ieSg
and consider the expression in brackets:
y;log-=Tgy log - ] +log

ki
ieSg ieSa
1
— YgH (y) + Yg log
where
(1.6) y, = I yt g = l,.-.,G
ieSg
(LI) ■' '."

i e Sg
On combining this result with (1.5) we conclude
(1.8) H(y)= £ r,log2+ £

9=1 D 0-1
The first term on the right is the between-set entropy, Yg being the income
share of Sg. The second term is a weighted average of the within-set entropies
H (y). Each of these deals with conditional income shares, all incomes being
measured as fractions of the corresponding set incomes rather than total in-
come. It will be clear that the decomposition (1.8) is quite similar to equation
(7.4) of Section 2.7.
We can now answer the question on the two individuals i and j. Let
consist of these two, so that its entropy is
Hx (y) = v log
Yi
fi 7,
ff^log
yj
where Y1 =yi+>’J = constant. Initially we have yi>yj, after which yj rises at
the expense of yt. It is clear that Hx (y) increases up to the point yt = Vj, where
it is equal to log 2, after which it decreases until it vanishes for yt — 0, yj = Yl.
Since the other terms on the right of (1.8) are not affected, we conclude that
the changes in H(y) are the same as those of H1(y) apart from the positive
multiplier Yv For the inequality measure (1.3) these changes are the same
apart from sign. The inequality decreases until the point yf=yj and increases
thereafter. Hence our question on rich i and poor j can indeed be answered
in the affirmative.
The decomposition (1.8) shows also that H(y) is not really a suitable
measure for income equality (as we alluded to in the third paragraph of this
section). The first term on the right takes its maximum value for Yg = l/G,
<7 G, which means that all sets have the same total income. However,
= 1, ...,
these sets will in general have different numbers of individuals, TVl5 ..., NG,
so that there may be sizable per capita income differences between the sets
even when the total set incomes are the same - and these per capita differ¬
ences are the things which count when between-set income comparisons are
made. For example, when we consider Whites and Negroes as two sets of
individuals, there should be a maximum of equality between the sets when
the average income of the Whites is equal to the average income of the
Negroes, not when total White income is equal to total Negro income.
Let us now consider the inequality measure (1.3) instead of the entropy
(1.2). We apply (1.8) and find:
log TV - if 00 = log TV - £ log 1 - X W>0

9— 1 *g 9~ 1
= log TV - X Yg logTVg - f Yg log 1

3=1 3=1 *g
+ I YgU°ENg - Hg(y)]
3=1
= £ y3 log T~ + I Yg [log TV9 - Hg(y)-]

3=1 iVs/iV 3=1
4.1 INEQUALITY AT THE INDIVIDUAL LEVEL 95
This is in view of (1.3) and (1.7) equivalent to

N G G
(1.9)
5>
i= 1
log
37
VN
= L
) niog njn +
9-1
Y„
.
L
9=1
Y,
[Ei
ieSg
i E Sn
log
ytfYg
The first term on the right, which deals with the between-set inequality, can
be regarded as the expected information of an indirect message. The prior
probabilities are the shares Ng/N of the various sets in the total number
of individuals (the population shares). The posterior probabilities are the
income shares Yg of the various sets. Clearly, when the per capita incomes
of all G sets are the same, the income shares and population shares are pair¬
wise equal:
T __ income of Sg Ng x per capita income of Sg

9 total income Nx total per capita income
= — if all per capita incomes are equal

N
In that special case the first term on the right in (1.9) vanishes. In all other
cases it is positive; its value depends on the degree to which the per capita
incomes differ. We can conclude that the objections raised to the first term
on the right of (1.8) as a measure of between-set equality do not apply to the
corresponding term of (1.9) as a measure of between-set inequality.
The other terms on the right of (1.9) and the left-hand side can also be
regarded as expected informations of the same type of message. The prior
probabilities on the left are all equal to i/N and hence equal to the population
share of each individual in the set of all individuals. In the separate within-set
expressions on the right we have 1 /Ng as prior probabilities, which are
population shares within the relevant set. In both cases the prior probabilities
are independent of i. The posterior probabilities do depend on i; on the left
they are of the form yh which is an unconditional income share, on the right
we have yJYg, which is a conditional income share. We thus obtain aggre¬
gation consistency when we interpret the inequality measure (1.3) as the
expected information of a message which transforms population shares into
income shares.
Our starting point was the entropy H(y) as a measure of equality, which
we modified slightly to an inequality measure by subtracting H(y) from its
maximum value. We now find that this measure has a simple interpretation
in terms of income shares and population shares; moreover, that it can be
aggregated in a straightforward manner. In that respect it is more attractive

than most well-known inequality measures such as Gini’s concentration ratio.
We refer to the Appendix of this chapter (Section 4. A) for a discussion of
this ratio and of some other inequality measures.
4.2. Some Simple Continuous Income Distributions; White and Nonwhite

Families in the United States
The distribution whose probabilities are the income shares is of course

quite different from the corresponding “ordinary” income distribution. The
former distribution identifies all N individuals separately and specifies a
probability yt for each of them. The ordinary income distributions do not
identify any individual at all. They just specify the number of individuals
in any given income interval. So we may ask the question: What form does
the inequality measure (1.3) take when we use an ordinary income distribu¬
tion, such as the lognormal or the Pareto distribution? After this question has
been answered we shall apply the measure to some empirical data. Natural
logarithms will be used throughout this chapter; hence information will be
measured in nits, not in bits.
Let us write z for the income of any individual and m for the average
income of all N individuals. Then total income is Nm and this individual’s
yt is equal to z/Nm. Hence:
z z
k; logiV^ = log
Nm m
We write/( ) for the density function of the income distribution, so that

the total number of individuals who earn an income between z — ^dz and
z+\dz is Nf(z)dz. We conclude:
Z Vi log Nyt = ( Z log - ) Nf(z) dz

its \Nm mj
= (-log V(z)dz
\m my
where S stands for the set of all individuals whose incomes are at most \dz
fromz. The measure of income inequality is then found by integration over z:
oo
(2'° l(^l0g3/(2)‘iZ = <f(lzl08Tz)

0
where S stands for the expectation operator corresponding to the income dis-
4.2
CONTINUOUS INCOME DISTRIBUTIONS 97
tnbution. We conclude that income inequality is measured by the expected

product of income and the logarithm of income, where it is to be understood
that income is measured as a fraction of its own expectation £z=m. It is
immediately seen that the measure takes the zero value when all incomes are
equal.
The application to the lognormal distribution is straightforward. Its densi¬
ty function is
(2.2)
1 1 Qogz - n)2\
/(“)
2 cr 2 ;
where exp{x} stands for ex, while pi and g2 are the mean and the variance,
respectively, of the logarithm of income. The mean of income itself is
(2.3) m = eV + ia2
We make the substitution C = log z, which implies z = e? and dz=ecdC. Then

(2.1) can be written as follows:
1 (log z - pi):
log exp < - dz
mu Jin, m, !
(( - pi- itr) exp e^dl

g y/2n H^j
e 1 (C - pi - <72)2
(C - pi - id ) exp 1 — - + pi + jG2 > d£
a yj2n
<
2\2'i
1 1 (C ~ pi~ c2)
= (Js d T*7 ) exP s - ■d£ = jG2
g y/2n a
We conclude that in the lognormal case inequality (in nits) is measured by

half the variance of the natural logarithm of income.
Next we consider the Pareto distribution. The probability of finding an
individual whose income exceeds z is then (z/d)~a, where a is the Pareto
constant (which we assume to be larger than 1) and d is the lower limit of
the interval for which the Pareto distribution is assumed to be valid. Hence
we deal with a conditional income distribution, the condition being that in¬
come z is not less than some given positive value d. We can assume without
loss of generality that units are chosen such that d= 1, so that the probability
just mentioned becomes z~a. This implies that the density function is
(2.4) / (z) = az_a_ 1 z>l
The conditional mean, given that income exceeds d = 1, is then
(2.5) m
7 1
For the inequality measure (2.1) we can now write:
a— 1
z/logz-log--jaz a 1 dz
« j
i
= (a - 1) z a log z - log I dz
a — l,
or after the transformation ( = log z:
-(a- IK I
(a - 1) C - log \dC
J a — 1
0
We apply another substitution, £' = (a —1)(, which gives:
' C a ' 1
(2.6) -- - log -- K = log
,a — 1 a — 1 a — 1 a — 1
This is easily tabulated. For some familiar a values the result is as follows:
a 1.2 1.4 1.6 1.8 2.0 2.2 2.4

Inequality (2.1) in nits 3.21 1.25 .69 .44 .31 .23 .18
We shall now apply the inequality measure (2.1) to observed income dis¬
tributions of White families and of Nonwhite families (Negroes, Indians,
Japanese, Chinese and other nonwhites) in the United States. The data are
taken from the Current Population Reports (Consumer Income) of the Bureau
of the Census. They are available for each year in the period 1947 through
1963 with the exception of 1953; for further details we refer to the Appendix
of this chapter (Section 4.B). An example of such a pair of income distri¬
butions (one for White and one for Nonwhite families) is presented in Table
4.1. It shows that in 1963 the percentage of White families with an income less
4.2 CONTINUOUS INCOME DISTRIBUTIONS 99
TABLE 4.1
INCOME DISTRIBUTIONS OF WHITE FAMILIES AND OF NONWHITE FAMILIES
IN THE UNITED STATES, 1 963
Non¬ Non¬
Income interval ($) White white Income interval ($) White white
(%) (%) (%) (%)
less than 1000 3.2 9.2 between 7000 and 8000 9.6 4.7
between 1000 and 1500 2.6 8.3 between 8000 and 9000 8.2 4.4
between 3500 and 4000 3.9 6.1 larger than 25000 1.1 .2
between 4000 and 5000 8.8 10.9
between 5000 and 6000 11.3 Total 100.0 100.0
8.7
between 6000 and 7000 10.5 7.3 Total number of families* 42,663 4,773
* In thousands.
than $ 1000 was 3.2, that the corresponding Nonwhite percentage was 9.2,
etc. The table shows also a difficulty in the application of the inequality
measure (2.1). When the income distribution is only specified in terms of
percentages of income recipients for rather wide intervals, we have to use
approximate methods to compute the measure. One method consists of
fitting a free-hand curve through the observed distribution, followed by
numerical integration in accordance with the left-hand side of (2.1). Another
method will be used here. It is equally approximate, but it has the advantage
that its results can be reproduced: One acts as if the income recipients within
each interval have the same income, viz., the amount which corresponds with
the midpoint of the interval.1 It will be clear that this procedure underestimates
the true inequality level, because it puts the inequality within each income
bracket equal to zero. We refer to Section 4.C of the Appendix, which con¬
tains an attempt to formulate limits to the inequality level.
1 For the open interval (larger than 325,000 in 1963) the common amount of income is
defined as a multiple 3/2 of the lower limit of this interval in the case of White families,
and as a multiple 4/3 in the case of Nonwhite families. These multiples correspond with
a values of 3 and 4, respectively, of the Pareto approximation to the observed distributions
in the highest brackets. [This approximation is not satisfactory for any wide range of
income; see below in the text.]
TABLE 4.2
AVERAGE INCOME AND INCOME INEQUALITY FOR WHITE AND NONWHITE FAMILIES, 1947-1963
Average income ($) Inequality (in nits)
Year Between
White Nonwhite Total White Nonwhite
groups
1947 3760 2033 .249 .236 .286 .010

1948 3892 2113 .244 .231 .280 .010
1949 3740 1952 .250 .236 .287 .012
1950 3982 2141 .249 .236 .282 .011
1951 4419 2359 .237 .223 .277 .011
1952 4741 2411 .258 .241 .329 .013
1954 4986 2784 .255 .243 .281 .010
1955 5257 2906 .241 .229 .252 .010
1956 5712 3087 .239 .226 .263 .011
1957 5805 3258 .226 .213 .270 .010
1958 6041 3363 .233 .219 .291 .010
1959 6449 3525 .240 .225 .288 .011
1960 6756 3954 .245 .232 .291 .009
1961 7030 4091 .256 .243 .310 .009
1962 7255 4055 .238 .224 .273 .011
1963 7541 4460 .226 .213 .266 .009
When the procedure just described is applied to the data of Table 4.1,
we find an average White family income of w = 7541 and a Nonwhite average
m = 4460 (in dollars per year). The computation of the inequality measures
is then straightforward. The results are presented in Table 4.2. The first
two columns show that the average family income increased by slightly more
than 100 per cent (in current dollars) during the postwar period, both for
Whites and for Nonwhites. Columns 4 and 5 indicate that the inequality
among Whites fluctuated between .21 and .25 nit and the inequality among
Nonwhites between .26 and .31 nit (apart from a high value in 1952 and a low
value in 1955). It is fairly plausible that part of the fluctuations over time is
due to the sampling errors of the survey. Also, there is probably some dis¬
tortion due to changes in the limits of the income intervals. The only safe
conclusion seems to be that the inequality among Nonwhites exceeds that
among Whites. We observe - for comparison purposes - that an inequality
value of .23 nit corresponds with an a value of 2.2 if the distribution is of the
Pareto type, and a value of .28 nit with an a value of slightly over 2. However,
4.3 STATES OF THE UNITED STATES 101
before identifying these a values with typical White and Nonwhite income
distributions one should note, first, that the distributions are not nearly of
the Pareto type, and second, that the present distributions deal with all
positive income values, not with income above a certain positive limit.
The third column of Table 4.2 specifies inequality among all families,
color being disregarded, and the last column contains the White-Nonwhite
between-group inequality. [These are the two expressions on the left and the
right of the equality sign in (1.9).] The between-group figures are fairly
constant around a value of 1 per cent of a nit. There is not much evidence of
a narrowing gap between the two population groups. A further analysis
would require a more detailed regional analysis. It would lead to a map of the
United States indicating the inequality among Whites and among Nonwhites
in each region and also the White-Nonwhite between-group inequality by
regions. It would further give figures for the inequality between regions, both
for the two population groups separately and for all residents in each region;
it would also be possible to take other characteristics such as age into account.
We shall now go one step in this direction by considering the regional
inequality problem, leaving personal characteristics such as race and age
aside. Our basic data are then in the form of per capita incomes and popu¬
lation shares, so that the problem of incomplete data - numbers of families
in income brackets rather than individual family incomes - does not arise.
4.3. Two-level Aggregation: States and Sets of States in the United States
In this section we shall be concerned with the incomes of the residents of

the states of the United States. Our starting point is again the decomposition
(1.9), which states that total inequality
N N
i= 1 l'=l
can be written as the sum of the between set inequality

G
(3.1)
9= 1
and the total within-set inequality

G
g —1 ieSg
The second component deals with inequality among individuals and will be
disregarded, because our present objective is the measurement of the ine¬
quality among states. This is given by (3.1), G being the number of states and
Yg and NJN the income share and the population share, respectively, of the
gth state. This notation is not convenient, however, since we shall want to
apply a second aggregation by combining states to larger areas. Let us write n
for the number of sets (states in this section), xt for the population share of
the /th set, and yt for its income share. Then the inequality between sets is
(3-2) I(y:x)=fjyi log—

i= 1 *i
which is precisely the same as (3.1) except for the notation. This notation
corresponds with that of equation (2.4) of Section 2.2 in which we introduced
the expected information of an indirect message.
We shall now give a more direct interpretation of the inequality measure
(3.2). Since xt and yt are shares, we have
population of the zth set

xt =-••-
total population of all n sets
total income of the ith set

y. =-
total income of all n sets
and hence:
yt per capita income of the ith set
(3.3)
x; per capita income of all n sets
In other words, the ratio yjxi is the “deflated” per capita income of the z',h
set, where “deflated” means that each set’s per capita income is divided
by overall per capita income. We conclude that / (y:x) is the logarithm of a
weighted geometric mean of deflated per capita incomes:
(3.4)
the weights being the income shares yt. These weights are comparatively large
for sets whose per capita income is high, which implies that the weighted mean
exceeds 1 unless all per capita incomes are the same. This is of course only
an intuitive argument; a more formal proof of this inequality was given in
Section 2.2.
We shall now apply the measure (3.2) to n = 49 states of the United
4.3
STATES OF THE UNITED STATES 103
TABLE 4.3
PER CAPITA INCOME INEQUALITY AMONG THE STATES OF THE UNITED STATES, 1929-1961
Total Between-set Within-set

Year Between-set
inequality inequality inequality as a percentage
1929 .0718 .0592 .0125 82.5

1930 .0795 .0665 .0130 83.6
31 .0823 .0671 .0152 81.5
32 .0888 .0721 .0167 81.2
33 .0804 .0648 .0155 80.7
34 .0712 .0589 .0123 82.7
35 .0607 .0503 .0104 82.8
36 .0630 .0526 .0103 83.6
37 .0576 .0475 .0101 82.5
38 .0581 .0470 .0112 80.8
39 .0588 .0477 .0111 81.2
1940 .0587 .0483 .0104 82.3
41 .0496 .0421 .0075 84.9
42 .0379 .0323 .0056 85.3
43 .0347 .0298 .0049 85.9
44 .0288 .0245 .0043 85.0
45 .0267 .0220 .0047 82.4
46 .0289 .0234 .0055 80.8
47 .0269 .0223 .0046 82.8
48 .0241 .0200 .0041 82.9
49 .0240 .0195 .0045 81.2
1950 .0256 .0210 .0046 82.1
51 .0244 .0204 .0040 83.6
52 .0233 .0192 .0041 82.5
53 .0242 .0202 .0040 83.6
54 .0223 .0179 .0044 80.1
55 .0220 .0176 .0043 80.3
56 .0227 .0180 .0048 79.1
57 .0227 .0179 .0048 78.7
58 .0198 .0156 .0042 78.6
59 .0201 .0158 .0044 78.2
1960 .0200 .0158 .0041 79.2
61 .0197 .0153 .0044 77.9
Note. All information values are expressed in nits.

States.1 These are the 48 “classical” states plus the District of Columbia. The
period to be considered is 1929-1961, which necessitates the exclusion of
Alaska and Hawaii for which no comparable data are available. The income
concept used is state personal income, which may be briefly described as the
total income, gross of taxes, received by the inhabitants of the state from all
sources, including income received in kind.2 The results are presented in the
first column of Table 4.3. They show that there was an increase from about
.07 to almost .09 nit in the first few years, after which inequality declined to
less than .06 nit during the later part of the thirties. There was a consider¬
able inequality reduction during the war, the 1945 figure being less than one-
half of the 1940 value. Thereafter there was a further gradual reduction of
inequality after the war. The figures of the years around 1960 are only one-
quarter of those in the early thirties.
All figures of the first column of Table 4.3 are larger than the between-
group inequality values of Table 4.2. This may seem surprising, given that
there is such a large difference between the White and Nonwhite average fami¬
ly incomes. It is appropriate, however, to take account of the relative size of
the two population groups. For example, consider the case x1 = .9, x2 = .l,
which corresponds approximately with the shares of White and Nonwhite,
respectively, in the total number of families. Take y± = .94, y2 —.06, so that
the ratio of the average incomes of the two groups is
y1/xl _ .94/.9 47
1.741
3^2 = d)6/Tl = 27
which also corresponds approximately with the figures of Table 4.2. The
between-group inequality is then
.94 .06
.94 log + .06 log = .0102 nit
.9 T
1 The same data were used by J. W. Hooper and H. Theil in “The Information Approach
to the Measurement of Income Inequality,” Report 6501 of the Econometric Institute of
the Netherlands School of Economics (1965). The approach followed in that paper is
slightly different. Its inequality measure is I(y:x) + I(x:y), which has a less obvious con¬
nection with individual income inequality and which has more complicated aggregation
properties. For a discussion of I{x:y), see Section 4.A of the Appendix.
For details see C. F. Schwartz and R. E. Graham, Jr., Personal Income by States since
1929, A Supplement to the Survey of Current Business (Washington, D.C., 1956), pp.
57-65. The data are contained in this publication (pp. 140-145) for all years through 1955;
for the later years they are taken from E. J. Coleman, “Personal Income by States in 1961,”
Survey of Current Business, Vol. 42, No. 8 (August 1962), pp. 8-17.
4.3 STATES OF THE UNITED STATES 105
Compare this with the case — x2 — .5 (equal number of White and Nonwhite
families) but in which the average-income ratio is the same as before:
yilxi = ki _ 47
yi!x2 y2 27
Hence = 47/74, y2 = 27/74. The inequality is now
47 , 47/74 27 27/74
74 °g -J- + 74 lQg 5 =-0370 nit
which is more than three times as large as in the case xx = .9, x2 = A. The
White-Nonwhite inequality would therefore be much larger if the poor group
were larger relative to the rich group. The dependence of inequality on popu¬
lation shares, given fixed per capita incomes, is not more than natural. Take
n — 2 and let x2 approach zero; then the inequality measure (3.2) decreases
with zero as limit, for whatever fixed values of the two per capita incomes,
because in the limit xx =y1 = 1 and x2 =y2 =0. This is as it should be, because
in that limit the inequality problem has ceased to exist due to the disappear¬
ance of one of the two groups. We shall return to this subject in Section 4.6.
We proceed to aggregate the 49 states to eight sets of states. The grouping
used is the one which was suggested by the U.S. Department of Commerce;
it is based primarily on homogeneity of the states as regards the income level,
the industrial composition of the labor force, and some non-economic charac¬
teristics.1 The sets consist of adjacent states and can be described as follows:
(1) New England, consisting of 6 states: Maine, New Hampshire, Vermont,
Massachusetts, Rhode Island and Connecticut
(2) Mideast, consisting of 6 states: New York, New Jersey, Pennsylvania,
Delaware, Maryland and District of Columbia
(3) Great Lakes, consisting of 5 states: Michigan, Ohio, Indiana, Illinois
and Wisconsin
(4) Plains, consisting of 7 states: Minnesota, Iowa, Missouri, North
Dakota, South Dakota, Nebraska and Kansas
(5) Southeast, consisting of 12 states: Virginia, West Virginia, Kentucky,
Tennessee, North Carolina, South Carolina, Georgia, Florida, Alabama,
Mississippi, Louisiana and Arkansas
(6) Southwest, consisting of 4 states: Oklahoma, Texas, New Mexico and
Arizona
1 See Schwartz and Graham, loc. cit., p. 139.

(7) Rocky Mountain, consisting of 5 states: Montana, Idaho, Wyoming,

Colorado and Utah
(8) Far West, consisting of 4 states: Washington, Oregon, Nevada and
California.
The aggregation procedure is precisely the same as that of Section 2.7. We
writes ,0 = 1, ...,(/( = 8), for the 0th set of states, Xg for its population share,
and Yg for its income share:
(3.5) ^ = X N I U 3 — 1, •••, G
ieSg
We write <T for the share of the ith state in the population of the set to which
it belongs, and similarly r\t for the conditional income share:
(3.6) & = 3 i = -‘ icSg, g = \,...,G

Ag *g
Then the inequality decomposition can be written as follows:
(3.7) I(y:x) = I0(y:x) + Z yg4(-f:x)

9= 1
where I0(y:x) is the between-set inequality and Ig(y'-x) the inequality with-
in Sg:
Yg
J0(y-x)= Z r9log x„
9- 1
(3.8)
Jg(y :x) V
L Vi log 7
1 'h
0 = 1, ■■■,&
ieSg Qi
The numerical results are given in the second and third column of Table
4.3. Both the between-set inequality and the (total) within-set inequality show
largely the same development over time as that of total inequality. During the
nineteen-thirties the between-set component accounted for slightly more than
80 per cent of the total (see the last column of the table). This percentage
increased to about 85 during the war, after which it declined to the prewar
level. An additional reduction took place in the mid-fifties; from then on the
percentage was slightly less than 80. We conclude that throughout the period
considered the between-set inequality accounted for an overwhelming part
of total inequality. This is not entirely surprising, since the grouping of the
states is partly based on the income criterion. In a way, we may regard the
present results as a test of the quality of the classification procedure.
4.4 INCOME INEQUALITY AMONG COUNTRIES 107
4.4. Income Inequality among Countries: 1949, 1957, 1976
Our next application deals with the shares of 54 countries in the total
income and the total population of their combined area. We shall distinguish
between the following six sets of countries:
(!) North America and Northwest Europe, for which data are available on
13 countries. United States, Canada, United Kingdom, Norway, Sweden,
Finland, Denmark, Germany, Netherlands, Belgium, Luxembourg, France
and Switzerland
(2) Southern Europe, 4 countries: Portugal, Spain, Italy and Greece
(3) Near East, 6 countries: Turkey, Syria, Iraq, Lebanon, Israel and Egypt
(4) Africa, 5 countries: Kenya, Uganda, Congo (formerly Belgian Congo),
Ghana and Nigeria
(5) Asia, 8 countries: Pakistan, India, Ceylon, Burma, Thailand, Malaya,
Philippines and South Korea
(6) Latin America, 18 countries: Mexico, Guatemala, Honduras, Costa
Rica, Panama, Colombia, Ecuador, Peru, Bolivia, Paraguay, Chile, Argen¬
tina, Brazil, Venezuela, Cuba, Dominican Republic, Puerto Rico and Jamaica.
Although these 54 countries are not a random sample in any reasonable
sense, they do account for a considerable part of the world outside the Com¬
munist area. The set of all 54 will simply be called “the world” from now on.
Data on these countries have been made available by Dosser and Peacock
for three years: 1949, 1957 and 1976.1 For the first two years these data
have the character of estimates of past realizations, just as the data on
American states which were used in the previous section. For the last year
the data are conditional forecasts under the assumption that the developed
countries are willing and able to donate all the economic aid which the less-
developed countries can absorb. The estimates dealing with this assumption
were taken from articles by Rosenstein-Rodan and Dosser.2 Obviously,
such forecasts are necessarily rough; the resulting figures on income ine¬
quality should merely be regarded as an approximation to the lower limit of
1 D. Dosser and A. T. Peacock, “The International Distribution of Income with ‘Maxi¬

mum’ Aid,” The Review of Economics and Statistics, Vol. 46 (1964), pp. 432-434. The
countries considered in this paper include South Rhodesia, which is excluded here,
because it was combined with Nyasaland after 1949.
2 P. N. Rosenstein-Rodan, “International Aid for Underdeveloped Countries,” The
Review of Economics and Statistics, Vol. 43 (1961), pp. 107-138; D. Dosser, “Allocating
the Burden of International Aid for Underdeveloped Countries,” ibid., Vol. 45 (1963),
pp. 207-209.
TABLE 4.4
POPULATION SHARES AND INCOME SHARES OF 54 COUNTRIES, 1949, 1957, 1976
Country 1949 1957 1976 1949 1957 1976
Population shares (%) Income shares ( %)

(1) North America and 28.93 28.13 24.47 74.10 72.28 71.59
Northwest Europe
United States 12.56 12.67 11.64 44.73 44.10 40.88
Canada 1.14 1.23 1.20 2.43 2.97 3.00
United Kingdom 4.24 3.82 2.96 10.44 7.85 6.53
Norway .27 .26 .21 .50 .51 .47
Sweden .59 .54 .44 1.46 1.49 1.31
Finland .34 .32 .27 .41 .50 .34
Denmark .36 .33 .27 .80 .64 .59
Germany 4.00 3.81 3.30 4.49 6.69 8.91
Netherlands .84 .82 .68 1.60 1.45 1.48
Belgium .72 .67 .53 1.27 1.25 1.20
Luxembourg .02 .02 .02 .04 .05 .05
France 3.47 3.26 2.62 4.91 3.84 5.94
Switzerland .39 .38 .32 1.02 .96 .90
(2) Southern Europe 7.60 7.03 5.91 5.35 6.04 6.81
Portugal .71 .66 .56 .70 .34 .31
Spain 2.35 2.18 1.83 1.14 1.79 1.39
Italy 3.87 3.59 3.00 3.17 3.45 4.57
Greece .66 .60 .52 .33 .46 .54
(3) Near East 4.23 4.67 5.52 1.77 2.79 2.18
Turkey 1.65 1.89 2.19 .76 1.71 .96
Syria .29 .30 .37 .11 .12 .11
Iraq .42 .48 .54 .12 .15 .17
Lebanon .10 .11 .13 .05 .09 .08
Israel .09 .14 .21 .11 .24 .27
Egypt 1.69 1.74 2.08 .62 .48 .60
(4) Africa 4.25 4.56 4.56 1.03 .94 .83
Kenya .47 .46 .43 .08 .12 .09
Uganda .43 .42 .43 .07 .07 .07
Congo (formerly
Belgian Congo) .95 .97 .96 .18 .18 .16
Ghana .36 .35 .45 .17 .14 .12
Nigeria 2.04 2.36 2.29 .54 .43 .40
(5) Asia 42.67 42.41 44.11 11.31 9.65 9.98
Pakistan 6.21 6.25 6.42 1.55 1.08 1.16
India 29.11 28.67 29.60 8.12 5.90 6.55
Ceylon .61 .68 .76 .18 .23 .20
Burma 1.45 1.48 1.48 .26 .24 .27
Thailand 1.51 1.53 1.69 .27 .43 .42
Malaya .44 .46 .55 .45 .41 .38
Philippines 1.63 1.68 1.88 .26 .81 .68
South Korea 1.70 1.65 1.73 .22 .56 .34
(6) Latin America 12.31 13.19 15.43 6.44 8.30 8.60
Mexico 2.06 2.33 2.91 .85 1.26 1.58
Guatemala .32 .26 .32 .08 .10 .09
Honduras .11 .13 .17 .03 .06 .05
Costa Rica .07 .08 .10 .03 .05 .04
Panama .06 .07 .09 .04 .04 .03
Colombia .93 .98 1.16 .39 .32 .59
Ecuador .28 .29 .36 .04 .10
Peru .09
.69 .73 .87 .25 .22 .27
Bolivia .34 .24 .27 .06 .04 .04
Paraguay .11 .12 .12 .03 .03 .03
Chile .48 .53 .55 .29 .41 .41
Argentina 1.39 1.47 1.45 1.77 1.73 1.71
Brazil 4.15 4.56 5.45 1.60 2.66 2.64
Venezuela .39 .45 .57 .31 .61 .43
Cuba .44 .47 .51 .40 .36 .31
Dominican Republic .19 .20 .24 .04 .10 .08
Puerto Rico .18 .17 .16 .17 .16 .16
Jamaica .12 .12 .13 .05 .07 .08
All countries 100 100 100 100 100 100
4.4 INCOME INEQUALITY AMONG COUNTRIES 109
TABLE 4.5
PER CAPITA INCOME INEQUALITY AMONG 54 COUNTRIES, 1949, 1957, 1976
1949 1957 1976
Total inequality .530 .526 .576

Between-set inequality .456 .462 .545
Within-set inequality .074 .064 .031
Inequality within
N. America and Northwest Europe .077 .062 .021
Southern Europe .034 .013 .061
Near East .041 .166 .083
Africa .051 .038 .012
Asia .040 .061 .027
Latin America .145 .089 .080
Between-set inequality as a percentage
of total inequality 86.0 87.9 94.6
the actual inequality that should be expected in 1976, the argument being that
there will be more inequality when there is less aid. We shall compare the
incomes of different countries in real terms; that is, the incomes are compared
by applying real exchange rates which are obtained by estimating the pur¬
chasing power of the currencies in terms of United States dollars. For further
details we refer to the articles quoted above.
Table 4.4 contains the income and population shares of the 54 countries
in the three years. The inequality measures are presented in Table 4.5. Total
inequality exceeds half a nit - the largest value obtained in this chapter. It
decreased slightly between 1949 and 1957, but is expected to increase by
almost 10 per cent in 1976. The between-set inequality accounts for an even
larger proportion of the total than it did in the case of the American states.
Moreover, this proportion is on the increase. The figures show a monotonic
increase of the between-set inequality and a monotonic decrease of the total
within-set inequality. The regional inequality values show a regular decline
in the case of North America and Northwest Europe, Africa, and Latin
America; in the other cases the picture is less regular. It will be clear that the
total within-set inequality is largely dominated by the inequality within North
America and Northwest Europe, because weighting takes place by means
of income shares - and the share of that region is of the order of 70 per
cent.
Our conclusions are that the inequality among countries is very sub¬
stantial, that the figures indicate that it is likely to increase still further, and
that the problem is largely of a regional nature: The total inequality within
regions is far less than the between-region inequality and this difference is
likely to increase in the near future. Of course, appropriate qualifications
are necessary because of the limited quality of the data.
4.5. Inequality Changes in the Case of Unchanged Population Shares
When comparing the results on White-Nonwhite between-group inequality

with those on the inequality among states in Section 4.3 we observed that the
inequality measure depends both on the ratios of the per capita incomes and
on the size of the sets involved as measured by the population shares. Our pri¬
mary interest is of course in the differences between the per capita incomes;
nevertheless, changes in population shares have their own effect on income in¬
equality. It is therefore of some interest to know how the development of in¬
equality would have been if there were no such effect, i.e., if the population
shares had been constant. The present section is devoted to this question.
In order to have a definite terminology we shall speak about “countries”
and “the world.” We write zt for the per capita income of country i, so that
ZbqZj is world per capita income:
V Population country i Income country i

2_, Xizi — )- x —
i= i World population Population country
i= 1
£ Income country /
l= 1
———--—:-— World per capita income
World population
Hence we can write deflated per capita income as follows:
Y. "
‘ x%-
j= 1
so that each income share can be expressed in terms of per capita incomes and
population shares:
(5.2)
X
j= i
xizj
4.5 UNCHANGED POPULATION SHARES 111
Our inequality measure can therefore also be expressed in the same determi¬
nants:
r -\
(5.3) I(r-x)= X X;
i= 1
X 1
xjzj X
j= 1
xjzj
J
It will be noted that this expression is essentially the same as
which we derived when considering continuous income distributions; see (2.1).

Both here and there we divide the incomes by the average income (ZxjZj or
S’z), multiply these ratios by their logarithms, and then take the average of
all such products.
Consider now any fixed base year in which the population shares are
xt, ..., xn. Using (5.2) we conclude that in any current year t the income
shares would be of the form
XjZj,
(5.4) yv*
n n
n
X
j= 1
xizijt
if the population shares in t were the same as those of the base year and if
the per capita incomes would take the values zit which are actually observed
in the current year. Hence
(5-5) i(y*-x) = X y* log }"

i= 1 X:
is the inequality level in t that would have been observed if the population
shares had been constant, given the actual per capita incomes of that year.
There is a decomposition similar to (3.7). When we introduce
X
W
x“
y* = y« 9 — 1, •••,G
II
ieSg ieSg
(5.6) *
Z. = 4 = yl icSg, g = 1,..., G
*gt
we have
G
(5.7) I(y*:x) = l0(y*-x)+ X YgVgiyf-x)

9=1
TABLE 4.6
PER CAPITA INCOME INEQUALITY AMONG 54 COUNTRIES, GIVEN FIXED POPULATION SHARES
Current Fixed population shares of

Year 1949 1957
values
Total inequality
1949 .530 .530

1957 .526 .522 .526
1976 .576 .540 .547
Between-set inequality
1949 .456 .456

1957 .462 .458 .462
1976 .545 .509 .516
Within-set inequality
1949 .0741 .0741

1957 .0638 .0638 .0638
1976 .0311 .0304 .0304
where I0 (y*: x) is the between-set inequality and 7g(yr*: x) the inequality with-
in Sg:
Io(yf:x)= £ Y*log/
3=1 A9
(5.8)
4 Of:*) = I 'h* lQg T g = 1, ...,G

i e S' 5 Si
The results are shown in Table 4.6 for the 54 countries. They indicate
that when the 1949 population shares are combined with the per capita in¬
come figures of 1957 and 1976, the resulting total inequality in the last two
years would be less than the value which corresponds to the population distri¬
bution in these years. The same applies to 1976 when we use the population
shares of 1957. The main determinant of this effect is the between-set in¬
equality, although the within-set inequality exhibits the same feature in 1976.
We conclude that, given the per capita income figures of the three years, there
would be a smaller increase of the inequality among countries if the distri¬
bution of the world population over these countries remained unchanged.
Table 4.7 contains similar results for the inequality among the American
states. Three base years are selected. The first is 1932, which was the year
UNCHANGED POPULATION SHARES 113
't(Snoon'c\oN\oosmooo^fiH
5 2 2 ^ ^ ^ ^
o
l
PER CAPITA INCOME INEQUALITY AMONG THE STATES OF THE UNITED STATES, GIVEN FIXED POPULATION SHARES
U (N — ^ IQ
*3 o os co
O OS >n
os co t}-’ in in
ON
q. ~ NT Tf Tf T* Tfr Tt
oocoTttNinNOoooocococoocsoinoN — soosfNoocNosr-—'Ostj-nooos
«R3oS»8og"5i$??K$a2222,
Tt Tf rt 2
Nf-!!!22'9Sii(J«
rt rj- Tt tJ- Tf
3a) 1)ft oo in os <n| ^ P* in co

u, j3 so in oi Tt co o —o
t^- in os co — rj-_ (N CN —< oq o os so *n co m co ro os m so
3 13 o — in so r-< m’ in o O on Ti- co oo <N co’ co’
u ►
so in <n o o
TpTf<NOO — Tt
to (N o os ■—
itsoinfooocoinsot'O
o OsOoocoqocosonososo
CN <N CN fN <N
co-ôNinmcoososi^Ttoo—iOconJosminO(N(NTi-
^^fO(NNdCS(N(NGNGMOI'“ —
OOCN(NOS'^-CNCO(NOO — — O — OsOOOSOst^-r^t^SO
hhh-hhh
soosco(Nm(Nr--ininosÔinincoOTpr-
ooosfNr-sor^r'-H*— ooTtiNcoiNOO — o
OihN(NNO\OS-<
m'tin'trJ-’t'tTfrO(SfS(NN(N(NNN(S COOOOSOnSOSOSOSO
8C£ j3§ — oo os co so m o co —< ro oo m o rj- co o O t}- (N fN OS so SO 00 00 CO

3 aJ
G'tCMOGLL
r- so in >n wo tT Tt
00(NONOsrt<Nro(NO
TfTj-rO(NCN(N(N(NtN
— o os o f" r-
fN N - N - - >n in in in
u >
osOcoTtt^-msoso — r-sou-i^ — o\m

cor^Tj-Tj-inTj-fOTtcoiNcoroo — OO
CN<N<N<N<NCN(N<N<N(N<N<N'NCN<N<N
h in so co os - insoosOmmiNTtoococoTj-cor^Tpo
ooosh'tcohoshsfin'oinl'infoco’ti'- - — —<
in'tcOfON(N(N(NrNNNNN(NCl(N(N(NN(SNN
•S
=\
OO <N os n in pi m -,osoooos£)Ttr-0<Nininor~coo'nO’--'ÔTp—< t-*

0O O oo(Nr^r'Oor~oosocooosoosr',cP'^-soTfcoincocoTtTp—i — — o
OO 00
3 co
SL- o3 oo^C'ir-'Oso-—<oor-soosr-'-oot'~ONOS’-*©
ooo—-ocoi^'ooooooost^'^toospoosp ^. ,
-ro<Ncoot'-r'-oo—'Or-~
JucOTf<N<NrSrlOSOOOs
3 ’rt oooor-sosO'n'nin'nTj- coco(N<N<NCN<N(N(N<N<N<N<N<N<N<N—<<N(N —
O >
(NrOTtmsOI^COONO-rC^co
rocococococococOTpM’MM
of the largest inequality (see Table 4.3); the second is the last prewar year
1940, the third is the first postwar year 1946. The effects are very small;
additional decimal places are needed to show the differences. It turns out
that, if we use the fixed population shares of 1932, the inequality among the
states decreases more than the observed pattern shows. This applies to a
period of more than ten years; after 1945 the situation changes and the fixed
pattern of the population distribution produces slightly higher inequality
values than those of the current population distribution. The between-set
and within-set components have the same development over time. Further¬
more, when we use the population shares of 1940, inequality decreases also
more rapidly until the end of the war. The use of the 1946 population shares
leads to slightly larger inequality values immediately (from 1947 onward). We
conclude that the effect of the shifts in the population distribution on the
inequality among the American states is very small and that, insofar as any
effect exists, it did not clearly work in the direction of more equality. Need¬
less to say, all these results refer to the case in which population shares are
supposed to take values different from those which are observed without
affecting the per capita incomes.
4.6. Migration and Its Effect on Per Capita Income Inequality;

Maxwell’s Demon on Ellis Island
The result just obtained may seem surprising. Isn't there a tendency to
migrate from poor states to richer states? Wouldn’t this lead to changes in
population shares which imply a lower inequality among states? The first
question can be answered in the affirmative. We refer to Figure 4.1, which
consists of two scatter diagrams, one corresponding to the decade 1930-40
and the other to the decade 1940-50. The variable measured vertically is
net migration (total immigration minus total emigration) during the decade,
expressed as a percentage of the average population of the state in the first
and the last year of the decade.1 The horizontal variable is state per capita
income in the middle year (1935 and 1945), measured according to a loga¬
rithmic scale with the national per capita income in the origin. The results
show that there is a clear positive association between net migration and the
relative income position of the state, particularly in the second decade. [The
1 That is, the average population of 1930 and 1940 for the first decade and that of 1940
and 1950 in the second. The figures on net migration are taken from U.S. Bureau of the
Census, Historical Statistics of the United States, Colonial Times to 1957 (Washington,
D.C., 1960), pp. 44-45.
Net migration (%)
1930-40 20
10
200 800
’Per capita income (S)
-10
1940-50 20-
10
800 2000
Per capita income (S)
x o®
°X ,
- -10 Population shares:

o Less than 1 per cent
x Between 1 and 2 per cent
• Larger than 2 per cent
Fig. 4.1. The relationship between net migration per state and state per capita income,
1930-40 and 1940-50. [States are indicated by different types of dots depending on their
share in the national population.]
figure shows also that the horizontal dispersion of the points is less in 1945
than in 1935, thus indicating that there is less inequality.]
The second question, which deals with the effect of migration on the
population shares and with the effect of these on inequality, is much more
complicated in nature. First of all, the effect of migration may be offset
by differential birth and death rates. But let us assume that there is no
such complication, so that migration from A to B raises the population
share of B at the expense of that of A. Then it is still not certain in which
direction inequality will move, even if the migration does not affect the
ratio of the two per capita incomes. To simplify the exposition we shall
confine ourselves to the two-country case; hence the analysis which follows
is a continuation of our considerations in Section 4.3 on the White-Nonwhite
inequality in relation to the inequality figures for American states.
Let country A contain a fraction p of the total population and country
B a fraction q=\— p. A is richer than B. We choose units such that the per
capita income of A is 1 and we write 0, O<0<1, for the per capita income
of B. The total income of A is therefore pN, where N is the total number of
persons in the two countries combined; that of B is qdN, and hence
p qd
(6.1) Vr —-
p + qO p + qO
are the income shares of A and B, respectively. The inequality of the two
countries is therefore
qd
p + qO qd p + qd
(6.2) / = log--+ log
p + qd p + q0 q
qd logd
log (p + q0)
p + qd
The partial derivative of / with respect to d is
16 31 dI _ pq loS0
dO (p + qd)2
which is negative (as it should be). The inequality decreases when the ratio
0 of B's per capita income to A’s increases, given fixed population shares.
The minimum / is zero (in the limit for 0 approaching 1), the maximum is
-logp (in the limit for 0 approaching 0). The latter result is in accordance
with what we found in Section 4.1 under (2).
4.6 MAXWELL’S DEMON ON ELLIS ISLAND 117
We proceed to consider the effect of population share changes under the

assumption of fixed per capita incomes, i.e., under the assumption of a fixed
9. It is easily seen that 7 = 0 both for p = 1, q = 0 and for p = 0, q = 1. When q
increases from 0 to 1, inequality rises first, then reaches a maximum and
decreases thereafter. To prove this we consider the partial derivative of 7
with respect to q (putting p = \ — q):
(6.4) d-r-l Z±+”!*»

dq p + qO (p + q6)
The first term on the right is always positive, the second is always negative.
The derivative vanishes for
l 9 log 0
(6.5) <? = i +
i - e we.
and the corresponding inequality value is
log 0 6 log0\
(6.6) Max / =
9 we
This must indeed be a maximum, because 7^0 for O^g^l, 7 = 0 for q — 0
and q=\, and there is only one solution (6.5) of dl/dq = 0. The maximizing q
values and the corresponding 7’s for some alternative 9 values are given in
Table 4.8. The results show that the maximizing q decreases from 1 to Jr and
the maximum inequality value decreases from co to 0 when 9 increases
from 0 to 1. Hence inequality is particularly high when the rich country has
a relatively small population.
Suppose now that, given an initial situation which is characterized by
the parameters p, q = l—p and 9, there is migration from poor B to rich A
by which the population share of B is reduced to q — dq and that of A is
raised to p + dq. [We confine ourselves to infinitesimal changes.] It will be
assumed that the migration does affect the per capita incomes. Those who
migrate had an average income of kB in the old country and they have an
average of kA in the new country; it is supposed that there is no effect on
the incomes of those who do not migrate. Hence the analysis of the previous
paragraph deals with the special case kA = 1, kB = 6. In the present more
general case the total income of country A becomes p + kAdq (which implies
a change in the per capita income unless kA = 1) and that of B becomes
qQ — kBdq (implying a per capita income change except when kB-9), apart
from the factor N, the total number of individuals in the two countries.
TABLE 4.8
SOME INEQUALITY MEASURES IN THE TWO-COUNTRY CASE
Maximum /
\-d
e q of (6.5) of (6.6)
log 8
in nits
0 1 00 0
.1 .827 .619 .391
.2 .747 .313 .497
.3 .691 .178 .581
.4 .649 .104 .655
.5 .614 .0597 .721
.6 .584 .0325 .783
.7 .559 .0159 .841
.8 .537 .00622 .896
.9 .518 .00139 .949
1 .5 0 1
The sum of these amounts is
P + qd + (kA — kB)dq
so that the income share of A becomes
p + kAdq
— JN + dq
p + q6 + (kA — kB)dq p + qd
and that of B:
y Bk A + v^/cg
yB - dq
p + qd
where yA and yB are the income shares before the migration as defined in
(6.1). On combining the new income shares with the population shares p + dq,
q — dq we find after some algebra that the new inequality value exceeds the
old value (6.2) by
(6.7) <//=-[! -0 + (ytkA + >■>„) log

p + qO
This agrees with (6.4) in the case kA = 1, kB = 0. [Note that q is not raised
but reduced by dq\ hence the minus sign after the equality sign.]
The question to be considered is: What is the condition under which the
migration from poor B to rich A reduces the inequality between the two
4.6
MAXWELL’S DEMON ON ELLIS ISLAND
119
(<//<0)? This amounts to a positive sign of the

expression in brackets in
(6.7), and hence to
(6.8)
On the lett we have a weighted average of the migrants’ earnings before and
after the migration. The weights are the income shares of the other country
(.Lb for kA, yA for kB). Now this weighted average should be less than the
positive value on the right of (6.8) in order that there be less inequality.
This right-hand side is. tabulated in the last column of Table 4.8. It is
approximately equal to the geometric average yJ{ly.Q)=yje of the two per
capita incomes.
Clearly, when the migrants are sufficiently poor, both in the old country
and after their arrival in the new country (kA and kB both rather close to
zero), condition (6.8) is satisfied. When they are sufficiently well-to-do,
either before they emigrated or thereafter or both, the inequality between
the countries will increase rather than decrease. These results are intuitively
plausible. The first case is the typical poor Negro case. His income was
relatively low in the South before he moved to New York {kB low) and it is
also relatively low by New York standards after he arrives there (kA low).
His migration raises the per capita income of his former state and lowers the
New York per capita income; it thus contributes to less inequality between
the two states. The second case is that of a wealthy German Jew who emi¬
grated to the United States in the nineteen-thirties and whose standard of
living was very modest in the new country. We then have a large kB and a
small or moderate kA. It is not sure whether this leads to a larger German-
U.S. inequality as long as the algebraic symbols of (6.8) are not replaced by
numerical values, but it is certainly conceivable that this migration raises
the inequality level due to the reduction of the German per capita income.
The third case is that of a poor Indian scholar who is attracted by a well¬
paying American university. Here we have a large kA but not a large kB. The
fourth is the case of a successful businessman who leaves Alabama to go
to sunny California. His kA and kB are both large and his move will certainly
raise the Alabama-California inequality.
It is instructive to consider the migration effect on income inequality
from the standpoint of Maxwell’s demon.1 This animal was born in 1871 and
1 The lines which follow are quoted from L. Brillouin, Science and Information Theory
(2d ed.; New York: Academic Press, 1962), p. 162.
first appeared in Maxwell’s Theory of Heat (p. 328) as “a being whose fac¬
ulties are so sharpened that he can follow every molecule in his course, and
would be able to do what is at present impossible to us ... . Let us suppose
that a vessel is divided into two portions, A and B by a division in which
there is a small hole, and that a being who can see the individual molecules
opens and closes this hole, so as to allow only the swifter molecules to pass
from A to B, and only the slower ones to pass from B to A. He will, thus,
without expenditure of work raise the temperature of B and lower that of A,
in contradiction to the second law of thermodynamics.” This second law
states that there is an inherent tendency of the entropy to increase (see the
end of Section 2.1), and our demon manages to reduce it.
There is no need to continue this physical exposition. [It was argued in
the later literature that the demon cannot really do what he is described as
doing.] Instead, we suppose that A is the United States, B the rest of the
world, and that there is a demon on Ellis Island where the migration from
B to A is controlled. For each potential immigrant the demon collects data
on recent earnings in B. Moreover, his faculties are so sharpened that he
can predict perfectly what each potential immigrant would earn in A. Thus,
by admitting and rejecting immigrants he can influence kA and kB. We shall
consider three different attitudes. One is that of an alpha demon, who favors
a high per capita income level in A and does not bother about B. He prefers
a high kA value to a lower kA and is indifferent about kB. We may expect
that this works in the direction of a large value of the left-hand side of (6.8).
[This is not necessarily true, since a large kA may be compensated by a small
kB, but that is probably not the typical case.] The beta demon favors a high
per capita income in B and is not interested in A. His preference is for low
kB values. He is therefore more likely to reduce the inequality between A and
B than an alpha demon is. The omega demon is not interested in either of the
countries in particular. He favors the economic well-being of the individual
migrants by admitting everyone whose income in the new country will
exceed what he earned in B. That is, he admits if and only if kA>kB. Since
the change in world income (per capita) is (kA — kB)dq, we may also say that
the omega demon maximizes world income. Our question is: Does this
demon’s marginal immigrant (kA&kB = k, say) contribute to a reduction of
inequality? Answer: That depends on k and 9, no longer on the income
shares yA and yB, because the left-hand side of (6.8) is now simply k. Since
the right-hand side is approximately J9 (see Table 4.8), we conclude that
this marginal immigrant reduces inequality if and only if his income is less
than the geometric average of the per capita incomes of A and B.
4.A. Notes on Some Alternative Measures of Income Inequality
This section consists of a brief discussion of four alternative measures of

income inequality. The first is Gini’s concentration ratio. We consider n
groups of individuals and write xt for the population share of the zth group
and 7; for its income share. Then Gini’s concentration ratio may be defined as
hi _ T/;
(A.l) G* = i i=u=i
Z Z ixNj ~ xjyi\ = iZ Z xixj
ij=i xt xj\
Hence G* is one-half of a weighted average of all absolute differences be¬

tween the deflated per capita incomes, the weights being the products of the
corresponding population shares. The minimum value is obviously zero,
which is attained when the per capita incomes are all equal. If one group,
say the first, earns all the income and hence the other groups nothing at all,
so that j>i = l, yj = 0 for j> 1, we have G* = l — x1. This approaches 1 when
the population share of the first group becomes smaller and smaller.
The concentration ratio can also be defined geometrically by means of the
Lorenz curve. Such a curve (consisting of upward sloping straight line
segments) is drawn in Figure 4.2 for the case n — 4. It is based on an arrange¬
ment of the n groups according to increasing per capita income values:
>r . yi .
-
. yn
... $5
(A. 2)
X, x„
When there is complete equality (yjxi independent of i), the Lorenz curve
coincides with the diagonal which cuts the square into two equal parts. We
shall now show that the concentration ratio is equal to twice the area between
this diagonal and the Lorenz curve. For this purpose we note that the area
below the curve consists of n triangles, the total area of which is
(A. 3) i i
£=1
121
and of \n{n— 1) rectangles with the following combined area:
(A.4) E E xtyj = i E xt(E yj +1 - E yj)

i=1ji
This is the area below the Lorenz curve. We find the area between the curve
and the diagonal by subtracting the expression from \ (the total area below
the diagonal):
(A-5) 2 E *j(e yj- E yj)

It should be noted that this expression is only seemingly asymmetric in the
x’s and y’s. The total area of the rectangles below the Lorenz curve can also
be written as „ „ ,
E E ypj = i E (E *j + 1 - E xj
i = 1 j>i i=1 \j>i jî
4.A ALTERNATIVE MEASURES OF INCOME INEQUALITY 123
When this expression is added to (A.3), we find for the area below the curve
so that the area between the curve and the diagonal becomes
This is indeed symmetric to (A.5). [Note the inversion of j i in
(A.6) compared with (A.5). This is due to the fact that the ranking according
to increasing values of V;/jc; is the same as that according to decreasing
values of xjy^]
We must now show that G* as defined in (A.l) is equal to twice the ex¬
pression (A.5) or (A.6); or equivalently, to the sum of these expressions.
Consider then:
n n
=i I I Ott;
I I (xjy - xiyj) + i x Z (x>yj - xjyd
which is indeed equal to the sum of (A.5) and (A.6).

The concentration ratio is rather popular, but it has the serious drawback
that it does not lend itself to between-set and within-set decompositions.
This is due to the use of absolute differences, which are awkward for such
purposes. An inequality measure which is superior in this respect is the
variance of the logarithms of the incomes. This is the second coefficient to be
considered. Let us write z; for the income of the zth individual, Ng for the
number of individuals of set Sg, g = 1, ... , G, and N=ZNg for the total
number of individuals. Consider the geometric mean income Z of the set
of all individuals and Zg of set Sg:
N
(A.7) £= 1
0 = 1,..., G
i eSg
which are connected by
(A.8) logZ = logZ,

N
9=1
Next we consider the variance of the income logarithms:
N N
1 ^ 1 V
^ (Iogz,-log Z.f = NL, log
1=1 i= 1
which can be decomposed as follows:

G
iy y [(log*, - iogz,)+(iogza - iogz)]2

9 = 1 ieSg
■im
9 1
= ieSg
l°4)+lwkl 9=1 >eS9
log
Z;
+2y*»i log^ log
/1 N N„
9= 1 ieSq
This can be simplified to
^9
(a-9)
»=1 9=1
+
9= 1
N iI(iogzX
ieSg
The first right-hand term is the between-set logarithmic income variance,

the second is a weighted average of the within-set variances, the weights
being equal to the population shares N^/N.
The variance decomposition (A.9) is elegant and straightforward; it is in
particular useful when income is approximately lognormally distributed. Its
disadvantage is that the variance of the income logarithms is the second-
moment extension of the geometric mean of the individual incomes. The
arithmetic mean(thepercapitaincome)ismoreconvenient,becauseitis directly
related to total (national) income. The information measure is preferable in
this respect, because it deals with (deflated) per capita income.
The third inequality measure to be considered is the squared coefficient of
4. A ALTERNATIVE MEASURES OF INCOME INEQUALITY 125
variation. It is not based on geometric means but on arithmetic means:

N
(A. 10)
ieSa
which is an advantage compared with the logarithmic variance; but its

decomposition has certain disadvantages as we shall shortly see. The variance
decomposition is
N G G
-\2
(Zi ~ Z)
A (4 - + (zi ~ Z,Y
i— 1 9 = 1
so that we obtain for the square of the coefficient of variation:
1 N g jy
N i%
A = i (Z‘-
A z2
ly9 9 E
Ag ieS,
(z<- - z«)2
(A. 11) + Nz2
-2
z2 9= 1
The first term on the right is the square of the between-set coefficient of
variation, the second is a weighted sum of the squared within-set coefficients
of variation. It is not a weighted average, because the weights do not add up
to 1. The sum of the weights is
A, A, G N
=\2
?2+.?a(^)2
(A. 12) = 1 +
z-2 z2
In words, the sum of the weights of the within-set terms is at least equal to 1,
and the excess of this sum over the unit value is precisely equal to the square
of the between-set coefficient of variation. In other words, these weights are
on the average larger when the between-set term is larger. This is a disad¬
vantage, because one should prefer a measure for which the within-set compo¬
nents, including their weights, are independent of the between-set component.
The fourth and last inequality measure to be discussed is I(x:y), the
expected information content of the indirect message which transforms the
income shares as prior probabilities into the population shares as posterior
probabilities. This is analogous to the measure discussed in the text, I (y:x).
except that the roles of the population shares and the income shares are
interchanged. The formula is
n X•
(A. 13) I(x--y)=
i= 1 yi
which implies
(A. 14)
Hence I(x:y) is the logarithm of the reciprocal of a weighted geometric

mean of deflated per capita incomes, the weights being the population shares.
These weights are less than the weights yt of (3.4) in the case of sets with a
large per capita income, and larger than yt for sets with small per capita
incomes. The weighted geometric mean in the denominator of (A.14) is
therefore <1, contrary to that of (3.4) which is ^1.
The decomposition of I(x\y) is as follows:
(A.15) I(x:y) = I0(x:y) + X Xglg{x:y)

0= 1
where
x„
h (x '■ y) — } xgiog
9=i
(A. 16) T( ,
Ig(X-y) = )
V Xi
“log
Xn " yJYg
XilX< g = 1, ...,G
i e Sq
x,= T.x, y,= 1 y, g = l ,-,G

ieSa i £ Sg
This is largely analogous to the decomposition (3.7), but there is one im¬
portant difference: The weight of the within-set inequality measure Ig(x\y) is
the population share Xg, not the income share Yg as in the case of Ig(y:x) in
(3.7). We recall in this connection that the total within-set inequality of the
54 countries is largely dominated by the inequality within North America
and Northwest Europe, because that area’s income share is so large (see the
end of Section 4.4). If one dislikes this feature on the ground that the within-
set inequality of the other areas should have a larger weight, one may prefer
(A. 15) to (3.7) and hence use the income shares as prior and the population
shares as posterior probabilities. This depends on whether one is “income
4.A ALTERNATIVE MEASURES OF INCOME INEQUALITY 127
oriented” or “population oriented” with respect to the weighting of the

within-set inequalities. It should be noted, however, that the simple intuitive
considerations which led to I(y:x) as an inequality measure do not apply to
I(x:y). Our starting point in Section 4.1 was the entropy H(y) of the income
shares of A individuals, which we then subtracted from the maximum
entropy, log A, to obtain a measure of income inequality. In the case of
I(x:y) we have to take the income shares yt, ..., yN as prior probabilities
and transform them to the population shares 1/A, ... , 1/A as posterior
probabilities. The simple entropy concept does not work; we have to apply
the indirect message idea immediately. Also, one may argue that it is against
intuition not to take the equal population shares as a starting point but to
consider them as “posterior.” But this is a matter of taste.
The formulas to which I(x:y) lead are certainly not more complicated
than those of I(y:x). Let us return to the second paragraph of Section 4.2
and write z for the income of any individual and m for the average income of
all A individuals. Then this individual’s y, is z/Am and his x; is 1/A. Hence:
x; 1 m
x; log - = - log -
k; N z
When/( ) is the density function of the income distribution, so that Nf(z)dz

is the number of individuals of the set S whose income is at most\dz from
z, we have
X. / I7l\
Z x; log' - = log- )f(z)dz
ieS ki V Z/
Hence the inequality measure I(x:y) can be formulated in terms of the

continuous income distribution in the following way:
(A. 17) ^log f)f(z)dz = <^log - j = log <Sz — £ (logz)

0
When the distribution is lognormal with logarithmic variance o2, (A. 17)
reduces to |er2. This is equal to the value of I(y:x) as derived in Section 4.2.
For the Pareto distribution with parameter a we find
a 1
(A.18) log
a a
which differs from the corresponding value (2.6). The derivative of (A.18)
with respect to oc is — 1/a2 (a — 1), which is negative as in the case of (2.6).
We also refer to the Appendix of Chapter 8, which deals with the Herfin¬
dahl index of industrial concentration. Since concentration and inequality
are essentially the same concepts, this index may in principle also be used as
a measure of income inequality.
4.B. Details on White and Nonwhite Family Incomes
The data on White and Nonwhite family incomes are presented in Tables
4.9 through 4.11 with the exception of the 1963 data, which are presented in
Table 4.1 in the text. It should be noted that the incomes of so-called
unattached individuals are excluded. For details we refer to the Current
Population Reports of the Bureau of the Census.
The percentages do not add up to 100 exactly in a number of cases due to
rounding errors. This was corrected by dividing the percentages by their sum
before the calculation of averages and inequality values. The income distri¬
bution for all families (color disregarded) is computed from the corre¬
sponding pair of White and Nonwhite distributions by weighting their per¬
centages by means of the number of White and Nonwhite families. The total
number of families is available for each year; the two separate numbers are
not available in a number of years.1 In such cases the population shares have
been interpolated linearly. The interpolation errors are of minor importance,
since the percentage of Nonwhite families increased very gradually from
about 8^ in 1947 to about 10 in 1963.
4.C. Limits on Inequality when Income Brackets are Used
The objective of this section is to derive upper and lower limits to the
inequality measure I(y:x) for the case in which the available data are in the
form of numbers (or percentages) of income recipients whose incomes are
located in a finite number of intervals. We shall not obtain a complete
success in this respect due to the fact that the interval containing the largest
incomes is usually an open interval. The procedure that will be described
has therefore an approximate character.
Our starting point is the decomposition (1.9), which will now be inter¬
preted such that Sg is the set of income recipients whose incomes are in the
gth interval, g = 1, ..., G. The procedure followed in Table 4.2 amounts to
neglecting the within-set differences, since all incomes of the gth interval,
g-1, •••, G, are supposed to coincide with the midpoint of that interval.
1 These are the years 1949 through 1952, for which the successive total numbers of
families are: 39,193; 39,822; 40,442; 41,020 (in thousands).
4.C
TABLE 4.9
INCOME DISTRIBUTIONS OF WHITE AND NON WHITE FAMILIES IN THE UNITED STATES, 1947—1951
LIMITS ON INEQUALITY
129
<N
(N
CO o 00 o ON 00 on CO (N o o <0
00 00 <N <N
00 of NO of
d ON OO o r- r- r- on on
r-
»/n
Os
(N
ON
,-H NO (N on o of O r-H ON of 1—H i/n un NO^
on r~ on
r4 oi CO of of of on NO r- r" on o ON
CO
r- of
ON
»/n
ON 00 of Of NO NO CO on ON on ON —- O ON
NO NO
o of of CO
o On ON 00 OO NO r- CO
r-H
r-l
NO
UO
Os
*/n
O O o Of NO NO of
H CO NO On NO r- CO <N oh CO
o ON
< (N <N CO of of of NO NO OO r- oh NO NO
H
c/3
CO
<3
Qw
H
Z 1
O
5 NO of OO ON on i-H (N O ''t r- OO OO NO o o ON
U3 — OO On OO o r- NO of- on Of CO CO
ffi un T—'
H >/n
ON
o
of
o NO <N on i-H CO — <N of NO oh ON ON CO O »n °\
co CO of of on on r^* 00 CO ON CO on r-H 00
i—t 1 Q> CO
< X.
Ph
W o
W g ^f
5 o
NO
m
< zoi ON
ON
OO
r—
(N
i—i
CO
o
CO
OO
00
OO o
—H
00
on
00
or
r-
oh
of
CO
r-H
CO
OO -h O §
CO
of
z ITS
ON
Q O
Z r-
< T—' on O o <N 1 CO On of NO O ON 00 1-H Tf
w r-~ r- <N T—, OO
H Of CO on on on NO OO ON Of
CO
"
u-
O CO r- CO O ON on r- of CO of on NO OO un
00 T—1 T—(
z NO d <N <N <N <N NO CO <N oh <N
O
H (N
to
P
m ON
2 r-
H OO of ON OO on r- OO 00 ON O
CO CO of of on NO ON ON 00 r- <N OO ON CO i-H
'—
w
S
o
u
z
o o o o O O O O o O o o o o o
o o o o o o O O o O o o o o o
<D w o on O on o on O on o o o o o o
(N <N cn co of of on NO r- o on on
| g 1 1 1 1 1 1
iii O O ct O o o o o
o o o O o O o o O o o o O'
on o on o on o on o o o o o o
(N CO CO of of on NO o on on
<N
4.C
LIMITS ON INEQUALITY 131
132 1111 MI ASURI MI NT OF INOOMF 1N1QI AL1TV
Hence the figures of Table 4.2 are between-set inequality values (sets in the
sense of income intervals); the true inequality values cannot be lower than
these, because the wit h in-set inequalities which are neglected are nonnegative.
Therefore, the figures of that table are lower limits to the true inequalitv
values.
Consider then the within-set inequality for an income interval (a, b). Let
the average income of this set be c. which of course satisfies
(C.l)
In our numerical applications we shall take
(C.2) c = 4(« + b)
but this is appropriate when the average income is actually given. This is not
the ease tor the data of Section 4.2, so that we shall confine ourselves to the
approximation (C.2).
One extreme is the case in which all incomes of this interval coincide with
c (the case of Table 4.2); the other extreme is the case in which all incomes are
either equal to a or to b. I he position of c within the interval (a. b) determines
the tractions of those w hose incomes are equal to these respective limits. It is
easily seen that these fractions are
b — c c — a
(earning a) (earning b)
b — ti b — a
l lit sc ate the two population shares. The corresponding income shares are
(b — c)a (c - «) b
and
(b - <i)c (b — a)c
so that the within-set inequality is
(.b — c) a a (c — ti) b b
log - + log
(/’— a)c c (b — a)c " c
1 his expression can be w ritten in a slightly more elegant form if we write
(c^ a = c(l — fij) b = c( 1 + S2)
1 he result is that the within-set inequalitv is
/C4x faO -'\Vtog(l -<0 + ^(1 + S2) log (1 -d.)

S| — Pj
4.C LIMITS ON INEQUALITY 133
which reduces to
(C.5) 4 Jog (1 -S2)+ [6 log ' 4f> St=52 = 5
in the special case when the mean c coincides with the midpoint of (a, h), so
that the two <5’s arc equal. When the relative difference between a and h is
small, the expressions (C.4) and (C.5) are approximately equal to {<5, <52 and
4 , respectively, third and higher-order terms being disregarded.
I he within-set inequalities thus obtained arc to be weighted, in accordance
with the decomposition (1.9), by means of the set income shares. A difficulty
arises, however, with respect to the open interval {a, oo) containing the larg¬
est incomes. Write again c for the average income of that interval and
imagine that a fraction p has an income equal to a and a fraction I —p an
income equal to
(C.6) C ” Pa
1 ~P
[It is easily verified that, for any p, the average income is indeed c.] The
income shares arc then pa/c and (c—pa)/c, so that the inequality value
becomes
pa a c — pa c — pa
— log + log
c c c 0 ~ P)c
pa c - pa c - pa 1
log c + log a + \og (c-pa) + log
c c c I — p
Suppose now that p converges towards 1, so that almost everybody’s income

is a and very few have an extremely large income (C.6). Then the first right-
hand term remains constant, the second and third converge to
a c — a
log a and log (c — a)
c c
respectively, and the fourth increases beyond bounds. Hence there is no

finite upper limit to the inequality in the open class.1
1 Strictly speaking, one coufd argue that the largest value which p can take is one minus
the reciprocal of the number of income recipients of the open class. The corresponding
inequality value is then approximately (1 ale) log N, where N is the number just men¬
tioned. The use of this value would make the total inequality value excessively large, par¬
ticularly also because the weight of the open class in total inequality is not very small.
TABLE 4.12
APPROXIMATE LOWER AND UPPER LIMITS TO THE INEQUALITY FIGURES OF TABLE 4.2
Year Total White Nonwhite
1947 .260 ±.011 .247 ±.011 .297 ±.011

1948 .256 ±.012 .243 ±.012 .291 ±.011
1949 .259 ± .009 .245 ± .009 .299 ±.012
1950 .259 ±.010 .246 ±.010 .292 ±.010
1951 .244 ± .007 .230 ± .007 .286 ± .009
1952 .265 ± .007 .248 ± .007 .336 ±.007
1954 .262 ± .007 .250 ± .007 .288 ± .007
1955 .248 ±.007 .236 ± .007 .258 ± .006
1956 .246 ± .007 .234 ±.008 .269 ± .006
1957 .233 ± .007 .220 ±.007 .276 ± .006
1958 .241 ± .008 .227 ± .008 .297 ± .006
1959 .247 ± .007 .232 ± .007 .293 ± .005
1960 .253 ± .008 .240 ± .008 .296 ± .005
1961 .264 ± .008 .252 ± .009 .316 ±.006
1962 .246 ± .008 .232 ± .008 .278 ± .005
1963 .233 ± .007 .220 ± .007 .273 ± .007
To solve this difficulty we shall approximate the distribution above the

lower limit of the open class (a, oo) by a Pareto distribution, using (2.6) and
the a values mentioned in footnote 1 on page 99.1 The total within-set
inequality is then obtained by weighting the separate terms with the income
shares of the sets. The results (in nits) are shown in Table 4.12 in the form
A±B, where A — B is the lower limit (which was already given in Table 4.2)
and A + B the approximate upper limit. The figures indicate that if we use
the midpoint A of such an interval, the error is in most cases of the order of
1 per cent of a nit.
1 For the group of all families (color disregarded) a is defined such that a/(a — 1) is equal
to the ratio of the average income above the lower limit of the open class to that limit.
This average income, in turn, is the weighted average of the corresponding White and
Nonwhite averages with weights equal to the respective population shares.
CHAPTER 5
A STATISTICAL APPROACH TO THE PROBLEM
OF PRICE AND QUANTITY COMPARISONS
5.1. Random Selection of Commodities
In this chapter we shall be concerned with expenditure groups such as

consumption, output, exports, and imports. The group as a whole is hetero¬
geneous. It consists of n commodities, each of which is considered to be
homogeneous. When we multiply price and quantity for any of the n com¬
modities we obtain the expenditure on that commodity, after which total
expenditure on all n commodities is found by summation. The main problem
that will be considered in this chapter is classical and dates back to the
nineteenth century. It can be briefly described as follows: Suppose we have
price and quantity data for the individual commodities in two different
regions or in two different periods; can we then argue in any meaningful way
that the price level of the second region (or period) is, say, 10 per cent higher
than that of the first? Can we do the same thing for the quantities? Note that
both questions refer to averages: The prices of the second region (or period)
exceed those of the first, on the average, by a certain percentage (similarly for
quantities). We may also be interested in the variation of individual price and
quantity ratios around such averages. For some commodities the price in the
first region may exceed the price in the second by very much more than the
average indicates; for other commodities the converse may be true. A Ford
is cheap in the United States compared with its price in India, but household
assistance is comparatively much cheaper in India. Such dispersion problems
will also be considered in this chapter.
The index number problem can be approached from the statistical angle,
but one can also take economic theory as a starting point.1 In the present
chapter we shall adopt the statistical viewpoint; the other approach will be
considered in the next two chapters. To make things specific, let us consider
1 Reference should be made to Ragnar Frisch’s classical article, “Annual Survey of

General Economic Theory: The Problem of Index Numbers,” Econometrica, Vol. 4 (1936),
pp. 1-38.
135
136 PRICE AND QUANTITY COMPARISONS
two regions (rather than periods), to be denoted by a and b. We shall write

pia for the price of the ith commodity, i= 1, in the ath region, qia for the
corresponding quantity, and
n
(1.1) nia = Z1
i—
PiaQia
for the total expenditure in that region. Furthermore, we introduce the share
of each commodity in total expenditure:
PiaPia
(1.2) i = 1, ...,n
m„
which will be called the ith value share in the ath region. These value shares will
play a very prominent role in the analysis which follows. If it is assumed - as
we shall do - that prices and quantities are all positive, these shares add up
to 1 and are also positive:
(1.3) Zwia=l wia > 0 i = 1,..., n

;= i
We mentioned in the first paragraph that our main problem is: Can we
argue in any meaningful way that the price level of the ath region exceeds that
of the bih by, for example, 10 per cent? We are thus interested in relative
differences, so that the logarithms of the prices pia,pib, i= 1, ..., », are the
obvious basic ingredients for the procedure that will now be described. For
each individual commodity the relative price difference is then measured by
Pia
(1.4) log pia - log pib = log
Pib
which will differ for different values of i. Our statistical viewpoint amounts to
the following. Suppose we are in the ath region; suppose that we draw the n
commodities at random in such a way that each dollar (or franc or guilder...)
of total expenditure has an equal chance of being selected. Then the chance
that we shall draw the /'th commodity is the value share vvia as defined in (1.2).
Hence wia is the probability of finding the logarithmic price difference (1.4).
This leads to the following average of all n logarithmic price differences:
(1-5) Z wia log ^

;=i Pw
We take the antilogarithm and conclude that
(1.6) n Pib.
5 1 RANDOM SELECTION OF COMMODITIES 137
is the price index number which is implied by this procedure. It is a weighted

geometric mean of the n individual price ratios, the weights being the value
shares of the ath region.
The price index (1.6) has certain merits. It is, for example, independent of
the units in which we measure the quantities of the various commodities
(tons, gallons, etc.). It has the disadvantage, however, of being one-sided in
the sense that it is based on the distribution of expenditure in the ath region.
We could equally well apply our random selection procedure to the bth region,
in which case wia is replaced by wib in (1.5) and (1.6). We must conclude that
(1.6) is an asymmetric index number, which is a disadvantage because the
question asked is implicitly symmetric: If the price level of the bth region
exceeds that of the ath by a factor 1.2, say, we should expect that the price
level of the latter region exceeds that of the former by a factor j 2.
To solve this difficulty we proceed again along statistical lines. Suppose
you live in Denver, Colorado, and want to make a price comparison between
New York (a) and San Francisco (b). Before selecting dollars at random in
New York or in San Francisco you should first make up your mind as to the
question of where to do this. Given our statistical viewpoint, you make this
choice at random. If x is the probability that you will do the job in New York,
and hence 1 — x' the chance that you will go to San Francisco, the probability
that you will draw a dollar which is spent on the ith commodity is
xwia + (1 - x)wib
Given that the question asked is symmetric, an obvious choice is x = ^, so

that the probability then becomes
Wia + W:ib
(1.7) Wiab = i = 1,.... w
In words: the average of the two values taken by the ith value share in the two
regions. Using these probabilities we obtain the following average of the
n logarithmic price differences:
n
Via
(1.8) nab = £ wiab log
i= 1 Pib
which corresponds to the price index

n Wib
(1.9) enab n
Hence the price index whose logarithm is nab is the (unweighted) geometric
mean of two weighted geometric means of price ratios. One of the latter
means uses the value shares of the ath region as weights, the other uses those
of the bth region. Note that the e before the first equality sign in (1.9) indicates
that our logarithms are natural logarithms. This applies to all logarithms
dealing with prices, quantities, expenditures, and value shares, both in this
chapter and in the two which follow, which implies inter alia that we shall
continue measuring information in nits, as we did in the previous chapter.
The price index number defined in (1.8) and (1.9) uses the n individual
logarithmic price differences as the basic ingredients. They are combined
linearly by means of a two-stage random selection procedure: First, we give
each region the same chance \ of being selected, and second, we give each
dollar spent in the selected region the same chance (l/m0 or 1 /mb) of being
drawn.1 The simple linear structure of nab will be of major help when we
extend the approach to dispersion problems in Section 5.5. It is fairly obvious
to apply the same structure to the quantities. This leads to a volume or
quantity index number which is the antilog of
C1-10) Kab = Z wiab log Cha

i= 1 Q ib
Formulas (1.8) and (1.10) contain the logarithmic index numbers with which
we shall work throughout this chapter.2
5.2. Coal Miners in the European Economic Community; The Factor Reversal
Test and the Allocation Discrepancy
We shall now apply the logarithmic price and quantity index numbers to
make pairwise (binary) comparisons for consumption per family by coal
miner families in six regions:3
1 An alternative interpretation is in terms of a third region in which the value shares are
all halfway between the corresponding shares of the first two. It will turn out in Section 5.6
that this interpretation has its merits in aggregation analysis.
2 We cannot claim any originality for these index numbers. As a matter of fact, it is very
difficult to invent new ones since Irving Fisher in The Making of Index Numbers (Boston
and New York: Houghton Mifflin Company, 1922) went systematically through all kinds
of averages (arithmetic, geometric, harmonic, ...) with all possible kinds of weights. The
price index (1.9) appears as No. 123 in his list on p. 473. There are fourteen index numbers
which he regards as superior to this one. Leo Tornqvist was a stronger advocate of this
index, which he derived from Divisia’s continuous approach. He even computed loga¬
rithmic price variances of the type that will be considered in Section 5.5. See “The Bank
of Finland’s Consumption Price Index,” Bank of Finland Monthly Bulletin, No. 10 (1936)
pp. 1-8.
The empirical analysis is based on T. Kloek and H. Theil, “International Comparisons
of Prices and Quantities Consumed,” Econometrica, Vol. 33 (1965), pp. 535-556.
5.2 COAL MINERS IN THE E.E.C. 139
- Federal Republic of Germany excluding Saarland (to be abbreviated as

Germany)
- Belgium
- France
- Italy
- Netherlands
- Saarland.
The last region - Saarland - was part of Germany before and during the
Second World War, after which it was occupied by the French army. In the
period 1947-1956 it was politically autonomous to some extent, but closely
tied to France in the economic sphere. On January 1, 1957 it was politically
re-united with the Federal Republic of Germany, but its economic integration
was postponed until the summer of 1959. Since the data to be used refer to the
period 1956-1958, this explains why Saarland is regarded as a separate region.
The data just mentioned are taken from a price survey and a budget survey.
The price survey was organized by the Statistical Office of the European
Economic Community in 1958, partly in cooperation with the national
statistical offices, and covers 178 commodities. Three countries were divided
into a number of sub-regions according to the areas where coal miners live.
These areas are indicated in Table 5.1 for all six regions except tiny Saarland,
TABLE 5.1
SUB-REGIONS AND THEIR WEIGHTS WITHIN THEIR REGION
Sub-region Weight Sub-region Weight
Germany France
Aix-la-Chapelle (Aachen) .1 North (near Lille) .6

Ruhr area .9 Lorraine (near the Saarland
border) .2
Belgium
Upper Loire Valley
Borinage (near Mons) .1 (near Saint Etienne) .1
Charleroi .3 Cevennes .1
Center (near La Louviere) .1
Italy
Liege .2
Campine (Northeast of the Sulcis (Isle of Sardinia) 1
country) .3
Netherlands
South Limburg (Southeast of

the country) 1
c
c$
X
'O
d>
S
COMMODITY GROUPS AND VALUE SHARES
<L>
U*
d
(N
o
c
X
X
o
X
£
d
X
uo
<L>
J3
o3
>
I
O
X
•o
a
<D
X
3bfi
X
I
together with their weights within the total region, which measure the
proportion of coal miners living in the sub-region. These weights are used to
convert the sub-regional prices of the survey to regional prices. It is interesting
to observe from the table that all coal miners except the Italians and a minor¬
ity of the French live within an equilateral triangle with sides of only 200
miles and angle points in or near Dortmund, Lille, and Saarbriicken. This is a
rather small area, but it is divided by several national and linguistic borders.
The budget survey was organized by the High Authority of the European
Coal and Steel Community in the period May 1956-May 1957 among coal
miner families with two children below 14 years. The value shares obtained
from that survey have been taken as a starting point for the computation of
the q’s and the w’s, given the p’s derived from the price survey. However,
many of the value shares are quite small, and they are only specified in units
of one tenth of one per cent. Therefore, the 178 commodities have been
aggregated to 21 commodity groups. A list of these groups is given in Table
5.2, together with their value shares in each region. The aggregation is such
that each group has a share of at least 2 per cent in at least four of the six
regions. We refer to the Appendix of this chapter (Section 5.A) for further
details, in particular for the logarithmic price and quantity differences of the
21 commodity groups.
Our basic ingredients are thus the logarithmic price and quantity differences
, Pia
log —
,log Qia
—
Pib ib
where /= 1, ...,«( = 21, the number of commodity groups) and a, b = l, ...,N

(= 6, the number of regions). Following (1.8) and (1.10), we weight these loga¬
rithmic differences by the average wiab of wia and wib, the latter two value
shares being given in Table 5.2. The results are shown on the first twelve lines
of Table 5.3 in the form of two skew-symmetric matrices, one for prices (nab)
and one for quantities (icab). The prices pla,pib are all expressed in local
currency in the original publication, but for the purpose of the present
analysis they have been converted according to the official rates of exchange.
This simplifies a comparison of purchasing power parities and the official ex¬
change rates. Actually, the choice of the exchange rates makes no real
difference. Going back to (1.9) we see that, when all prices of the a,h region
are multiplied by k>0, the price index enab which compares that region with
the bth is raised by the same factor. The effect on the logarithmic index nab is
additive, the new value being equal to the old value plus log k. The quantity
index is not affected at all, see (1.10).
g m m m os G m T-H m O »n <N Os vo G O m G"

jg o i-H m q OS ov <N <N VO G 00 Os q q g T-H O ^ VO
g
G
GO
i-H
i-H
<N <N ni o o
.1—1 1 i—i m
G1
1
o rn o
i—i 1 fN
o 1—1 <n r~i «n <N Os
i—i 1 1 m <N r r f r f1
i i 1 i i
*0
g
jg m rn <N m os <N VO m o o G »n G vo <N G G rn
(N q OS in Os q vo q Ov G q «n °o G O *—i G <N
CD o q in GOO vd o q q o q q q rn q
q (N i-H <N i—i m i-H i r* i—i (N <N i
1 i
<D
£
53
j>% oo in
rn (N G;
m
q q
in
G;
r^- oo
o (N
o m
Os vo e
in oo 00
q vo q
G" Os
in
co O ^ m
s
BINARY COMPARISONS OF PRICE AND QUANTITY LEVELS
O VO (N (N
s
G- i-H Os O G1 <N
T-H i—t
q
T—1
o q o q o
m <N
s vd o vd o q q
m <N m
PS
Vj ro
11 A 00 St
C3 o Q
PS
Vi .cn
.VJ ■*— "3
3 St
rN m m <N m Ol r- Os oo m G’ Os oo in _( ■*-«» m r- m
00 nj G; On in G; o <N q cs <N q Os oo q C3 o — (N r-
vd Os O Os q q q q o q q q o q O vd q q
| (N <— i—i **«* <N (N
iI 1l ii
i i rs 7 1 l
6 in m y—1 m m G- Os r- vo m m OS oo G* (N
g oo VO r- O rt O
q (N o Os o o vo (N q q vo m Os T-H ^ VO i^ O
j3) q O Os i—i q q vd o q o O d q o q d q q 1
i
i-H i—i i-H i-H m T—( i—i 1 m (N 1 I I I
I I I I I
Note. All figures are to be multiplied by 10~2.
G (N 00 in m in <N G*
c3 G* m G in in vo «n oo <N G
oo cn <N o Ov G; G q ON Os (N q q 00 oOO
S vd G* o q o vd q q vd G* o q d vd q q
-H
0) 1 <N ,_h i-H i^ H I I

o
C/5
TJ T5 *G
>T >T G G G
_§ XI G "G G yG
s G p G g G *G
G 6 <D 0) 2 c g
t-< 2 G 5 ^<D ^G
<D 5 Uh <D "C §
G O
E 5b:G
<D ’
0) G
>Y "2 *—1
^ G o
£'5b c q>"5 b> iE •-bfl O
S
<D G
b I •-
5 M
O <1> G
n pq«r2 3% 8 r^13 2 rt <D ’

rh »> 2 S « S a)
,-K % r2 ^ £ 8
pn j—i Z oo J—( i {j CQ q £5 Z <z>
The table contains the logarithmic indices multiplied by 100. Hence the
figures indicate (approximately) percentage differences. We thus find that the
German price level exceeds that of Belgium by almost 3 per cent and that in
real terms the German coal miners consume about 15 per cent less than their
Belgian colleagues. We shall postpone a further discussion of the numerical
results until the next section. Here we shall consider the following problem,
which is of a more general nature: Do the indices satisfy the factor reversal
test?
This test, as is well known, states that the product of the price and quantity
index numbers should be equal to the ratio of the total expenditures. In the
present case of logarithmic index numbers this condition is formulated in
additive rather than multiplicative form:
, ma
nab + *ab = -
mh
Is this condition satisfied? To answer this question we go back to the
definitions (1.8) and (1.10):
*ab + Kab = E Wiab(lOg P'a + lOg q‘a)

i= 1 V Pit (lib)
Z -, Fiaîa v-\ ,
Wiah log-= E Wiab log
i=i PibPib i= i wibmb
= ^g — + Z Wiah lOg
mb i= i w ib
We conclude
mn
(2.1) nab + Kab = log — + <5,ab
m.
where
(2.2) dab = E *iab log - = i E ^ log - -iXwib log -

i— 1 wib i=l VVib i=l wia
I(Wa'-Wb) - I(Wb:Wa)
Here I(wa:wb) stands for the expected information of the indirect message
which transforms the value shares of the bth region (considered as prior
probabilities) to those of the ath region (the posterior probabilities). The
interpretation of I(wh\wa) is symmetric.
Equation (2.1) shows that the index numbers do not satisfy the factor
reversal test. When we add the logarithmic price and quantity index numbers
we obtain the logarithmic difference of the expenditures plus something which

may be positive, zero, or negative. It is zero in the special case when the coal
miners of both regions allocate their expenditures in the same way: H’ia = wi6,
i = l, In that case we have the same “disorder” in the entropy sense (see
Section 2.1) in the two regions. In general, when travelling from one region
to another, we shall find a change in disorder, some value shares going up,
some going down. The sum of the logarithmic price and quantity index
numbers will then differ from the logarithmic expenditure difference by an
amount 3ab, which will be called the allocation discrepancy of the two regions.1
We see from (2.2) that this discrepancy is one-half of the difference between
the information content of the o-allocation, given the ^-allocation, and the
information content of the ^-allocation, given the o-allocation. As stated
above, the allocation discrepancy vanishes when the value shares are pairwise
equal. An impression of its magnitude in the more general case can be gained
by expanding the logarithms [in a way similar to equation (5.7) of Section 2.5],
We have
YV:n Wjab + Hwia ~ Wib)
log — = log
Wib Wiab ~ - Wib)
1 wia - wib\ 1 Wig ~ Wib\

= log - log
2 wiab J 2 wiab J
Wia ~ Wib + 1 / Wia - WibV
Wiab 12 \ Wiab J
where terms of the fifth degree and higher are disregarded. We then obtain
for I(wa:wb):
1 wia - Wib- w,-
I(wa:wh) = W iab I 1 + log
w iab W:ib
i— 1
n
i =1
i =1
This discrepancy (but with opposite sign) is also known as the information component
or the information difference component of the two sets of value shares. See H. Theil,
“The Information Approach to Demand Analysis,” Econometrica, Vol. 33 (1965), pp
67-87.
where again filth-order terms are disregarded. For /(w6:wa) we have in the
same way:
i =1
and hence, in view of (2.2):
n
(2.3) + fifth and higher order terms

i =1
We conclude that the allocation discrepancy can be interpreted in two ways,
viz., as the weighted average of the logarithmic value share differences [see
(2.2)] and also - approximately - as a fraction °f the weighted third
moment of their relative differences. The latter differences are measured as
fractions of the averages wiab, which also act as the weights in both inter¬
pretations. It is clear from the second interpretation that the differences
wia — wib should be substantial in order that the allocation discrepancies take
large absolute values.
This conclusion is corroborated by the bottom part of Table 5.3. All dis¬
crepancies are less than 1 per cent of a nit in absolute value, and most of them
are less than .2 per cent. This is small compared with the logarithmic
differences in price level, quantity level, and expenditure. Note that these
small values are not due to small differences between wia and wib. This is
clear from Table 5.2. The information values I(wa:wb) and I(wb:wa) are of the
order of .03 to .14 nit; what matters is one-half of their difference, and this is
quite small.
Note also that the discrepancies would have been much larger if we had
used the weights wia of one region instead of wiab. The sum of the logarithmic
price and quantity index numbers would then have been
n Wia
PiaQ ia ma , r/ \
(2.4) X wia log = log — + X Wia log = log — + I(wa:wb)
mb wib mb
i =1 PibQib ;= i
In other words, the sum of the logarithmic index numbers would then exceed
the logarithmic expenditure ratio except when wwib, i 1, ..., n, in which
case the excess vanishes. This result is similar to the wellknown “bias” of the
Laspeyres and Paasche index numbers.1
5.3. One-dimensional Price and Quantity Scales
In the previous section we confined ourselves to binary comparisons. Each

price or quantity comparison deals with only two regions; the other four do
not enter into the computations at all. The picture of prices and quantities
would be greatly simplified, however, if we could give it a one-dimensional
character in the following circular sense. Suppose Saarland is 10 per cent
more expensive than Belgium, and Belgium 20 per cent more expensive than
the Netherlands; then Saarland is exactly 32 per cent more expensive than
the Netherlands because 1.1x1.2=1.32. If such circular relations hold
systematically for whatever set of regions which we care to select, we can
measure the logarithmic ratios of the price levels for all pairs of countries as
distances on a straight line, each region being represented by a point on that
line. Hence the term “one-dimensional scale.” The same applies to the
volume levels, mutatis mutandis. We shall shortly see that our index defi¬
nition does not satisfy this rule. Also, we shall see in the next chapter
(Section 6.7) that there is no theoretical reason to expect this to be true.
Nevertheless, the observations may fit a one-dimensional scale closely, in
which case one may argue that it is empirically justified to work with this
simplified picture.
In logarithmic terms the scale amounts to
(3.1) nab + nbc — nac for all triples (a, b, c)
This rule is not satisfied by the nab definition (1.8). [It would be if we decided
to replace wiab by a weight which is independent of a and b.] A necessary and
sufficient condition for (3.1) to be true is that, for each pair (a, b), we can
1 The Laspeyres price index number PL for the 6th region with the ath as base can be
written as a weighted arithmetic average of the price ratiospn/pia with weights uya. The
price index considered in (2.4) is the reciprocal I/P' of a weighted geometric average of
the same price ratios with the same weights. We have P' ^ PL because a geometric average
does not exceed the corresponding arithmetic average. Similar statements can be made on
the quantity indices Q' and QL. The Laspeyres bias implies that we usually havePiQL> V,
where V-- nib/ma is the ratio of the expenditures in the bih region to those in the nth. If we
had P' =PL, Q' - QL, we would expect to find (1/P') (1/Q')< 1/K, but we see from (2.4)
that < is to be replaced by 3s. This is due to the “geometric” effect P'<PL, Q' < QL.
[I am indebted to T. Kloek of the Econometric Institute in Rotterdam for his remarks
on this point.]
5.3 ONE-DIMENSIONAL PRICE AND QUANTITY SCALES 147
write nab in the form Tia — nb, which in matrix notation amounts to
(3-2) II = ni - in'
where n is the N x N matrix [nab\, N being the number of regions, and n and i
are column vectors of N elements:
TCi 1~
(3.3) n
7T, 1
LîvJ 1_
If condition (3.1) is not satisfied, we must replace (3.2) by
(3.4) n = Jti' - in' +V
where V= [vab] is the Nx N matrix of the residuals from the one-dimensional

scale which is implied by the vector n. A rather obvious procedure is to
select n such that the sum of the squares of all residuals is minimized. This
amounts to minimizing tr V'V, where tr (trace) stands for the sum of the
diagonal elements of the square matrix V'V:
N N
tr V'V = X I v2ab
0=16=1
We apply the ordinary rules of trace algebra1 and use the skew symmetry
of IT. This gives
tr V'V = tr n'n + 4i'IItc - 2(i'rc)2 + 2Nn'n
Note that a logarithmic index vector like n always has one additive “degree
of freedom” in the same way that the ordinary index numbers are free as to
the question of whether the base value should be put at 1 or 100 or any other
positive value. We shall make use of this freedom by imposing a linear con¬
straint on the elements of n. The most convenient choice is i'tc = 0, so that the
arithmetic mean of the logarithms vanishes and the geometric mean of the
index numbers is 1. In that case the function to be minimized is simplified
to
tr V'V = tr rrn + 4i'Eln + 2Nn'n
1 The rules employed are tr (A+B) = tr A + tr B and tr CD = tr DC, where A, B, CD

and DC are all square.
TABLE 5.4
ONE-DIMENSIONAL SCALES FOR PRICES AND QUANTITIES
Nether¬
Germany Belgium France Italy Saarland
lands
Logarithmic scale
Price .0160 -.0100 .0815 -.0219 -.1812 .1155

Quantity -.0325 .1350 .0047 -.1639 .0282 .0286
Antilogs
Price 1.016 .990 1.085 .978 .834 1.122

Quantity .968 1.145 1.005 .849 1.029 1.029
This is minimized subject to i'n = 0,1 which gives the following simple result:
N
(3.5)
b= 1
Hence we take for each region the average of all its binary logarithmic price
indices (including naa = 0).
The procedure is completely analogous for the quantities, the tt’s being
replaced by k’s. The numerical results corresponding to the data of Table 5.3
are given in Table 5.4. We conclude that the Dutch prices are low on the
average. This is due to the low wage-low price government policy which was
induced by a desire to stimulate exports. The German prices are higher. As a
result it has been profitable for many years to work and to earn in Germany
and to buy in the Netherlands. Many people from both sides of the frontier
have benefited financially from this situation. The table shows also that prices
are high in France and in Saarland. It is interesting in this connection to
know that the French franc and the Saarland franc were devaluated by
17.55 per cent at the end of 1958, three months after the price survey was
held on which the present data are based.
As to the quantity indices, Belgium is clearly first and Italy clearly last.
The values of the four other regions differ from each other by at most 6 per
cent. Needless to say, these one-dimensional scales will change when one of
1 The simplest procedure is by means of a Lagrangian multiplier, which turns out to be

zero because of the skew-symmetry of [nab\.
5.3 ONE-DIMENSIONAL PRICE AND QUANTITY SCALES 149
the six regions is deleted or when a seventh is added. In principle this is not
different from the effect of changing the set of observations in regression
analysis. Formula (3.5) is so simple that changes in the set of regions can be
easily taken care of.
We must still answer the question to what extent our one-dimensional
scales can account for the observed binary values nab and Kab. The obvious
answer is in terms of the residual matrix V of (3.4) and the corresponding
matrix for the quantities. They are presented in Table 5.5. The median values
(disregarding signs) are of the order of .005, the largest absolute values are
less than .015. This holds both for prices and for quantities. The range of the
one-dimensional scale is about .3, again both for prices (Netherlands-
Saarland) and for quantities (Italy-Belgium); see the first two lines of
Table 5.4. We conclude that the residuals are small in absolute terms and also
small compared with the range of the one-dimensional scales.
A comparison of the price and quantity residuals of Table 5.5 shows that
corresponding values are mostly of the same order of magnitude but of
opposite sign. For example, in the Germany-Belgium comparison: .0025 for
TABLE 5.5
DISCREPANCIES OF THE BINARY COMPARISONS FROM THE ONE-DIMENSIONAL SCALES
Nether¬ Saar¬
Germany Belgium France Italy
lands land
Prices
Germany 0 .25 -.27 .59 .53 -1.10

Belgium -.25 0 -.08 .02 -.09 .40
France .27 .08 0 -.88 -.35 .88
Italy -.59 -.02 .88 0 -1.40 1.13
Netherlands -.53 .09 .35 1.40 0 -1.31
Saarland 1.10 -.40 -.88 -1.13 1.31 0
Quantities
Germany 0 -.19 .25 -.68 -.55 1.17

Belgium .19 0 .06 .18 -.03 -.41
France -.25 -.06 0 .42 .72 -.83
Italy .68 -.18 -.42 0 1.31 -1.38
Netherlands .55 .03 -.72 -1.31 0 1.44
Saarland -1.17 .41 .83 1.38 -1.44 0
Note. All figures are to be multiplied by 10 2.

prices, -.0019 for quantities. To explain this effect, let us indicate Germany
by 1, Belgium by 2, so that the corresponding price residual is v12. Using (3.4)
and (3.5), we can express this residual as follows:
1 1
V12 — n12 TTj + 7T2 — 7^12 ^ nl a + ^ ^2a
a= 1 a=1
When adding the corresponding quantity residual we obtain, using (2.1):

N IV
7t12 + K 12 — ^ ^ (nla + K1 a) + ~ ^ (n2a + K2a)

a= 1 a= 1
N
= log ,ni
m2
+ S12 - '
N
V (log "7l + Sla) +1N\ (log ”?2 + d2a
\ ma J ^ L\ ma j
a— 1 a- 1
a= 1 a =1
N N
= Si: l ^1 a + ^ $a2
N
a=l a=1
In the last step we made use of the skew-symmetry of the matrix [(5flb].
We conclude that the sum of the price and quantity residuals is equal to the
allocation discrepancy of the two regions, corrected for row and column
means. In our case these discrepancies are smaller on the average than the
residuals from the one-dimensional scales (the median absolute <5 is only
.0014), which explains why corresponding price and quantity residuals are
usually, though not always, of the same order of magnitude but of opposite
sign.
5.4. Variation over Time: Consumption in the Netherlands, 1921-1963
We shall now apply the same index procedure to the variation of prices
and quantities over time. The problem of the one-dimensional scale will be
avoided by the use of chain index numbers. So we shall compare the n prices
pit and the n quantities qu in period t with the values pi:t-i, of the
preceding period. This approach will also reduce the importance of the factor
reversal test problem, because we must expect that in most cases the value
share wit=pitqit/mt, mt — lpjtqjt, will be closer to x than to s> 1. In
5.4 CONSUMPTION IN THE NETHERLANDS, 1921-1963 151
words, the allocation discrepancies will generally be small when we compare

successive periods.
The basic ingredients for the price index are now of the form
log —— = A (log P;t)

Pi, f — 1
where A is the operator of taking first backward differences. The corre¬

sponding weight is the average of wit and Hence the change in the
logarithm of the price index number is
n
(4.1) A (log p,) = ^ A (i0g p;r)
i= 1
The notation, here as well as in the sequel of our time series operations, can
be simplified considerably if we write w* for the average of the zth value share
in t and t— 1, and D for the operator of taking the change in the natural
logarithm (to be called the log-change from now on):1
(4.2) wf, = +2Wu~-1 D = A (log )
Then (4.1) becomes

n
(4.3) Dp, = w*Dpit

i= 1
and the log-change in the analogous quantity or volume index becomes

n
(4.4) Dqt = X w*Dqit

i= 1
If we add (4.3) and (4.4), we obtain a result similar to (2.1):

n »
Dpt + Dqt = X K(Dpit + Dqit) = X w*D(witmt)

i = 1 »=1
n
= Dm, + X w*Dwn
i= 1
Hence:
(4.5) Dp, + Dq, = Dm, + 3,
i A table which transforms log-changes into relative changes and vice versa is given at
the end of this book.
TABLE 5.6
COMMODITY GROUPS AND VALUE SHARE AVERAGES, THE NETHERLANDS 1921-1939, 1948-1963
Commodity groups Prewar Postwar
1. Groceries 58 53
2. Dairy products 71 76
3. Vegetables and fruit 45 45
4. Meat 78 73
5. Fish 7 7
6. Bread 43 27
7. Confectionary and ice cream 28 36
8. Tobacco and tobacco products 32 43
9. Beverages (alcoholic and nonalcoholic) 31 23
10. Clothing and other textiles 125 145
11. Footwear 12 18
12. Household durables 72 75
13. Other durables 16 25
14. House rent 95 58
15. Water, light and heat 53 54
16. Services and other commodities 235 240
Total 1000 1000
Note. All figures are to be multiplied by 10-3.
where <5, is the allocation discrepancy between the successive periods:

n
(4.6) 5t= £ w*tDwit = - |J(wf_1:wr)

i= i
These concepts will now be applied to data, constructed by A. P. Barten,

on per capita consumption in the Netherlands during the prewar period
1921-1939 and the postwar period 1948-1963. Sixteen commodity groups are
used. These are listed in Table 5.6, together with their value share averages in
the two periods. For details we refer to the Appendix of this chapter (Sec¬
tion 5.B). Table 5.7 contains the log-changes in the price level, in the quantity
level, and in total expenditure, and also the allocation discrepancies. The
figures refer to pairs of successive years in the prewar and the postwar period
as well as to the war transition. For that transition t should be interpreted as
1948, t-1 as 1939. We conclude that prices declined in 1922 from the 1921
level by about 10 per cent, that the quantities went up on the average by
about 7 or 8 per cent, so that on balance total consumption expenditure per
capita decreased by 2\ per cent in 1921-1922. Prices decreased on the average
TABLE 5.7 5.4
LOGARITHMIC PRICE AND QUANTITY CHANGES AND ALLOCATION DISCREPANCIES FOR CONSUMPTION PER CAPITA IN THE NETHERLANDS, 1921-1963
CONSUMPTION IN THE NETHERLANDS, 1921-1963
Note. All figures are to be multiplied by 10

153
by about 2\ per cent per year in the prewar period; quantities went up on the
average, but only by 1 per cent per year. The booming postwar period gives a
different picture, both prices and quantities increasing by 3 per cent per year
on the average. The war transition 1939-48 has the nature of an outlying
observation. Prices were considerably higher after the war than before.
Consumption per capita in real terms was about 9 per cent less in 1948 than
in 1939 owing to destruction and disorganization during the war.
The allocation discrepancies are all very small with the natural exception
of the war transition. [Note that the fourth column contains 100 dt, not <5t.]
There is only one case in which the absolute value is larger than \ x 10~4. The
absolute median is less than 1 per cent of the absolute median of the <5’s of
Table 5.3 in the cross-section case. We conclude that in the present case-apart
from the war transition - the factor reversal test is almost exactly satisfied.
5.5. Variances and Covariances of Logarithmic Price and Quantity Differences
The crucial difficulty of all price and quantity comparisons which we

considered in the previous pages is the variability of the value shares: wia
versus wib in the cross-section case, wit versus w;>(_ t in the time series case. If
these shares were pairwise equal for each i, the allocation discrepancy would
vanish, so that the factor reversal test would be satisfied exactly. If, moreover,
wia = wib were true for each i and for all pairs (a, b), the one-dimensional scale
of Section 5.3 would fit exactly. It is therefore appropriate to consider the
differences in value shares for different regions or time periods in more detail.
This will at the same time lead to interesting results on the dispersion of
prices and quantities around their respective logarithmic averages.
We recall that the allocation discrepancy is the weighted mean of the
logarithmic differences between corresponding value shares, see (2.2) and
(4.6). An obvious measure for the degree to which the value shares differ is
the variance of the same distribution. Using the time series notation, we can
evaluate this variance as follows:
n n
Z w*t(Dwit - S,)2 = X w* (Dph + Dqit - Dm, - S,)2

i=1 t= 1
= Z w* [.(Dpit - Dp,) + (Dqit - DqJ]2

i= 1
= Z vv* (DPit ~ Dp,)2 + £ w* (Dqit - Dq,)2

i= 1 i= 1
n
+ 2Z <(Dpu-Dp,)(Dqit-Dq,)
i = 1
5.5 VARIANCES AND COVARIANCES 155
where use has been made of (4.5). So we have the following decomposition of
the variance of the log-changes in the value shares:
(5-D t < (Dwa - S)2 = nt + Kt + 2rt

i= 1
where 17t is the variance of the log-changes in the individual prices:
(5.2) n, = Y w* (Dpit - Dp)2

i= 1
Kt the variance of the log-changes in quantities:
(5.3) Kt= jl w* (Dqit - Dq)2

i= 1
and r, their covariance:
(5.4) rt=i w* (DPit ~ DP,)(Dqit ~ Dq)

i= 1
We have thus succeeded in decomposing the variance of the log-changes in

the value shares into three elementary price-quantity concepts. The first is the
variance 71, of the logarithmic price changes, which vanishes when all prices
change proportionally from t— 1 to t. In that special case the log-change Dpt
in the price index describes the behavior of all individual prices completely.
In general, of course, some prices will increase more than others and there
may even be changes of opposite sign. We then have a complete distribution
of log-changes in prices of which Dpt is the mean. The variance 77, is in that
case a convenient measure for the dispersion of the individual changes around
their mean; we may say that it is a quantitative measure of the change in the
price structure, whereas the mean Dpt is concerned with the change in the
price level. In cross-section analysis we have an analogous concept, to be
written 17ab rather than 77,:
(5.5) nab = j) wiab(log — - nab]

i= i V Pib )
It measures the extent to which the price structure differs in the two regions -
see the example on Fords and household assistance in the United States and
India at the end of the first paragraph of Section 5.1.
The second component of (5.1) is the logarithmic variance K, of the
quantity changes. It measures the change in the volume structure. The third
component is the covariance term. One should expect that it will usually be
negative, given the buyers’ tendency to substitute in favor of those com-
156
LOGARITHMIC VARIANCES OF VALUE SHARE, PRICE AND QUANTITY CHANGES AND LOGARITHMIC COVARIANCES OF PRICE AND QUANTITY CHANGES
FOR CONSUMPTION PER CAPITA IN THE NETHERLANDS, 1921-1963

PRICE AND QUANTITY COMPARISONS

5.5
VARIANCES AND COVARIANCES
157
LOGARITHMIC VARIANCES Of VALUE SHARE, PRICE AND QUANTITY DIFFERENCES AND LOGARITHMIC COVARIANCES OF PRICE AND QUANTITY
R3
C
d oo r- o on o n vn in o h o O o r- cn i-i ©
d
O <n o m
vo On on h- ON On O N On ^Tj-vooomo
d <N
nmrHTtr-, O rn in no <n
in h (N
*d
d
Th vOMnooOh ^ Is On NO O H
<D ^ oo o oa
MOnvo r- m
Hnmm
CN y—i t"- as
t-h , cl ^ rt
»n oo oo t-i o m
H(NONO m
£ m o oo O rt — rj* m
"cd ^ 1-i <N
On O O oo ON OO OO rt O AO o
c3 AO <N rt
n rib
On O
AO OA
co
mm
rt rj-
r-- r-
m rj-
d- r- ao o rt <n
O T-H oo r}- o l~~ OS CT\ O >—i OO
O0«o AO AO
rt- <N -rt i-h (N O NO m
tN fN t-h <N fN
M- (N T-l CO
I I
DIFFERENCES IN THE CROSS-SECTION CASE
<D
o £2 O O O vo O
d ao r-* o i-H On »n T|r-OAO(Nh ON Tt o ON OO NO
d £ "JO rt o <N mm i-H o <N rt- CO o (N
Irt AO OA h m OA •n m Tf m h ^ c- O o m
TABLE 5.9
M ON ^ CO rt m ^ rt Ttrt
i i i
E
d o o r- r- r- M O h OO h Vi JrthAOO
OA O <N oo o *3! cn d- <N oa rt-1 -rt ( Tt
g\ o d* oa oo rt
'Sb O (''' rt— fNj r—I
<v
r- OA rt O no
CN -rt
rt m m m m OA OAMOxf m rt- (N rt- rt-
fN -rt
pq
>S
Note. All figures are to be multiplied by 10 4,
d
a O n m Oa ao oo OclNOoccm Ort IDi-rt O O On oa c* >n m
oa rj- ao m m
h ao m ci ao
m m h On
d* m m — m
MNO-nh OincOrtO
OA fN © m rt m m rf rt-
<D rtH (N
o
_
/5
C /5
C
*o -d *d
>> c >> d
d -d d’d d "d >a d T3
d d d c g'g
t: g d S <D T2 g *
d s t: g
E •- 8 te J
£ 3 8 <d d g § O <d d CD d
g.S
E m = £ w) E w> d ., rt nnCaO b Z'S 15
.—o ^ S |»
~ Sj
or <l> 13 2 d i_z,
, „ rt’ ,rt ju rrj rs«s ctf i-7
<D 2
O CQ IL — Z 00 Q PP Ph £- Z C/3 O CQ P- a—i Z C/3 0 pq Ph hZw d
modities that have become relatively cheaper. Since the approach of this
chapter has its roots in statistics rather than in economics, we shall not pursue
this aspect further. We confine ourselves to measurement here, but we shall
come back to this topic in Chapter 7 (particularly Section 7.6).
Table 5.8 contains the decomposition (5.1) for pairs of successive years in
the case of per capita consumption in the Netherlands. Taking square roots,
we conclude that the standard deviation of the logarithmic price changes in
1921-22 is about 7 per cent, that the standard deviation of the logarithmic
quantity changes is slightly larger than 9 per cent, and that the two sets of
changes are negatively correlated:
749.45 x 87.22
Our expectations on the sign of the covariance turn out to be justified. There
are only two positive values, both in the prewar and in the postwar period.
Table 5.9 contains similar data for the cross-section case of the coal miner
families. It shows that the covariances are also negative with only two
exceptions. The cross-section variances are considerably larger than the
variances of the year-to-year changes in Table 5.8, which is understandable.
In particular, the cross-section quantity variances are very large. They exceed
the corresponding price variance in each binary comparison. It is instructive
to compare Ilab and Kab for Saarland and France. The price variance is very
small, which is due to the fact that Saarland was under the French economic
regime at the time. The quantity variance is more than ten times larger in
spite of the fact that the quantity levels of the two regions differ by less than
3 per cent (see Table 5.3). The conclusion is tempting that the Saarlanders
continued to eat their Bretzel and did not switch to croissants.
5.6. Partial Index Numbers and the Decomposition of Price and Quantity
Variances and Covariances
This section and its successor will deal with aggregation problems. We
imagine that the 11 commodities are combined to G sets Sl5 ..., SG such that
each good belongs to exactly one Sg. We define value shares for sets:
(6.1) Wgt = Z vv„ g = l,...,G
so that the average set value share in the periods t and / —I becomes
5.6 PARTIAL INDEX NUMBERS 159
We define the log-changes in the partial price and quantity index numbers
of Sg as follows:
DPn,= w*DPit
Wrat
ieSa
(6.3) g = 1
DQatgt = w* w*Dqit
rrgt
ieS„
This means that the individual log-changes are weighted by means of the
conditional shares w*/W*. The partial index numbers are consistent in
aggregation when these weights are used. This is verified by multiplying both
sides of the two equations (6.3) by W*t and summing over g:
I W*DPgt = i <DPit = Dp,

(6.4) 9;1
(j n
Z W*tDQg, = £ w*,Dqit = Dq,

9=1 i =1
We conclude that the overall indices can indeed be obtained from the partial
indices by means of exactly the same weighting procedure as that of (4.3)
and (4.4).
Note that we could also have proceeded in a different manner. We intro¬
duced w* as the average of the zth value share in two successive periods. One
can then argue that wit/Wgt is the conditional share in t and wiJ^1/WgJ_l the
conditional share in t— 1, so that the weight of Dpit in DPgt and of Dqit in
DQgt should be
w;, Wj,t- i\
(6.5) +
at Wa,t-J
which differs from w*/Wgr If this procedure is adopted, there is no consistent

aggregation in the sense of (6.4), which is the reason why (6.3) is preferred.
However, the use of this definition implies that we have to modify our two-
stage random selection interpretation.1 When prices in New York and San
Francisco are compared, we should no longer fly to these cities with proba¬
bility j each, because this would lead to weights of the partial index numbers
which are of the type (6.5). We should go to a third city, not necessarily on
Earth, where the value share of each commodity happens to be the average of
1 See footnote 1 on page 138.

TABLE 5.10
160
LOG-CHANGES IN PARTIAL PRICE INDEX NUMBERS FOR FOOD (1), VICE (2), DURABLES (3), AND REMAINDER (4) IN THE NETHERLANDS, 1921-1963
PRICE AND QUANTITY COMPARISONS
Note. All figures are to be multiplied by 10 2

TABLE 5.11
5.6
LOG-CHANGES IN PARTIAL QUANTITY INDEX NUMBERS FOR FOOD (1), VICE (2), DURABLES (3), AND REMAINDER (4) IN THE NETHERLANDS, 1921-1963
PARTIAL INDEX NUMBERS
Note. All figures are to be multiplied by 10-2

161
the corresponding shares of New York and San Francisco. That leads to the
definition (6.3) with its consistent aggregation property.
Tables 5.10 and 5.11 contain an application of the partial index numbers
to the time series data. The sixteen commodity groups of Table 5.6 are
combined to the following four sets:
(1) Food, consisting of 1. Groceries, 2. Dairy products, 3. Vegetables and
fruit, 4. Meat, 5. Fish, 6. Bread. The prewar average Food value share is .302,
the postwar average .281.
(2) Vice, consisting of 7. Confectionary and ice cream, 8. Tobacco and
tobacco products, 9. Beverages. The prewar value share average is .091, the
postwar average .102.
(3) Durables, consisting of 10. Clothing and other textiles, 11. Footwear,
12. Household durables, 13. Other durables. The prewar value share average
is .225, the postwar average .264.
(4) Remainder, consisting of 14. House rent, 15. Water, light and heat,
16. Services and other commodities. The prewar average value share is .383,
the postwar average is .352.
The logarithmic variances and covariance of the price and quantity changes
also have simple aggregation properties. Consider
G
nt= ^^ w* (Dpit — Dpt)2

g= 1 ieS9
The term in square brackets, which is the contribution of Sg to 77„ can be

written as follows:
^ w*(Dpit-DPgt + DPgt-Dpt)2
eSg
= £ <(Dpit - DP,,)2 + V <(DP„ - Dp,)2

ieSg i e Sq
+ 21 ieSg
-HS1
wit
W*
rr gt
ieSg
We conclude:
G
(6.6) nt-= n, I gt
9= 1
5.7 THE DISPERSION OF VALUE SHARE CHANGES 163
where /70, is the between-set logarithmic price variance:
(6-7) n0t= £ W*(DPgt-DPt)2

9= 1
and f[gt the logarithmic price variance within Sg:
<6'8> n„= £ ~i(Dp„-DP„)2

icSg
We have completely similar results for the quantity variance and the
covariance:
Kt = K0t + £ W*Kgt
9= 1
(6.9)
r, = r0, + l w*rt,
9= 1
where
K„,= I W*(DQgt-Dqt)2
9= 1
(6.10)
G
r„,= E W*{DPgt-Dpt)(DQgt-Dqt)
9= 1
and
*
VV-*
Kgt = - oe„)2
is So
(6.11) *
VV-,
rgt - DP„)(Dq„ - DQ„)
ieS„
The decompositions (6.6) and (6.9) are shown in Table 5.12 for the four
sets Food-Vice-Durables-Remainder. It appears that on the average roughly
one third of the variances and the covariances is accounted for by the
between-set variation, and two thirds by the total within-set variation. The
negative sign pattern of the covariances applies to both components of T( in
all years with few exceptions.
5.7. The Decomposition of the Dispersion of Value Share Changes: An

Informational Approach
In Section 5.5 we derived the logarithmic price and quantity variances and
covariance as parts of the logarithmic variance of the value share changes.
We considered these separate parts in the aggregation analysis of Section 5.6,
NO co Vi C- d OO Vi OO NO ON Vi o
,i3 o 00 Vi CO V) co Cl CO d ’'t d O 00 o
G_ 00 On — cd Vi Vi ci 00 Vi ON 'd cd
o 1 1 1 1 1 1 1
G £ 7 7 7 7 7
C3
> G c- o c~ c- c- oo NO O Vi co O NO
O on NO C| o o CO Cl CO d CO CO On O 1
<u -H
u CO Cl ci vi cd ci c- d -H cd d
£ |
d | | | | | | |
<u co
m i
_G CO oo c- Cl Vi oo CO
O co
G vn NO d- vo d- NO CO r- CO d o d co V) d V)
vi NO cd d ■'d vH CO ■d vi
<u Tfi o d d
a £ to
h G
G C
a
co Q
X
F 73 co
z G 03 § V.
©
O
ON ON co Vi Cl o co NO d vi ON d ON
< a
ON co r- OO r- co NO NO CO CO CO co
& 1 no ci _ d NO d ’'d ^H ON vi —i d d
a co
d
Cl
o m
z
m
tzi so c_ oo oo oo Cl Cl 00 co NO Vi co c c~ Cl On
uj o\ o G o On Vi d- NO vi Vi Vi o O d CO NO
no NO ci cd cd o cd ON —. o ci o d o
c- T_* Cl 1 Cl 1—1 d 1 y~l
£ d-
fS al
>
OS
O G
<D co V) Vi 00 NO Cl NO d CO o d
vn d- co d- c- d o Cl o d o NO NO o Vi 00
a £ —1 co 00 NO ci vi ci Tf vi o vi vi
o co
z w in
c
a
yox
H 00 On o Cl CO d- Vi NO c- oo ON O d CO u
o W T t
Vi
1
Vi
1
Vi
1
Vi
1
V,
1
V) Vi
1
V)
1
Vi
1
Vi
1
NO
1
NO NO NO
1 s
00
a3
-J Z On OO O Cl CO 4 Vi NO c- 00 o d
(N
w co
ON
T
ON ON
Vi
On
Vi
ON
Vi
ON
Vi
ON
Vi
ON
Vi
On
Vi
ON
V)
ON
V)
ON
VI
ON
NO
ON
NO
ON
NO
ON >
o X
i/S
i/a
H
W U W
m z< <H NO <N Vi Vi o Vi O ON 00 OO ON d
w 00 r-~ 00
O co O NO CO NO Cl d- CO NO c- •rt Vi l-H CO CO •^H
< ■G
H < s
OO ON 00* Vi -d oo NO r- NO d — NO On o d
> <
o
$ 1 1 1 1 1 i I- 1 1 1 i 1
o o
G 1 1 | |
nJ
u W Oh cd
> IT) O c| V) CO co o oo o ON ON 1-H CO Vi OO NO NO
p O O O 00 NO On 1-H1 o H-l d c- NO -'t O — d ON d
z z O £ o‘ <N NO — vi o d — ■'d vi d
< o <D
c/a H CQ
H Oh
U s
z D _G
< oo G
On
co ONr-
o O CO Vi
On
CO
Vi
co
oo
CO
c~
CO
CO
d
o
d
NO d
Tf
Vi
o
"3-
o O
co
O
NO
o
5 Z •G — d i—(
< O
cd ON d- ci c~ ci o d cd cd NO vi tj-’ d d vi
4i Jk •ct d Cl
> CJ
d *■" d
o £
Note. All figures are to be multiplied by 10 4
G £
G _d
M, «
O o
P4
G oj G
<D £ co NO C'l NO O Vi ON Tf CO co Vi CO CO
_
o
a a>
z oo Vi OO CO O CO ON CO d d Vi HI NO O'. CO d
o <u
cd Ol d 00 Tf ci ci d NO vi 00 00*
CQ
o
a.
_G
s NO d-
■—1
co o OO cl r- o d 00
O <D G o C'l c
1) G
<u r- ON 00 Cl CO V) NO 1
o <D ON d f- 00 O
co ON o CO O (M O 00 d o 00 NO Vi d c- CO Tt- d
£ y-4
Oh d CO 00 cd ci OO oo" vi vi T-l 00 vi 00
co d
03
c| co ■0- Vi NO 1 00 ON O d CO Vi
cl c-l <1 cl <S Cl (S Cl CO co CO CO CO co CO CO <u
1 1 1 1 1 1 1 1 1 1 1 1 a bo
cfi
<N CO 4 Vi NO h- oo ON O 4 CO 4 Vi 4 %
C| ci Cl Cl CO CO co co co
ON On ON ON On ON ON ON ON ON On ON ON Nh
0H a
5.7 THE DISPERSION OF VALUE SHARE CHANGES 165
so that an obvious question is whether we can obtain similar aggregation

results for the value shares. This is indeed possible as is shown in the Ap¬
pendix (Section 5.C). Here we shall consider a slightly different problem.
Since the value shares are, in contrast to prices and quantities, nonnegative
numbers which add up to 1, we can apply a decomposition in informational
terms.1 This decomposition is not symmetric in the periods t and t— 1; but,
when compared with the variance decomposition, it has the advantage of
simplicity. Consider the expected information of an indirect message which
transforms the allocation proportions of t—\ into those of t:2
W;,
(7.1) 7(wt:wt_1)= £ w.fiog-T7L_= £ witDwit
i— 1 Wf, (_ i i= i
We know from equation (7.4) of Section 2.7 that this information expec¬
tation can be decomposed as follows:
(7.2) J(wt:wt_1) = /0(wt:wt_1)+ Z WgtIg(wt:wt^)

9=1
where
4K:w(-,)= Z WgtDWgt
9= 1
(7.3)
4(w<:wi-i) = g = 1,..., G
w
rr gt w
’’ gt
ieSg
Hence we can write (7.2) in the equivalent form

n G G
Wit Wit
(7.4) y witDwit = y wgtDwgt+y wsgt —D —
w
rvgt
w
vvgtJ
i= 1 9=1 9=1 i E Sn
The left-hand side of (7.4) is of course not the variance of the value share
changes, (5.1). Nevertheless, the two are closely related. To show this we
start as follows:
1 Awit
(7-5) Z witDît = Z w* 1 + A —sr ) Dwn = $t + l Z
2 w
-i'DWi
i= 1 i= 1 i=l Wit
1 The similarity of variance analysis and information analysis was pointed out by W. R.
Garner and W. J. McGill, “The Relation between Information and Variance Analyses,”
Psychometrika, Vol. 21 (1956), pp. 219-228.
2 Remember that we considered the same information expectation in relation to the
factor reversal test at the end of Section 5.2.
Now the change Awit measured as a fraction of w* is normally very close to

the log-change Dwit. This is verified by expansion:
w* + }Awit 1 dwir\ 1 dwi(N

Dwit = log log - log
2 w*J 2
Awit 1 (Awit\3
< 12 V w* )
On combining this result with (7.5) we conclude that the information expec¬
tation is approximately equal to half the second moment plus the allocation
discrepancy St:
1 ^ A XV:
Dwit -
X wirDwit — <5f + I w*(Dwlt)2 2 4 I W:
1=1 i= 1 i= 1
We know also, however, that St is of the third order in the ratios Awit/w*.
This follows directly from (2.3) when that equation is written in time series
notation. Therefore, we may conclude that the decomposition (7.4) must be
approximately equivalent to a variance1 decomposition apart from the
factor j.
This conclusion is verified in Table 5.13. The first column contains the
variances of the value share changes (taken from Table 5.8). The second
column contains the left-hand side of (7.4), the third and fourth contain the
two terms on the right; the figures of these columns are all multiplied by 2 in
order to facilitate the comparison with those of the first column. It turns out
that the first two columns are indeed practically the same. The other columns
indicate that on the average the between-set variation accounts for about one
third and the total within-set variation for about two thirds of the dispersion
of the value share changes - a result which should not surprise us in view of
the earlier results obtained in Table 5.12.
5.8. Concluding Remarks
The informational application of the previous section concludes the

statistical approach to the problem of price and quantity comparisons. The
analysis will be continued in the next two chapters, which have a more
economic flavor. Chapter 6 deals with the theory of consumer demand; that
1 We may interchange “variance” and “second moment around zero” freely, because the
difference between the two is Sf, which is of an even higher order of smallness than the
third.
TABLE 5.13
LOGARITHMIC VARIANCES OF VALUE SHARE CHANGES AND THEIR INFORMATION DECOMPOSITION FOR CONSUMPTION PER CAPITA IN THE NETHERLANDS, 5.8
1921-1963
CONCLUDING REMARKS
167
Note. The variances are to be multiplied by 10“4, the information values (unit: 1 nit) by \ X 10
chapter is completely theoretical. In Chapter 7 we shall use some of the data

described in the previous pages to illustrate the ideas explained in Chapter 6.
Several of the concepts and symbols used in this chapter will also be used
in the next two. In addition, more symbols will be introduced in the sequel.
The Appendix of Chapter 7 (Section 7.C) contains a summary which should
facilitate the reading of Chapters 5, 6 and 7.
5.A. Details on the Cross-section Data
The data of the cross-section analysis of coal miner family expenditures

have been taken from two publications of the Statistical Bureau of the
European Community. One deals with family expenditures of workers in
the European Coal and Steel Community, 1956-57 (Statistical Communi¬
cations, Social Statistics Series, No. 1, 1960, published in Dutch, German,
French and Italian). This publication will be referred to as Family expenditures.
The second is concerned with real incomes in the European Coal and Steel
Community (same series, No. 2, 1960, same languages) and will be referred
to as Real incomes.
The price data, house rents and value shares are all taken from Real in¬
comes (Appendix IV, Table 3, and Appendix V, respectively). The amounts
of total consumption are computed from the amounts given in Family
expenditures (Tables III-l, III-10, III-14, III-24, III-39, III-47) by subtracting
the items which were not considered in the price survey.1 The results are as
follows:
Germany 6,571.52 (German marks)
Belgium 89,920.52 (Belgian francs)
France 730,572.92 (French francs)
Italy 826,264.92 (lire)
Netherlands 5,185.85 (guilders)
Saarland 772,730.81 (Saarland francs)
The official exchange rates are taken from Family expenditures, Tables 9-14.
The matrix of their logarithms is not exactly skew-symmetric, but this pattern
was imposed by replacing the matrix by the average of its transpose
and itself. After that the single-scale procedure described in Section 5.3 was
applied to this average.
1 The items excluded are 0620, 09,10,11,14,15 and 16 (in the code of Family expenditures).
Neither social insurance premiums nor taxes were included.
169
CO o tJ- CM in co c~ so os 00 00 © OS
in in © t|- OS O
C/5 I/s SO CO 1/1 CO
Z CO CO CO CO CO CO CO CO
7
CM
1 1
r-
1 1 1 7 7 1 1 1 1 1 1 1 77 1 1 1 1
O CO o CM m CO o c- c- •n Os c~
C/5 OS o 00 oo r- Os CO l''
so I/S © Cl 00 o CM CO CO
co CM "rf CO ■'t
1
CM
1
CO
1
CO CM
1 1 7 1 1 i I 1 1 i 1 i 7
,
■rt CO
z SO 00
r- r-
CO O
OS 1n
1*-
’t
O
co
OS
OS
in
r-
m,
so r-
CO
oo
CO •n OS
oo
CO ©
"t
CM O
co ’t CM CO
1
1 1
7
CO CO o r- SO 00 oo 00 in in SO r- •<t CO
FS
CM CM OS so CO so OS 00 c- Cl in
oo so in CM CM CO r- so r— in CM
1
7 1
7
CM
1
i
7 1 i
i
i
1
z «n •n CM oo o so CM CO o SO m co 00 CM
COMMODITY GROUPS
r- oo SO 00 >n r- OS oo 00 CM Os os •n co ©
CO so 00 C- r- oo m, m o r- m, o
CO CM SO (N CM CO CO co CM CO CO CM co CO CM CM
7 1
co CO O oo o Os OO Os 1^- C- m CO Os O SO © •n
O 'n r-~ co O Os so O r- CM «n ■*t o so CM CO "rt
E SO in Cl CM so O Os Os in O so CO co CO 00 OS
(N CO CO CM CO CO co CM -t CO CO
77 1 1
i
1
C/5 CM O CM oo OS I/s SO OO SO CM Os O in 00 ■'d' OO m,

OS OS O r- m, o 00 </) C- Os 00 co c- Cl O co Tt OS so
•rt os r-~ co 00 OS ■'t SO f" so co CO so "t OS CO 1/, 00 cm
CM I/I CM CM CM co CO CM
1 1 I 1
21
1 i HI 1 ( 1 1 1 1 1 1 1 | 1 l
z 00 CO o CO
LOGARITHMIC PRICE DIFFERENCES FOR
Os os CM CM CO so in SO Os o Cl CO CM co
m OS C" Tf o c- Mb r*» CO 00 Cl Tt SO oo O CM ©
os OS rt CM os Os OS OS m CM so CM c- r- in
■'fr "'t CO CM CO 1 1 1—1
-
1
2
r— CM CM O CO o CM CM c- CO CM m OS 00 r-~ so o
CM Os O I/s r- -t SO OS so co OS SO CO © •n ©
— i/i CO 00 OO O in SO CM o SO co SO
■'d- 1/1 CM CM CO
l
1 7 1 1
1
7
,
Uh OO so o Os •n in Os 00 00 Tt" SO Cl o CO
cq co OO CM OS oo oo rooo c- 00
Tf CM OS i/i o 00 CO O so co Os
■*t CM CM co "It
7 1
1 i
1 | 1
1
7 i
7 1 1 ! 1 7 1
1 1
C/5 so <n 't Os so 00 Os SO SO o

CO 00 00 OS CM ’t CM O t"- co Cl CO
0 ■'t OS t-* r- CO t" m, m CO
CO CM CO CO
1 777 1 77 1 1
Tf
1 i
1
7 I 7 |
z
r- CO so co ICS 00 oo in o o Os 00 CO
CM o r- •n os CO SO CM
Note. All figures are to be multiplied by 10~4
0 >n Os in m,
«n CO CM CM
r- CM so o ! Tf Tf OS co r-
Tf- oo o "cb 00 O Os
oo >n co
CM CO ■rt 1
77 1
1 1
7
CM
|
CM
U< OS O Os CM o o oo CO co r- so
0
•n r- Yf co
CO «n in, OS CM Cl co OO so
i co CO co CO
1
1 1
1 1 7 I 1 1 | | | | |
1 1
CQ rt cm c- C" SO 00 o , co CM co SO CO
f- OS CM O Cl C- OO oo CO CO 00
0 co oo CO t*- O in © 00 "t o in >n Tt 00 00 ■'t so CM
1
CM CM o CM
1
1
7 1 1
7
Cl
1
1 7
o
z — (N CO «n SO r- 00 Os o — CM CO Tt •n SO r- oo © _ .©
CM CM
5.A
LOGARITHMIC QUANTITY DIFFERENCES FOR 21 COMMODITY GROUPS

CROSS-SECTION DATA

171
Table 5.14 contains the logarithmic price differences for the 21 commodity
groups, defined as follows:
(A.l)
[ W(r)a + W(r)b
log
P(r) a
P(r) b
where Ft is the set of commodities that are part of the z'th group (7=1, ..., 21),
p{r)a the price of the rth commodity in the ath region, and w(f.)a its share in that
region’s total expenditure per family. For quantities (Table 5.15) the loga¬
rithmic differences are obtained by subtracting (A.l) from the logarithmic
difference of the expenditure on the commodity group:
wiama Pia
(A.2) = log - log
<hb
wibmb Pib
This amounts to neglecting the allocation discrepancy within the ilh com¬
modity group. It will be noted that the left-hand notation of (A.l) and (A.2)
suggests that these are partial logarithmic index numbers which satisfy the
one-dimensional scale:
Pia Pia
log log
Pib Pic
but this is actually not the case, witness the right-hand sides of (A.l) and
(A.2) and also Tables 5.14 and 5.15. The 15 columns of these tables refer to
15 pairs of regions which are indicated by their initials (GB = Germany-
Belgium, etc.). The last row of the tables contains the overall indices nab
and Kab.1
5.B. Details on the Time Series Data
The time series data for consumption per capita in the Netherlands were
constructed by A. P. Barten. From various sources, both published and un¬
published, he computed prices and total expenditure series for 99 basic
commodities before the war and for 108 commodities after the war. This leads
immediately to the value shares wit for the 16 commodity groups and Wgt for
the sets, which are given in Table 5.16. Logarithmic price indices Dpit for the
16 groups are then defined in a way similar to (A.l), and logarithmic quantity
indices Dqit in a way similar to (A.2); the only difference is that a and b are
to be replaced by t and t— 1, respectively. The results are given in Tables 5.17
Table 5.14 was also published in the article by Kloek and Theil, quoted in the text.
However, rows 6 and 7 should be interchanged in the table of that article.
5.B TIME SERIES DATA 173
TABLE 5.16
VALUE SHARES OF SIXTEEN COMMODITY GROUPS AND FOUR SETS OF GROUPS,
THE NETHERLANDS 1921-1963
Wit W2f W3f Wit W5t wet W7t
1921 553 770 555 979 67 451 296

22 531 731 510 1003 69 391 267
23 594 745 474 998 72 400 269
24 597 755 500 908 98 416 272
25 579 765 461 878 79 451 286
26 611 720 447 883 76 453 286
27 600 736 459 798 86 442 277
28 579 736 464 816 88 428 271
29 572 741 445 803 96 383 255
1930 560 693 411 789 87 390 261
31 546 671 452 736 73 357 247
32 561 647 473 660 67 352 246
33 594 672 420 640 51 372 254
34 574 661 406 675 61 436 279
35 602 675 405 650 52 459 291
36 578 666 379 649 56 479 306
37 600 682 386 651 78 491 305
38 614 686 418 691 80 491 318
39 622 690 400 652 70 461 305
1948 439 893 534 393 104 315 393

49 551 846 424 520 66 324 415
1950 520 761 482 726 69 297 382
51 539 773 461 773 63 306 364
52 579 806 456 823 61 349 378
53 564 815 457 840 61 332 383
54 582 800 467 820 58 301 358
55 569 779 430 772 67 274 346
56 567 732 444 794 66 248 345
57 515 765 449 774 64 239 348
58 522 748 438 776 71 239 352
59 520 736 450 774 67 234 347
1960 503 699 435 720 71 220 338
61 512 695 444 720 74 210 337
62 517 684 458 704 71 210 325
63 527 684 447 675 71 204 326

TABLE 5.16 (continued)
wst W9i Wiot Wilt Wl2t Wl3t Wilt
1921 281 370 1225 147 795 176 579

22 288 353 1427 132 771 165 631
23 310 343 1297 109 758 151 703
24 330 345 1311 119 748 142 735
25 295 348 1305 120 730 150 767
26 325 353 1287 128 718 150 804
27 321 350 1330 127 763 152 848
28 324 350 1297 133 766 157 874
29 338 317 1322 134 780 154 893
1930 353 309 1322 131 751 156 924
31 365 324 1267 120 722 156 1010
32 356 306 1207 106 741 150 1120
33 340 288 1205 109 708 163 1164
34 321 274 1142 101 647 162 1198
35 308 269 1133 104 624 163 1216
36 301 260 1154 102 668 151 1195
37 298 256 1179 107 647 170 1142
38 300 259 1148 102 591 170 1127
39 309 254 1250 125 666 176 1064
1948 394 266 1492 218 625 210 577

49 421 237 1759 183 616 196 534
1950 446 196 1807 189 589 221 493
51 445 213 1666 192 654 196 524
52 475 198 1443 177 618 183 541
53 463 206 1413 170 633 210 532
54 455 206 1435 171 686 236 567
55 433 201 1457 169 748 269 540
56 419 207 1504 176 771 278 517
57 454 229 1393 165 845 243 549
58 460 236 1314 169 824 217 618
59 449 249 1291 170 809 271 613
1960 429 242 1327 185 869 288 662
61 417 259 1333 179 878 338 666
62 402 271 1317 177 911 344 668
63 374 279 1312 180 958 376 673
TABLE 5.16 (concluded)
Wl5t Wl6t Wu Wn W3t Wit
1921 414 2343 3374 948 2343 3336

22 430 2300 3235 909 2495 3362
23 446 2332 3283 922 2315 3481
24 472 2249 3275 947 2321 3457
25 463 2324 3212 929 2305 3554
26 492 2266 3191 964 2283 3562
27 484 2228 3120 948 2372 3560
28 476 2241 3111 945 2354 3590
29 508 2260 3040 910 2390 3660
1930 519 2345 2929 922 2360 3788
31 543 2409 2835 937 2265 3963
32 574 2435 2759 908 2204 4128
33 588 2432 2749 883 2185 4184
34 589 2474 2814 875 2052 4260
35 592 2458 2842 869 2024 4265
36 603 2454 2806 867 2075 4252
37 623 2385 2888 858 2103 4150
38 609 2395 2980 877 2011 4132
39 594 2363 2894 867 2217 4021
1948 425 2724 2678 1052 2544 3726

49 457 2451 2732 1073 2753 3442
1950 482 2340 2854 1024 2806 3316
51 536 2294 2915 1022 2708 3354
52 580 2333 3074 1051 2421 3454
53 574 2347 3070 1052 2425 3453
54 550 2309 3027 1019 2528 3425
55 555 2391 2890 980 2643 3487
56 575 2356 2851 971 2730 3448
57 582 2387 2805 1031 2646 3518
58 586 2431 2794 1049 2524 3634
59 560 2460 2781 1045 2542 3633
1960 549 2465 2647 1008 2669 3676
61 536 2400 2656 1013 2729 3602
62 564 2377 2643 998 2749 3610
63 601 2312 2608 979 2827 3586
TABLE 5.17
PRICE LOG-CHANGES FOR SIXTEEN COMMODITY GROUPS
Dpu Dpzt Dpzt Dpu Dpu Dp 6( Dpu Dpst
1921-22 -1381 -2257 -1807 -1366 -924 -1333 -1061 0

22-23 368 -674 -698 -111 483 -436 -296 29
23-24 222 403 304 -588 1451 98 -137 -271
24-25 -58 324 967 8 -650 1029 -434 201
25-26 -182 -1118 -1351 -442 -635 -467 -1865 -266
26-27 -99 -200 -414 -828 300 -151 85 -171
27-28 21 294 299 4 101 -195 -228 -46
28-29 -280 -15 -520 645 586 -530 -299 49
29-30 -753 -1120 -1078 100 -1162 -580 -521 -3
1930-31 -755 -1470 -242 -2115 -1125 -1227 -569 -267
31-32 -711 -1877 -1031 -2500 -55 -746 -467 -1105
32-33 392 627 -2404 278 -3224 379 -18 -1216
33-34 17 -702 27 592 653 377 248 -688
34-35 -262 -27 -647 -788 -409 -157 -663 -785
35-36 -53 263 -567 -268 -1116 79 -428 -450
36-37 285 358 820 1223 2554 587 503 -50
37-38 3 171 1837 498 1224 -107 -32 -100
38-39 160 398 -1254 -252 1707 -411 61 42
1939-48 6430 7911 12402 8149 9363 4221 9357 10008
1948-49 988 418 -74 912 -1295 1487 1188 1165

49-50 656 856 1135 2570 147 272 512 435
1950-51 1739 153 581 722 806 926 704 842
51-52 116 265 -177 820 -103 1105 149 -250
52-53 -25 168 122 -676 136 -27 29 -196
53-54 379 165 1242 94 342 161 698 -41
54-55 -332 674 61 -37 -12 133 41 49
55-56 -446 71 1659 692 565 103 -213 21
56-57 429 618 728 191 951 452 264 862
57-58 -1 -202 -597 -215 129 -57 406 415
58-59 -73 284 155 363 237 -125 4 -58
59-60 -267 95 150 -512 470 340 -7 -71
1960-61 -178 199 164 427 392 366 142 4
61-62 -20 -38 1549 -41 925 385 153 78
62-63 188 656 -225 380 495 629 169 135

5.B
TIME SERIES DATA 177
Dpat Dpiot Dpiu Dpiu Dpi3t Dput Dpi 61

Dput
1921-22 -846 -1328 -600 -1445 -1680 585 -1074 -363

22-23 -121 -972 -834 -1071 -504 335 -198 -180
23-24 421 283 -2249 66 -461 324 -569 -7
24—25 -66 46 81 66 -10 211 -1143 -17
25-26 38 -886 -245 -466 -798 206 -35 -200
26-27 -4 -94 157 67 -496 498 -85 -81
27-28 64 -93 728 66 -20 474 -367 14
28-29 -1167 -108 -249 255 -247 274 180 -106
29-30 79 -1327 -1268 90 -283 179 -327 -210
1930-31 95 -718 -1495 -406 -675 175 -911 -329
31-32 -187 -1025 -2109 -1247 -1378 0 -52 -522
32-33 -63 -637 -427 -1108 -647 -175 -247 -388
33-34 85 36 -446 -709 -480 -269 -403 -164
34-35 -135 -299 -539 -308 -586 -370 -277 -248
35-36 77 -471 -590 -1830 -655 -385 -189 -417
36-37 -137 537 1727 849 927 -198 417 346
37-38 282 538 -755 495 124 0 -86 26
38-39 0 643 531 206 779 0 -183 99
1939-48 7704 10829 11168 11883 9340 0 4848 6906
1948-49 -114 429 527 -270 346 0 684 407

49-50 16 1046 1259 489 878 68 1090 528
1950-51 1373 1607 1485 942 1092 1165 1474 799
51-52 853 -1442 -1649 216 -114 149 943 293
52-53 85 -181 164 -196 -164 118 -298 -0
53-54 -8 87 -152 201 -72 1456 651 472
54-55 47 -140 94 133 66 101 484 365
55-56 -36 -575 -23 878 -4 336 386 261
56-57 1053 504 47 -604 -12 805 810 557
57-58 305 -186 -12 81 39 1084 66 474
58-59 29 -92 456 -21 170 146 15 144
59-60 4 171 231 93 193 1420 80 293
1960-61 -8 111 -165 81 92 501 21 226
61-62 1 153 33 49 -16 497 36 308
62-63 76 181 66 46 73 855 285 292
TABLE 5.18
QUANTITY LOG-CHANGES FOR SIXTEEN COMMODITY GROUPS
Dqu Dq-u Dqat Dqu Dqu Dqu Dqu Dqat
1921-22 723 1488 710 1354 1010 -359 -206 -13

22-23 150 247 -641 120 -650 53 -269 91
23-24 -198 -301 207 -391 1594 276 220 859
24-25 -300 -237 -1821 -388 -1607 -277 910 -1373
25-26 578 371 878 348 105 371 1722 1089
26-27 31 525 788 -64 1051 12 -298 164
27-28 -51 43 142 543 446 202 317 480
28-29 249 175 214 -706 396 -All -208 471
29-30 484 394 222 -328 96 699 690 376
1930-31 -142 513 556 111 -1174 -296 -588 -43
31-32 24 539 513 441 -1778 -360 -541 -101
32-33 -262 -700 761 -1037 9 -273 -123 305
33-34 -813 92 -823 -507 667 757 233 -344
34-35 274 -226 185 -42 -1601 207 637 -83
35-36 -457 -506 -212 133 1693 237 818 113
36-37 426 217 -294 -843 1110 3 -210 271
37-38 352 32 -907 225 -857 242 597 308
38-39 581 256 1411 271 -2345 392 110 855
1939-48 -3051 1527 -2656 -6362 1397 -1173 33 -732
1948-49 2141 -96 -1360 2763 -2469 -340 227 370

49-50 -324 -993 1052 1681 1274 -245 -423 1065
1950-51 -627 733 -272 643 -953 130 -434 -125
51-52 517 87 -14 -260 -354 122 137 828
52-53 159 337 284 1273 323 -59 498 339
53-54 841 556 -101 574 78 -256 -451 773
54-55 821 -227 -188 150 2199 -362 327 177
55-56 1262 146 -487 427 101 -251 1037 474
56-57 -1010 219 -224 -58 -849 -435 208 334
57-58 222 50 420 322 987 151 -212 -196
58-59 422 -51 495 -5 -446 292 245 187
59-60 647 106 251 517 769 -235 447 340
1960-61 949 342 629 164 672 -223 439 303
61-62 706 455 -650 399 -764 172 70 147
62-63 816 152 784 10 274 -94 660 -43
Note. All figures are to be multiplied by IQ-4.

Dq9t Dqiot Dqnt Dqi2t Dq13t Dqut Dqist Dq16t
1921-22 112 2599 -755 887 789 12 3210 -74

22-23 -782 -591 -1644 289 -1016 140 -66 -291
23-24 -390 -208 3044 -224 -141 91 1121 -388
24-25 112 -137 -41 -350 459 164 912 300
25-26 -59 594 792 146 655 114 474 -204
26-27 46 537 -130 650 777 156 33 27
27-28 267 176 76 306 676 149 530 372
28-29 266 395 410 19 114 38 567 288
29-30 -397 1269 990 -525 401 116 497 525
1930-31 -240 -343 -69 -626 24 69 729 -40
31-32 -1372 -424 -107 545 -17 66 -366 -337
32-33 -981 163 286 197 1078 113 29 -74
33-34 -1030 -1027 -780 -643 -71 105 -37 -121
34-35 -502 -230 430 -515 182 63 -116 -268
35-36 -542 543 257 2409 -218 105 254 289
36-37 317 20 -913 -836 610 82 263 -295
37-38 -24 -677 436 -1264 47 2 -14 154
38-39 402 819 2067 1601 123 29 546 370
1939-48 -401 -2208 1252 -5675 -690 732 -1349 1370
1948-49 -152 2078 -1412 990 -191 87 906 -603

49-50 -1008 144 -11 -9 1240 62 369 -68
1950-51 183 -1681 -583 841 -1524 173 313 -259
51-52 -1658 -72 768 -865 -672 96 -220 -203
52-53 696 369 -191 828 1919 110 585 456
53-54 946 984 1157 1521 2163 99 -176 275
54—55 398 1001 495 1449 1971 129 334 699
55-56 1185 1737 1276 266 1157 68 800 432
56-57 360 -881 -316 1913 -929 179 -296 -31
57-58 81 -321 293 -254 -1094 180 74 -217
58-59 866 297 22 225 2449 150 -70 359
59-60 450 830 1308 1347 1136 75 445 454
1960-61 1293 529 466 609 2087 152 327 96
61-62 1020 311 442 900 799 131 1056 186
62-63 1030 589 913 1268 1619 12 1157 238
and 5.18. It should be noted that the quantities are all per capita (obtained by
subtracting the log-change in the mid-year population from the log-change
in national expenditure).
5.C. The Decomposition of the Variance of the Value Share Changes
In this section we consider the variance analogue of the information de¬

composition of the value share changes that was analyzed in Section 5.7.
Our starting point consists of equations (5.1), (6.6) and (6.9), from which we
derive
(c i) t w* (Dw» - KY = nt + Kt + 2rt
i= 1
= n0, + K0t + 2T0r + X K (n9t + K9t + 2r9t)

9= 1
Using (6.7) and (6.10), we can write the sum of the first three terms on the
second line as follows:
G
I K l(DPgt - Dp,)2 + (DQgt - Dq,)2 + 2 (DPgt - Dp,)(DQg, - Dq,)]

9=i
= X K(DP,. + DQ„-Dp,-Dq,y
0=1
= X K(DP„ + DQ„-Dmt-S,)2
9= 1
where use is made of (4.5) in the last step. Furthermore, applying the
definition (6.3), we obtain
(C.2) DPgt + DQgt = * (Dpit + Dqit) =

Wit
V77* D (witmt)
W.gt VYgt
ieSg ieSa
Wit
= Dm, +
wf°Wit
ieS„
and hence:
(C.3) Wi, N
Dot + K0, + 2rn,
ot = w.gt
0= i ieS„
The right-hand side is the weighted variance of
(C.4) wn
w*Dw"
9 =
ieSa
5.C THE VARIANCE OF VALUE SHARE CHANGES 181
around its weighted average 5t, the allocation discrepancy. It should be noted
that the G expressions (C.4) need not necessarily be close to zero. When total
expenditure remains unchanged (Dmt — 0) but the expenses on all com¬
modities of some set Sg increase, we have Dwit> 0 for each ieSg, so that (C.4)
takes a positive value for this g which may be substantial. On comparing this
with (C.2) we find that it is a natural result. When the expenses on Sg increase,
we must have a positive sum of DPgt and DQgt; when Dm, behind the last
equality sign of (C.2) vanishes, one should add a positive value to Dm, in
order that the sum be equal to DPgt + DQgt. This value, (C.4), is a partial
weighted average of the value share log-changes Dwit, in the same way as
DPgt and DQgt are partial weighted averages of Dpit and Dqit,respectively,
and in all three cases we have the same weights.
We proceed to consider the last term on the second line of (C.l) and use
(6.8) and (6.11) for the following derivation:
V W-*
^(Dpit + Dqit - DPgt - DQgt)

gt
ieS„
.* r
W;, W it
D(witmt) — Dm,
W,gt w*DWjt
rrgt
ieS a jeSg
\2
W{,
Dwit -
W,gt K°Wjt
ieSa j ESg
where use was made of (C.2) in the second step. The variance decomposition
as a whole is therefore
G
g=l ieSg jeSg
The first term on the right is the variance of the partial averages of the value
share log-changes around their mean, the allocation discrepancy 5,. This is the
between-set variance. The term on the second line is a weighted average of
the separate within-set variances, defined in a completely analogous manner.
The decomposition (C.6) is therefore straightforward, but it has a much more
complicated appearance than the information decomposition (7.4).
CHAPTER 6
THE CONSUMER’S ALLOCATION PROBLEM
6.1. The Consumer’s Optimality Conditions and the General Form of Demand
Equations
The preceding chapter was devoted to an empirical analysis of the way in

which total expenditure on some set of commodities, and in particular the
change in this total expenditure, can be decomposed into certain elementary
components. There was no economic model behind the procedures followed;
they were based on some simple statistical considerations. The scene will now
change to the extent that there is a consumer for whom m is the given
(positive) amount which he has to spend on n commodities; we shall call m
his “income” for the sake of convenience. The prices pu ...,pn of the n com¬
modities are also given (positive) numbers for him. The consumer’s choice is
therefore subject to the budget constraint
(1.1) =m
i=l
where the q s are the quantities of the n commodities which he decides to buy.
This decision problem is the central problem of this chapter: How much will
the consumer buy of each of the commodities, given income m and prices
Pu C>r equivalently: What is the proportion w^pfijm of income
allocated to the zth commodity, / = 1, ..., n, given income and prices? It will be
noted that, since the p’s and m are taken as given, knowledge of the value
shares wt is fully equivalent to knowledge of the quantities qt.
Even though income and prices are assumed to be given from the con¬
sumer’s point of view, they are not really constant over time, since they are
subject to changing factors beyond the consumer’s control. When prices and
income change, the consumer will want to revise his position by adjusting
the q’s and hence the w’s. This means that the quantities become certain
functions of income and prices, to be called the demand functions. The form
of these demand functions is the main object of the present chapter. It is in
this respect instructive to consider the infinitesimal change in the zth value
182
6.1 THE CONSUMER’S OPTIMALITY CONDITIONS 183
share:
c/w; = d —dpt + — dqt — Pl-\l dm

m m m2
which can also be written in the following more convenient form:
(1.2) dWi = wtd (log + wtd (log q,) - wtd (log m)
The change in the zth value share is thus written as a weighted sum of three
changes. The first and the third deal with price and income changes. The only
thing which consumer demand theory has to say about these terms is that it
takes them as given (as exogenous, i.e., as determined “from the outside”).
It has much more to say about the second term, which is the quantity
component of the change in the /th value share. In fact, we shall formulate our
demand equations in such a way that this term is the dependent variable.
That will be our immediate goal. It will be reached in equation (4.12) of
Section 6.4.
Let us start with the case in which income and prices do not change at all.
The consumer’s choice problem concerns the n quantities qt subject to the
constraint (1.1). The conventional approach implies that the consumer is
supposed to maximize a utility function
(1.3) u = u(q1,q2,--;dn)
which measures his satisfaction when he buys qx units of the first commodity,
q2 units of the second, and so on. For example, in the case of three com¬
modities we may have:
(1.4) U(quq2,q3) = 20qy + 10q2 + 20q3 -\{q\ + dl + dl + 209^2)
where 0 is some constant between -1 and 1. [We shall use this example on
several occasions for illustrative purposes and will then adjust 0 according to
our needs ] The quadratic example (1.4) is of course much more restrictive
than the general theory that will be presented. However, it is appropriate to
mention at this early stage that a number of assumptions will be needed.
First we suppose that the q's can be varied continuously (which is restrictive
with respect to indivisible goods). Second, the utility function is supposed to
have positive first-order derivatives in the relevant region:
wvll
so that there is no saturation with respect to any of the n commodities. [The

184 THE CONSUMER’S ALLOCATION PROBLEM
derivative du\dq{ is known as the marginal utility of the z'th commodity.]

Third, we assume that the utility function has continuous second-order
derivatives, so that the cross-derivatives are pairwise equal according to
Young’s theorem:
d2u d2u
(1.6) i,j = 1
Sqtdq j dqjdqi
Fourth, the utility function is supposed to be such that, when it is maximized
subject to the budget constraint, the solution consists of a unique set of
positive values of the q's, for all values of income and prices which we shall
consider to be of interest.1 Therefore, the approach to be followed is com¬
pletely classical except for certain quasi-cardinal features which will be
discussed more fully in Section 6.9 below.
The conditional maximization of (1.3) subject to (1.1) is carried out
conveniently by means of a Lagrangian multiplier X. So we consider
which is differentiated with respect to each of the q's and then put equal to
zero. The result is
(1.7) T =kPi
°<h
On the left we have the marginal utility of the ith commodity, which thus
turns out to be proportional to the zth price in the consumer’s optimal point,
X being the proportionality constant. To interpret X we note that (1.7) can
also be written as du/d(piqi) — X, i — 1, ...,«, sincep{ is a constant from the
consumer’s point of view. This means that in the optimal point an additional
penny spent on whatever commodity leads to the same utility increase X. In
other words, if the consumer’s income m increases by one penny, he can raise
the utility level by X units (by spending that penny on any of the n com¬
modities). Therefore, the Lagrangian multiplier X is known as the “marginal
utility of income.” Note that its value in the consumer’s optimal point is
necessarily positive:
(1.8) X > 0
in view of (1.7), (1.5), and the positive sign of the p's.
1 For the case of zero values of the q s, see H. Wold, Demand Analysis (New York: John
Wiley and Sons, Inc., 1953), pp. 84-87.
6.1
THE CONSUMER’S OPTIMALITY CONDITIONS 185
On combining the n equations (1.7) with the budget constraint (1.1) we

obtain the n + 1 optimality conditions, which are exactly sufficient in number
to determine the n+ 1 unknowns of the problem: the n optimal quantities and
the associated marginal utility of income.1 When applied to the numerical
example (1.4) we find that the n + 1 =4 conditions take the following form:
*7 i T + P\2 = 20
(19) ^<7i + <?2 + p2X = 10
#3 + P3X — 20
Pidi + P2P2 + P3P3 = m
These are tour linear equations in the optimal quantities and the associated
1, to be written q?, 2°. The solution is
(p 1 -- 9p2) m + 0t (P)
Qo (p)
Q5
(p2 -
1
02 (P)
Qo (p)
(1.10)
(1- 92)p3m + 03 (P)
00 (P)
-(1 — 92) m + Mp)

Qo (p)
where £>;(p), * = 0, 1, 2, 3, is a quadratic form in p, the column vector of the

prices of all commodities, and LA(p) a linear form in p:
0o(p) = Pi + P2 + (1 ~ 02)pI — 29piP2

0i (p) = 10[2p2 + (2-0)pl- pvp2 - 2p3p3 + 26p2p3]
(1.11) Q2(p) = 10[pi + (1 - 2e)p23 - 2PlP2 + 29PlP3 - 2p2p3]
Q3 (p) = 10 [2p\ + 2 p\ - 4 ePlp2-(2-9)Plp3-(l-29) p2P3]
L,(p) = 10[(2 - d)Pl + (1 - 2.9)p2 + 2(1 - 92)p3j
Our comments on these results are as follows:
1 The n + 1 optimality conditions (1.1), (1.7) ensure that the utility function is stationary
subject to the budget constraint, but it does not guarantee a maximum (rather than a
minimum or a saddle point). This guarantee can be formulated in terms of alternating
signs of certain principal minors; see, for example, J. R. Hicks, Value and Capital (2d ed.;
Oxford: Oxford University Press, 1948), Appendix. It will be assumed below that the
Hessian matrix of the utility function is negative definite, which is a sufficient condition
for a maximum.
(1) Equation (1.10) specifies the q’s to be bought by a consumer whose

utility function is (1.4), when income is m and prices are p. The equation thus
shows immediately how the three optimal quantities vary when income and
prices vary. That is, the equations for q° are demand equations, the general
form of which is
(1.12) = qi(rn,p1,...,pH) i =
or in vector notation
(1.13) q° = q(m,p)
where q° is the column vector of then optimal quantities, while q( ) indicates

its dependence on the determining factors. These factors are income m and
the prices py, which are the coefficients of the budget constraint subject
to which the utility function is maximized.
As stated above, the demand equations are the main object of the present
chapter. We observe that they have the form of a ratio in the special case
(1.10). The denominator is a quadratic form in the prices, which is positive
when |0|<1. The numerator is another quadratic form in the prices plus a
bilinear form in income and prices. It follows that, if income and prices
increase or decrease proportionally, the optimal quantities are not affected.
The last conclusion holds quite generally, not only for the quadratic utility
case (1.4):
(1.14) q(/cm,/cp) = q(m,p) for any k > 0
This follows directly from the fact that the demand functions are obtained by
maximizing the utility function (1.3), which is completely independent of in¬
come and prices and hence of k, subject to the budget constraint (1.1), which
is not affected when its coefficients on the left and on the right of the equality
sign are all multiplied by k.
(2) Example (1.10) shows that, in general, each q(- depends on all prices
Pi,p2, ■■■, not only on the pricept of the same commodity. The example shows
also that it is not necessarily true that the consumption of each commodity
increases when income increases. Take py=2p2, 6 = .7, in which case q2
varies negatively with m. The second commodity is then said to be inferior.
Note further that in the quadratic utility case, given fixed values of all prices,
each optimal quantity varies linearly with income. Hence, if a commodity is
infeiior, its optimal quantity according to (1.10) becomes negative when in¬
come is sufficiently large. Similarly, if a commodity is not inferior, its optimal
6.1 THE CONSUMER’S OPTIMALITY CONDITIONS 187
quantity according to (1.10) may be negative when income is very low. All
this simply means that the solution (1.10) is only applicable to a limited range
of income and prices.1
(3) The marginal utility of income in the consumer’s optimal point is also
a function of income and prices:
(1-15) A° = A(m,p)
We conclude from (1.10) that in the quadratic utility case X° has the form of a
ratio and that its denominator is the same as that of the q?. Its numerator is,
however, linear in income and prices. Hence the marginal utility of income is
homogeneous of degree — 1 in these variables, which is again generally true:
(1-16) X(km, /cp) =-X(m,p) for any k >0

k
This follows from (1.14) and (1.7). When income and prices are all multiplied
by k, the optimal quantities are not affected, so that the same is true for the
left-hand side of (1.7). Since pt on the right becomes kpt, the multiplier X
(to be written ?S> in the present notation) becomes X°/k.
(4) In equation (1.10) the marginal utility of income in the optimal point
is a decreasing function of income. [Remember that we assume \9\<1
throughout.] A similar statement can be made on the marginal utilities of the
three commodities:
du
t— = 20 - qi- 0q2
dq i
du
(1.17) — = 10 -Oq,- q2
dq2
Hence the marginal utility of the zth commodity is a decreasing function of qh

i = l, 2, 3. We also conclude from (1.17) that in this special case the marginal
1 It is possible to take account of the nonnegativity constraints. The procedure then

amounts to quadratic programming in the quadratic utility case. See H. S. Houthakker,
“Sur la forme des courbes d’Engel,” Cahiers du Seminaire d’Econometrie, No. 2 (1953),
pp. 59-66.
utility of the third commodity depends only on the corresponding quantity

(q3) and that the marginal utilities of the first two commodities depend only
on the first two quantities (qx and q2). Such special cases will play a prominent
role in a later part of the analysis (see Section 6.4).
6.2. The Fundamental Matrix Equation of Consumer Demand Theory
The solution (1.10) specifies a complete set of demand equations for the
utility function (1.4). This special case is quite restrictive, of course, and even
for this simple case we cannot really claim that we have succeeded in showing
the effect of price changes on the optimal quantities in a way which is
intuitively easy to understand. We shall be more successful in this respect if
we concentrate on the effect of infinitesimal price changes. That will be our
goal in this and the next section.
We return to the budget constraint (1.1), which can be written as p'q = m in
vector notation. Suppose that income changes by an infinitesimal amount dm
and that the prices remain unchanged. Then the optimal quantities will
change from q° to q° + dq. The budget constraint has to be satisfied both
before and after the change:
p'q° = m
p' (q° + dq) = m + dm
By subtracting the first equation from the second we obtain P dq = dm or

equivalently p'(<3q/Sm) = 1, where dq/dm is the vector of first-order derivatives
of the demand functions (1.13) with respect to m. It will prove convenient to
introduce a short-hand notation for such derivatives. We shall write qm for the
vector of income derivatives and Qp for the matrix of price derivatives of the
demand functions, both evaluated at the prevailing levels of income and
prices:
dq i dqi dqi dqx
dm dpi dp2 dpn
dq 2 dq2 dq2 dq2
dm QP = dpi dp2 dpn
dqn dq„ dqn dqn

dm JPi dp2 dpn_
Hence the result just derived can be written as p'qm = l. Note that this
equation is actually obtained by differentiating both sides of the budget
constraint p'q = m with respect to m. [The derivative of a function/( ) is
6.2
THE FUNDAMENTAL MATRIX EQUATION 189
obtained in the same way: subtract/(x) from f(x + dx) and divide the
difference by dx.] We may proceed in a similar way for prices. It is easily
verified that differentiation of p'q — m with respect to m and p gives
(2-2) p'qm = l P'Qp + q'= 0
The proportionalities (1.7) are handled in the same way. They also hold
both when income is m and when it is m + dm. So we differentiate both sides
of (1.7) with respect to m :
d2u dq- dX
= PiW~ i = 1,..., n
dm
j=i
and also with respect to pk:
'P d2u dqj dX

if k # i
L
j= 1
ddidqj dpk
= Pi x~-
dPk
dX
— X + pi -— if k = i
dpi
which takes the following form in matrix notation:
(2-3) Uq,„ — Amp UQp — p2p + 21
where U = [d2u/dqidqJ] is the Hessian matrix of the utility function, while
8X
dp i
dX
(2.4)
dm
dX
_d Pn_
are first-order derivatives of X° as defined in (1.15) with respect to income and
prices. We can now combine (2.2) and (2.3) in the following partitioned form:
O
O
k-H
~U p" q Qp"
(2.5)
1_
Li
■o '
q0,J
1
i
y o_ -
3
i
where the superscript 0 on the right indicates that we take the optimal
quantities and the corresponding value of the marginal utility of income.
Equation (2.5) will be called the fundamental matrix equation of consumer

demand theory.1 On the left we have the Hessian matrix of the utility
function, bordered by prices. This bordered Hessian matrix is postmultiplied
by the gradients of demand functions (1.13) and of minus the marginal-utility-
of-income function (1.15). Both matrices have n +1 rows and columns. On
the right we have a matrix of the same order of which the first column
consists of n zeros, followed by a unit element. The other n columns consist
of an n x n unit matrix, multiplied by the marginal utility of income in the
optimal point, and of the n optimal quantities apart from sign.
We are particularly interested in the derivatives of the demand functions,
for which we can obtain explicit expressions by solving the fundamental
matrix equation. For this purpose we need the inverse of the bordered
Hessian matrix, which can be written as
(p'u- 1P)u_1 — (u_ 1p)(u~ ‘Py u-y

1
p'UHp
-1
under the assumption that U is nonsingular. The derivatives of q° and 2°

with respect to m and p can then be found by straightforward partitioned
multiplication. The result is:
(2.6) - l.
q" pr'pu
(2.7) Q = A°U_1 -
P U
A ,p (u-‘P)(u-y-- 1 irW'
pU'p
(2.8) A,„ —
p'u-'p
A°
(2.9) 4 =
pU ‘pU P p'U'V
We can simplify these results by writing A,„ for the reciprocal of the quadratic
form p'U_1p in accordance with (2.8), so that qm can be written as
Am U 'p. Furthermore, we note that the vector U_1p occurs three times
1 The equation in the form presented here was first given by A. P. Barten in “Consumer
Demand Functions under Conditions of Almost Additive Preferences,” Econometrica,
Vol. 32 (1964), pp. 1-38. The present exposition, in particular the formulation of the
demand equations, is largely based on H. Theil, “The Information Approach to Demand
Analysis,” Econometrica, Vol. 33 (1965), pp. 67-87.
6.3 INCOME EFFECT AND SUBSTITUTION EFFECTS 191
in (2.7) and once in (2.9), so that additional simplifications are possible by

substituting (1/Am)qm. This leads to
(2.10) qm = 2„,U 'p
(2.11) Qp = x°U-‘
(2.12) K = ~ Am - 2,„q°
which, together with (2.8), will be the solutions of the fundamental matrix
equation with which we shall work from now on. Note that the derivatives of
the demand equations are subject to a number of constraints:
(2.13) p'qm = i
(2.14) 2°U_1 is a symmetric matrix
(2.15)
The first constraint is part of the fundamental matrix equation and follows
directly from the budget constraint. It states that the weighted sum of the in¬
come derivatives is 1 if we use the corresponding prices as weights. The
second constraint deals with the first component of the matrix of price
derivatives, see (2.11); it follows directly from the symmetry of the Hessian
matrix, see (1.6). The third constraint deals with the same component and
states that, if we postmultiply the matrix 2°U~1 by the price vector, we obtain
the vector of income derivatives apart from a scalar multiplier.
6.3. The Income Effect and the Substitution Effects of Price Changes
The solution of the fundamental matrix equation, in particular equation

(2.11), enables us to obtain a better insight into the various ways in which
price changes affect the optimal quantities. Let us consider (2.11) in scalar
terms:
(3.1) =
dpj dX/dm dm 8m dm
where uij is the (i,j)th element of U_1. We shall discuss the right-hand side
term by term, starting with the third, which is the income effect of a price
change dpj on the optimal value of the ith quantity. If pj increases by dpp the
batch of commodities q° which was originally optimal becomes more ex¬
pensive. An income compensation q]dp} is needed in order to enable the
consumer to buy q°. He does not receive this compensation, however, so

that this aspect of the price change dpj is equivalent to a negative income
change dm — — q^dpj. The effect of this income change on q® is the product
of dpj and — <7°(dqjdm), which is the third term on the right of (3.1); hence
the name “income effect of dpj on q°” of this term. Note that the income
effect of any price increase on q° is always negative except when the ith com¬
modity is inferior or when qf does not vary with m at all, in which cases the
effect is positive and zero, respectively.
To explain the other two terms of (3.1) we imagine that the consumer does
receive the income compensation q]dpj which enables him to buy the old
batch q° in spite of the price change dpj. So he faces two simultaneous changes,
dpj and dm = q°dpj, the joint effect of which on qf is
(3.2)
This joint effect is not zero in spite of the fact that the consumer can afford to
stick to the old quantity pattern q°. That is, the right-hand side of (3.2),
which is equal to the sum of the first two terms of the decomposition (3.1), is
in general different from zero. This is due to the fact that the change dpj
affects the price ratios. The right-hand side of (3.2) is the well-known substi¬
tution effect, which will be called here the total substitution effect, because it
consists of two terms. The second is
2° dqt dqj
(3.3)
dX/dm dm dm
which will be called the general substitution effect of the price change dpj
on qt. It is related to the effect of a change in 2° in the same way as the in¬
come effect is related to a change in m. Since the marginal utility of income is
the Lagrangian multiplier associated with the budget constraint, this part of
the total substitution effect can be regarded as describing the general compe¬
tition of all commodities for the consumer’s dollar; hence the name “general
substitution effect.”1
The precise interpretation of (3.3) can be given as follows. When Pj in¬
creases by dpp this affects the proportionality of marginal utilities and prices:
(3.4)
1 The term is due to H. S. Houthakker; see “Additive Preferences,” Econometrica

Vol. 28 (1960), pp. 244-257.
6.3 INCOME EFFECT AND SUBSTITUTION EFFECTS 193
see (1.7). This is trivially true for the yth proportionality which involves pj
directly, but it is also true for the other n — 1 proportionalities, because X° is
affected by the change dpy We have in view of (2.12):
dX dX
(3.5)
dPj dm dm
and conclude that the X change induced by dpj consists of two parts. The
second is -q°(dX/dm) multiplied by dpp which is clearly the income effect
of dpj on the marginal utility of income. If we would give the consumer an
income compensation dm = q)dpp so that he would be able to buy the old
batch q° if he chose to do so, that second term would be removed. Suppose
now that we decide to give the consumer another income compensation, this
time to remove the first term on the right of (3.5) as well, so that the marginal
utility of income is restored at the old level in spite of the change dpp Given
the proportionalities (3.4), this means that all marginal utilities remain un¬
changed (both of income and of the commodities) except the marginal
utility of they111 commodity whose price is changed. Now this second income
compensation is
dpj
dm
(3.6) dm
dX
dm
This follows directly from the fact that a unit increase in income leads to
dX/dm units increase in its marginal utility, and that the first term on the right
of (3.5) indicates that the number of units to be compensated is X°(dqj/dm)
multiplied by dpj.
We now return to the second term of the decomposition (3.1). Clearly, this
is precisely the effect on the ith quantity of the income compensation (3.6)
apart from sign.1 * On the assumption that the marginal utility of income is a
decreasing function - which holds for our example (1.4) when |0|<1, see
(1.10) - this general substitution effect is positive in the normal case when
both q° and q°j vary positively with income. This means that an increase in
the yth price is a stimulus for the consumption of the ith commodity as far as
the general substitution effect is concerned. The effect is also positive when
1 The difference in sign is simply due to the fact that the consumer does not really receive
the income compensation (3.6). The only change which he faces in (3.1) is dpj.
both commodities are inferior; it is negative when one of them is inferior but
the other is not.
Finally, we have the first term X°u'j in the right-hand side of (3.1), which
will be called the specific substitution effect. It is immediately evident that this
effect is indeed specific for each (/, j) combination, since the indices i and j
cannot be separated in such a simple multiplicative manner as is true for the
general substitution effect. When the marginal utilities of all n commodities
depend only on the corresponding quantities [which applies to (1.17) if we
put 0 = 0], both U and U_1 are diagonal. The specific substitution effects are
then all zero except for the relationship between qf and its own price pt. We
shall consider such special cases in more detail in the next section. Here we
confine ourselves to mentioning that, from now on, the Hessian matrix of the
utility function will be supposed to be negative definite. This implies that the
marginal utility of each commodity is a decreasing function of its own
quantity and that the matrix 2°U_1 of the specific substitution effects is also
negative definite.1
6.4. Demand Equations in Infinitesimal Changes; Preference Independence and

Block-Independence
We shall now formulate demand equations, in the first instance in infini¬

tesimal changes, and we shall use as the dependent variable in the ith equation
the quantity component of the value share change dwt as given in (1.2):
(4.1) w fi (log qf) = - dqt

m
The differential dqt is determined by the income and price differentials:
, dqt V dqL
dqi — - dm + ) — .dPj
dm U dPj
0Pj
j=
7 = 1i
n
dqt _ dQi dqj _ dpi Q

dm + X °u,J —
dm dPj
dXfim dm dm dm^j
7=1
wheie use is made of (3.1). On multiplying both sides by p^m in accordance
1 The negative definiteness of the Hessian matrix is a sufficient condition in order that the
optimality conditions (1.7), (1.1) imply a maximum, not a minimum or a saddle point.
See footnote 1 on page 185.
6.4 DEMAND EQUATIONS IN INFINITESIMAL CHANGES 195
with (4.1) we obtain:
(4.2) Wid (log qt) = -pd

m cm
2° dqt dq - dq(
A°uij - dPj
m dX/dm dm dm dm ^'
j= i
The various terms on the right will now be simplified. The income term can
be written as follows:
(4.3)
Pi dch ,
-
d(Pidi),n ,
dm = —--rt(logm) = md(logm)
,, ,
m dm dm
where
(4.4)
diPidi)
&=■ i = 1, ...,n
dm
is the marginal value share of the ith commodity. The price term of (4.2) can be
written as
PiPj . J_dlLdll Jliq° d (log pj)

m dXjdm dm dm dm 1_
J®1
so that the coefficient of d( log pi) is
*°PiPju‘J 51 m\~1 d(jpiqi) d(pjqj) d(p^) Pjq0j

= Vij - (Phihj ~ PiWj
m dm 2° cm Cm cm m
where
^PiPjU' l
(4.5) Vij 4> =
m d (log 2)
d (log m)
On combining this result with (4.2) and (4.3) we obtain:

n
(4.6) w,d(logqi) = ptd(logm) + £ (vy - 4>PiPj ~ d(logpi)

7=1
Before simplifying this equation still further we shall interpret the p’s, v’s
and 0 and formulate the constraints which they satisfy. First, the marginal
value shares add up to 1:
(4.7) X>,= 1
i= 1
This follows from (2.13). Second, the v’s form a symmetric matrix because
of(2.14):
(4.8) vu = vJt i,j=l,...,n
and this matrix is negative definite when the Hessian matrix U is negative
definite. This follows from the definition (4.5):
(4.9) M— PU^P
v m
where P is the n x n diagonal matrix of prices. Furthermore, </> is defined

in (4.5) as the reciprocal of the income elasticity of the marginal utility of in¬
come. It will be called the income flexibility for short.1 Note that we have
n
(4.10) Z vy - i = l,..., n
j=i
which follows directly from (2.15):
Pi /,0 ij\ Pi ^Qi i

VU =
- ) (X uJ)pj = — — J- =
^ m l_j m Xm cm
J= i
Note finally that summation of both sides of (4.10) gives
(4.11) 0=ZIvy<O
i=1j=1
where use has been made of (4.7). The inequality sign in (4.11) is based on the
negative definiteness of the matrix of v’s.
We shall now write the demand equation (4.6) in a much more elegant
form. We start by noting that the third component of the price term
- MiWjd (log pj)
is the income effect of the price change d (log pi) on the quantity component
1 There is unfortunately no uniform terminology in this respect. Houthakker (loc. cit.)

defines the income flexibility as the product of our cf> and income. Its reciprocal measures
the relative change in the marginal utility of income resulting from a small absolute (not
relative) income increase. R. Frisch uses the term money flexibility, which is defined as
the reciprocal of our cf> (i.e., as the income elasticity of the marginal utility of income).
See “A Complete Scheme for Computing All Direct and Cross Demand Elasticities in a
Model with Many Sectors,” Econometrica, Vol. 27 (1959), pp. 177-196.
wid 0°g <k) of the change in the ith share. We combine it with the income term,
which gives:
n
Hi [d (log m) - X Wjd (log pjJ]

j= 1
The general substitution elfect of the change in the kth price (the second part
of the price term) is
n
~<t>kikkd (logp*) = ~ X VfjHkd (log Pk)

j= i
see (4.10). We combine this expression with the specific substitution effect
(the first part of the price term):
X
j =1
vu 0°g Pj) ~ X
k=1
M(logpfc)]
so that (4.6) takes the following form:
(4.12) wtd(log qt) = ^ [d(log m) - X wkd(ôgpJ]

k =1
n n
+ j=I1 vl7[d(logp,.)- k=X1 MOogp,)]

This is the demand equation in infinitesimal changes which we shall use from
now on. The left-hand side is the quantity component of the change in the
fth value share. The first term on the right is the marginal value share of the
Ith commodity multiplied by the logarithmic income change, from which a
weighted average of the logarithmic price changes is subtracted. The weights
are the value shares vtq, ..., vv„; the subtraction procedure implies that we
combine the income effect of the price changes with the income change. The
term on the second line gives the total substitution effect of the price changes.
It also consists of two parts, the first of which, IVjjd (log Pj), represents the
specific substitution effect and the second the general substitution effect. The
latter effect amounts to the subtraction of a weighted average of individual
logarithmic price changes, just as the income effect on the first line, but the
weights are now the marginal income shares. These weights do add up to 1,
see (4.7), but they are not necessarily nonnegative. When the lcib commodity
is inferior, its price has a negative weight in this average.
The total number of h’s and v’s in the n demand equations (4.12) is
n(n+1). They are subject to the constraints (4.7), (4.8) and (4.10). If we
subtract the number of constraints from the number of /Ts and v’s and add 1
for the additional (f) which appears in (4.10), we obtain
n(n + 1) — 1 — in(n — 1) — n + 1 = \n(n + 1)

198
This is the number of “free” coefficients to be estimated if we specify the

demand equations in the form (4.12). [The actual specification should be in
finite rather than infinitesimal changes, but this will be considered in the next
section.] This number is very large unless n is small. It is reduced substantially
when we assume that the utility function can be written as the sum of n
functions, each involving only one of the q's:
n
(4.13) u(qu...,qn) = X
i= 1
This is the case of additive or independent preferences} There is additivity in

the sense that the utility function can be written in additive form, and
independence in the sense that the marginal utility of each commodity is
independent of the quantities of all other commodities. The latter feature
implies that the Hessian matrix of the utility function is diagonal. We can
thus conclude from (4.9) that vy = 0 holds whenever i =£j. Hence the term on
the second line of (4.12) is then simplified to one single price term:
n
V/i [d (log Pi) - XM (1 * * * * * * * * lo§ PkJ]

it= i
The symmetry condition (4.8) is now irrelevant. Constraint (4.10) is simplified

to
(4.14) = i = l,...,n
which amounts to a proportionality of the own-price coefficients and the

marginal value shares, the income flexibility being the proportionality
constant. The total number of p?s and v’s, disregarding the zero off-diagonal
v’s, is now In. If we add 1 for f and subtract 1 for (4.7) and n for (4.14), we
1 See the articles by Houthakker and Frisch quoted above. Reference should be made
to several other and related attempts to obtain manageable systems of demand equations:
R. H. Strotz, “The Empirical Implications of a Utility Tree,” Econometrica, Vol. 25
(1957), pp. 269-280; R. Stone, “Linear Expenditure Systems and Demand Analysis;
An Application to the Pattern of British Demand,” The Economic Journal, Vol. 64 (1954)
pp. 511-527; C. E. V. Leser, “Family Budget Data and Price Elasticities of Demand,”
Review of Economic Studies, Vol. 9 (1941-42), pp. 40-57; I. F. Pearce, “An Exact Method
of Consumer Demand Analysis,” Econometrica, Vol. 29 (1961), pp. 499-516; A. P. Barten,
“Consumer Demand Functions under Conditions of Almost Additive Preferences,”
Econometrica, Vol. 32 (1964), pp. 1-38. Reference should also be made to theoretical and
empirical work by W. H. Somermeijer (in Dutch); for a brief description see H. Theil,
“Some Developments of Economic Thought in the Netherlands,” American Economic
Review, Vol. 54 (1964), pp. 34-55.
obtain n for the number of free coefficients. This is a very substantial re¬
duction compared with %n(n+\).
The price to be paid is, of course, that a very restrictive assumption is intro¬
duced. One of the implications is that inferiority is excluded, since iit = vu/4>
in view of (4.14), and hence /i;>0 because vih </>< 0. One may decide to 'go
only halfway by assuming that the n commodities can be divided into G sets
Sx,...,SG such that (1) each commodity belongs to exactly one S , g = 1,..., G,
and (2) the utility function can be written as the sum of G functions, each
involving only one set:
(4J5) «(«i> •••»«»)= £ ug(qg)

9= 1
where qg stands for the column vector of those q’s that belong to Sg. The
marginal utility of the itb commodity will then depend on qj only if i and j
belong to the same set. Hence, when the commodities are arranged in a
suitable order, the Hessian matrix becomes block-diagonal:
UJ 0 ... 0~
0 U2 ... 0
(4.16)
0 0 ... UG
where \Jg is the Hessian matrix of ug( ), g = l,..., G. This is the case of block-
independent preferences. It implies that U-1 is also block-diagonal (with
blocks Uj"1, ..., UG *), so that the same applies to [vi7]. Hence the second line
of (4.12) becomes „
Z vtJ[d(logpj)~ ^ M0ogpfc)]
j e Sg k= 1
where Sg is the set that contains the /th commodity. We conclude that the
price term of (4.12) deals, as far as the specific substitution effect is concerned,
only with the prices of the group to which the /th commodity belongs. For
constraint (4.10) we can now write:
(4.17) YJvij = 4>Hi ieSg,g = l,...,G

J E sa
The total number of /ds and nonzero v’s is now
G G
n + X ng = £ ng(ng + 1)
9=i 0=1
The total number of constraints (4.7), (4.8) and (4.17) is
1+i Z na(n9 -l)+« = 1+ il na(n9 +0

0=1 3=1
If we subtract the latter number from the former, and add 1 for the additional
parameter <p, we obtain a total of
1 1
2 I ng(ng+l) = 2
free coefficients. Given the number n of commodities and the number G of

sets, this total number of free coefficients takes its minimum value when all
sets have the same number of commodities, ng = n/G, g = 1, ..., G. [This is
easily proved by minimizing Ing subject to Ing = n.] The minimum value is
\n(l + n/G), which varies between n (for G = n, the case of “complete”
independence) and $n(n +1) (for <7=1, the case of a full matrix [viy-]).
The utility function (1.4) is an example of block-independence. Its Hessian
matrix is
- 1 -e O'
U = - e - i o
o o -l
so that the first two commodities form one set and the third another, inde¬
pendent, set. Preferences are completely independent in the special case 0 = 0.
Note that block-independence does not exclude inferiority for those com¬
modities that are part of a set which consists of more than one commodity.
[A counter-example was provided in the discussion under (2) at the end of
Section 6.1.] If a set consists of only one commodity, this good cannot be
inferior because of the negative definiteness of the Hessian matrix of the
utility function.
6.5. Demand Equations in Finite Changes; Real Income and Two Kinds of
Relative Prices
We now return to the notation of the previous chapter (Section 5.4) and
suppose that we have time series observations on the value shares wit and the
log-changes Dpit, Dqit, Dm, for n commodities in a number of successive time
periods. The demand equation (4.12) is taken as a starting point for the
following equation in finite changes:
n
(5■ 1) w*,Dqit = pfim, + £ vuDp'jt i = l,...,n

j= i
where w* is the average of the successive value shares wit and wit_1 [see
equation (4.2) of Section 5.4], while Dm, is the log-change in real income:
(5.2) Dm, = Dm, — Dp,

6.5 DEMAND EQUATIONS IN FINITE CHANGES 201
which is obtained by subtracting from Dm, the log-change in the cost of living
price index, defined as follows:
(5.3) Dp, = £ w?,Dpu

i= 1
Furthermore, Dpj, is the log-change in the relative price of theylh commodity:
(5-4) Dp'jt = Dpjt - Dp\
which is obtained from Dpj, by subtracting the log-change in the marginal

price index, defined as
(5-5) Dp= £ p{Dpit

i= 1
In the special case of preference independence we can simplify (5.1) as follows:
(5-6) w*Dqit = pfim, + vuDp'it i=l,...,n
and in the case of block-independence:
(5.7) w*Dqit = PjDrn, + £ vipPjt ieSg; g = 1,..., G

jsSg
It is immediately seen that the log-change Dpt in the cost of living index as
defined in (5.3) is precisely the same as Dpt of equation (4.3) of the previous
chapter. The marginal price index is new. We shall consider both in Sections
6.7 and 6.8, which are devoted to the economic theory of price index numbers.
At this stage we confine ourselves to the following four remarks, which serve
to clarify the demand equation (5.1) and the concepts introduced in (5.2)
through (5.5):
(1) The variable on the left of (5.1), w*Dqit, is nothing else than the
contribution of the ith commodity to the volume index Dqt as defined in
equation (4.4) of the previous chapter. It is the finite-change approximation
to H’i<i (log #,) of (4.12). The relation between the two can be explained
conveniently by means of the following lemma. Let/( ) be some real-valued
function of a column vector of arguments which takes the value x in the
present period and y in the preceding period. We assume that all desired
derivatives exist and write/* and fy for the gradients evaluated at the levels of
the present and the previous periods, respectively. Consider then:
(5.8) / (x) ~ / (y) ~ HA + fy)' (x - y)

That is, the change in the value of the function is approximately equal to the
inner product of the change in the arguments and the average of the two
gradients. The lemma states that this is a local quadratic approximation in
the sense that the error is of the third order in the elements of the difference
vector x — y. It will be referred to as “lemma (5.8)” in the sequel and it is
proved in the Appendix of this chapter (Section 6. A).
We shall apply this lemma to the change Awit in the itb value share, it being
understood that this share is regarded as a function of the logarithms of ph
and m. So we have
Vi
log Pi, log Pi,,-1
x= log qit y = log?i,,-i
_log m,_ _log mr_! _
The three derivatives of w, with respect to its logarithmic arguments are:
dw, d\Vi d\Vi

- = VV- - = XV: ---- = — XV:
<3 (log Pj) ‘ d(log q,) ^ (log m)
so that the average of the two gradients in (5.8) is in this case a column vector
whose successive elements are w*, wf„ — w*. Hence lemma (5.8) amounts in
the present case to
(5.9) Awit « w*Dpit + w*Dqit - w*Dm,
On comparing this result with (1.2) we conclude that it is still possible, even
in the finite case, to decompose the change in a value share into parts
attributable to the changes in price, quantity and income, and that the left-
hand variable of (5.1) is the quantity component of Awit. In contrast to (1.2)
the present decomposition has an approximate character. However, the
approximation error is of the third degree in the log-changes Dpit, Dqit, Dmt,
so that it will usually be negligible when successive years (say) are compared.
(2) When we sum both sides of (5.9) over i (from 1 through n) we obtain
zero on the left. The right-hand side becomes Dp, + Dq, — Dmt, which is the
allocation discrepancy d, according to equation (4.5) of Section 5.4. We meet
the same 5t when we add all n demand equations (5.1), which is shown as
follows. The right-hand side becomes after summation:
D,Th + i (i Dp'jt
j= 1 \i=l /
rt / n
= Dm, + <t> £ pj ( DPj, - X hkDpkt

j= 1 \ k= 1
= Dm,
where use is made of (4.8) and (4.10). Since the sum of the left-hand variables
of the demand equations is Dq„ the implication is Dq, — Dm,. This is a
contradiction, since Dq, = Dmt + <5r. However, it will be obvious that the
demand equations do not hold exactly anyhow. If we recognize this fact by
adding a disturbance to the right-hand side, the implication is simply that the
sum of all n disturbances is equal to 5,. This subject will be discussed in more
detail in the next chapter (Section 7.1) when the statistical structure of the
disturbances will be considered.1
(3) We shall work with constant parameters ph vtj, which are supposed to
be subject to the constraints (4.7), (4.8) and (4.10). This constancy is re¬
strictive, of course. It implies, as far as pi = d(piqi)/dm is concerned, that the
optimal quantities are linear functions of income. Although this is probably
not too serious when real income and relative prices are subject to moderate
changes, it should be realized that equation (5.1) - or, for that matter, any
other form of demand equation - is a Procrustean bed which fits empirical
observations imperfectly. The main justification of (5.1) is its simplicity. We
shall meet the /t’s frequently in what follows; the analysis would be much
more complicated if they were not constant but dependent of income and
prices. Also, the symmetry condition (4.8) on the v’s would be much more
difficult to handle if they were not constant. For example, demand equations
with constant elasticities present their problems in this respect.2
Since our starting point was formulated in terms of first-order effects
(infinitesimal changes), the implications of the demand equations should also
be confined to first-order changes. It is interesting to see what happens if this
advice is not followed.3 Take the income derivative dqjdm = pjpf of the
zth optimal quantity. Since it does not depend on pj,jî, when /q is fixed,
we have
(5.10)
1 Another subject of interest is the relationship of Dqt and Dmt with the economic theory
of quantity index numbers. This theory, which is closely related to the price index theory
that will be described in Sections 6.7 and 6.8 below, is considered in the Appendix of this
chapter (Section 6.B).
2 Write Sij for the elasticity of the 7th quantity with respect to they'111 deflated price; then
condition (4.8) amounts to w>i£y = WjEji. If the e’s are constant, the ratios Wi/wj must be
constant too. Given that the value shares add up to 1, they must then be individually
constant, which is not realistic. Reference is also made to Section 8.6 below, which shows
that the corresponding problems of the demand equations for production factors cannot
be solved in the same simple manner.
3 The author is indebted to Professor Daniel McFadden of the University of California
at Berkeley, who made him aware of this point.
Let us compare this with the symmetric cross-derivative. For that purpose we
multiply both sides of (4.12) by m/ph which gives
djk
Pk
dqi="i{dm~lqidp)+l'i7{dpj Pk
k= 1 j =1 k= 1
We conclude that the first-order derivative of q° with respect to pp j=£i, is
Hi Pidj , vum v „
-— =-1-2-> Vih
Spj Pi PiPj h= 1 PiPj
(pPiPj)
Pi
where use has been made of (4.10). If we now differentiate with respect to m,
we obtain
dq: 1
(dqA =_ + (Vij - <t>PiPj)
dm\dpjj Pi dm ptPj
- (1 + (p)PiPj
i*j
PiPj
If we then equate this derivative to (5.10), we find that the following restriction
is implied:
(5.11) Vy = (1 + (fiflillj i #j
Flence all cross-price coefficients are determined by the marginal value shares
as soon as the income flexibility is given. Moreover, preference independence
(v;j. = 0, zVi) implies that the flexibility should be —1.
These results show that the assumption of constant marginal value shares
is a simplifying approximation. In general, the q’s and also the v’s should be
regarded as functions of income and prices. [If that is the case, the left-hand
derivative in (5.10) is no longer zero.] However, the price of such a generali¬
zation is too heavy, since it would complicate the analysis considerably. We
shall therefore stick to the assumption of constant p’s and v’s and regard
equation (5.1) as a linearized form of the “true” equation for w*Dqit (line¬
arized in Dm, and Dp'jt).
(4) Given that our starting point in Section 6.1 was the decomposition (1.2)
of the change in the ?th value share, an obvious question is what this de¬
composition looks like when we replace the quantity component by the right-
hand side of the demand equation, so that dw, is expressed in income and
price changes only. We shall answer this question in terms of finite changes,
using (5.1) and (5.9). So we obtain:
n
Awit « w*Dpit + HiDm, + £ VijDp'Jt - w*Dm,

j= i
The right-hand side contains both absolute and relative prices (Dpit and Dp),)
and also both money income and real income (Dm, and Dm,). This is
remedied easily:
w*DPit ~ w*Dm, = w*Dpit - w*Dm,
where
(5.12) Dpt, = Dpit - Dp,
Hence the decomposition becomes
(5.13) Awit « - w*) Dm, + xv*Dpit + £ vyD^jt

j= i
The first term on the right specifies the effect of the change in real income on
the value share. The log-change Dm, is multiplied by the excess of the
marginal value share pt over the average w* of the “ordinary” shares in t and
t— 1. This excess is positive when the zth commodity has an income elasticity
larger than 1 (in which case the commodity is said to be a luxury good),
negative when the income elasticity is less than 1 (the case of a necessity).
Proof:
d (Pidi)
dqt m dm Pi
(5.14)
dm qt VAi W:
We conclude that an increase in real income has a positive effect on the value
shares of luxuries, and a negative effect on those of necessities.
The other terms on the right of (5.13) deal with changes in relative prices.
There are two kinds of relative prices, one of which is obtained by deflating
by means of the cost of living index (Dpit) and the other by using the marginal
price index (Dpjt). The former is concerned with the direct effect of the /*h price
on the /th value share, the latter with the indirect effect via the zlh quantity.
Which effect will be more important? We shall answer this question for the
special case of preference independence. Then vu = 0 for z/y, vfi = 0î,so that
(5.13) takes the form:
(5.15) Awit * (pi - w*)Dm, + w*Dpit + (t)p,Dp'it
The price term is then confined to two terms in the zth price with different
deflators. Suppose, however, that Dpt&Dp't, which means that the two
weighting schemes produce essentially the same result during the transition
from t — 1 to t. The answer to our question will then depend on the sign of
+ or equivalently on
ihlK
(5.16) ii
-V4>
The numerator on the left can be regarded as the value of the income
elasticity of the ith commodity in the transition from t — 1 to t. The de¬
nominator is the absolute value of the income elasticity of the marginal
utility of income, because the income flexibility 4> (which is negative) is
defined as the reciprocal of this elasticity. When we have > in (5.16), i.e.,
when the income elasticity of the ith commodity exceeds that of the marginal
utility of income in absolute value, an increase in the relative price of that
commodity has a negative effect on its value share. In that case the indirect
effect via the quantity change dominates the direct price effect. Whether this
or the converse is true is an empirical matter. We shall come back to this
problem in the next chapter (Section 7.3).
6.6. The Marginal Utility of Income as a Function of Income and Prices
We just met the marginal utility of income again. In the fundamental matrix
equation (2.5) it played a role which was largely the same as that of the
optimal quantities, but we lost sight of it due to our concentration on the
demand functions. We shall now consider the analogous function (1.15) of
the marginal utility of income.
Equations (2.8) and (2.12) specify how this marginal utility varies when
income and prices are subject to changes. In particular, if we write (2.12) in
the form (3.5) and multiply both sides by Pj/X°, we find
3 0°g A) _ _ dqj d(log/l) pjq° _ Wj

5 (log pj) Pj dm <3 (log m) m ^j (f>
The logarithmic change in the marginal utility of income due to income and
price changes is therefore
3 (log X)
d (log X) = d (log m) + V d(\ogPk)
5 (log m) 5 (log pk)
fc= i
1
= , d (log m)
<P
^ (pk + y)^(logpk)
k= 1
6.6 THE MARGINAL UTILITY OF INCOME 207
which can also be written as
(6.1) d (log A) + Y pkd (log pk) = 1 [d (log m) - Y xvkd (log pkJ\

k= 1 <P k= 1
When we consider finite changes, the equation becomes simply
(6.2) DA’, — - Din, where Dl't = Da, + Dp',
We conclude that the effect of price changes on the marginal utility of in¬
come is completely determined by the behavior of the cost of living index and
the marginal price index. The former index acts as the deflator of money in¬
come, the latter as the deflator of the marginal utility.1 Hence (6.2) states that
the log-change in the marginal utility of income in real terms is equal to the
log-change in real income multiplied by the reciprocal of </>, this reciprocal
being - as it should - the income elasticity of the marginal utility of income.
It is quite instructive to compare (6.2) with the demand equations (5.1). If
we disregard for a moment the specific substitution effects in the right-hand
side of (5.1), the expression which remains is
PiDm, — Y V;jDp;= Y vtJ(~Dm, - Dp',) = Y vuDA,

7= i 7=i \<P J 7=i
where use is made of (4.10) and (6.2). After adding the specific substitution
effects we find that the demand equations (5.1) can be written in the following
equivalent form:
n
(6.3) w*Dqit = Y vij(Dpj, + DA,) i = l,...,n

7=1
This implies that the reciprocal of the marginal utility of income is used as a
deflator of all prices.2 We can simplify this result even further by using the
proportionality (1.7) between marginal utilities and prices. This propor¬
tionality implies that the sum of the logarithms of pj and A0 is equal to the
logarithm of the marginal utility du/dqj in the optimal point. Hence:
n
(6.4) w*Dqit = Y vuDujt i =

7=1
1 If we measure utility in dimensionless “utiles” per unit of time, the marginal utility of
income has a dimension equal to the reciprocal of money. Hence we should add rather
than subtract Dp[ to deflate the log-change in this marginal utility.
2 See the previous footnote for the plus sign before DAt.
where Dujr stands for the log-change in the marginal utility of the jth com¬
modity from the optimal point in t — 1 to the optimal point in t. So it turns
out that the quantity component of the value share change Awit is a weighted
sum of the log-changes in marginal utilities, the weights being elements of the
negative definite matrix [v^-] of price coefficients. In the special case of
preference independence this weighted-sum expression reduces to a simple
proportionality between the quantity component of Awit and the corre¬
sponding log-change Duit in the /th marginal utility.
The demand equation (6.3) does not contain income explicitly. The influence
of income on the quantity bought is taken care of by the log-change DXt in the
marginal utility of income. Hence, when the reciprocal of this marginal
utility is used as a “price index” to deflate prices, we should expect that this is
not an “ordinary” price index. There is nothing special when real income
does not change. As (6.2) shows, it is then equivalent to the marginal price
index. Suppose, however, that all prices increase by 10 per cent, that money
income remains unchanged, and that $=-$. Then any “ordinary” price
index will also go up by 10 per cent. But (6.2) implies that DXt is positive,
hence the index change -DXt is negative. This is the natural consequence of
the fact that the marginal utility of income depends not only on how much
one can buy for a dollar, but also on how many dollars one has.
6.7. The Indirect Utility Function, Its Income Solution, and the True Cost of
Living Price Index
The purpose of this and the next section is to consider the cost of living
index and the marginal price index from the standpoint of classical index
number theory.1 The simplest starting point is the so-called indirect utility
function. We recall that the consumer’s problem is to maximize the utility
function u( ) subject to the budget constraint p'q = m. This leads to a set of
demand equations q° = q(m, p) which describe the optimal quantities in in¬
come and prices. We now substitute these demand equations into the utility
function:
u = u (q(m,p)) = W/(m,p)
The function uf ) is the indirect utility function. Its arguments are income
and prices, not the quantities. The interpretation of u,(m, p) is that of the
1 The analysis which fo,l°ws owes much to discussions with T. Kloek of the Econometric
Institute as well as to the monograph by V. Rajaoja, A Study in the Theory of Demand
Functions and Price Indexes (Helsinki: Societas Scientiarum Fennica, 1958).
6.7 THE TRUE COST OF LIVING PRICE INDEX 209
highest utility level that can be attained when income is m and prices are p —
a level which is actually attained when the consumer’s purchases are in ac¬
cordance with the demand equations q° = q(m, p).
The derivatives of the indirect utility function, whose existence and
continuity follow from the analogous properties of the “direct” utility
function u( ) and the demand functions q( ), can be evaluated in a straight¬
forward manner. We have
n n
duj y du 8qj duI W1 du dqj

dm dqj dm dpt dqj dpt
J=1 J=i
Since du/dqj = X°pj in the optimal point, this leads to the following income
derivative:
(7.2) ,0 V n JO
dm j= i cm
which confirms our earlier statement [see the discussion below equation (1.7)
of Section 6.1] that the Lagrangian multiplier can be regarded as the marginal
utility of income. For the price derivative we obtain:
(7.3) mo
A qt i = 1,..., n
dPi M JdPi
where use is made of p'Qp = — q0', which is part of the fundamental matrix
equation (2.5). Given the positive signs of the optimal quantities and the
associated value of the marginal utility of income, we conclude from (7.2)
and (7.3) that the income derivative of the indirect utility function is positive
and that each price derivative is negative. In words: The highest attainable
utility level increases when income increases or when a price decreases. We
also conclude from (7.2) and (7.3):
dujtjdpi
(7.4) i = 1,..., n
diij/dm
which is an explicit expression for the optimal quantities. It is due to Rene

Roy 1 and it states that q° is equal to the ratio of the /th price derivative of the
indirect utility function to the income derivative, apart from sign.
Our next step implies that we put utility at some fixed level U, while prices
are also assumed to be fixed (at the level p, say). We imagine that income
takes a very low level initially, and that it is then raised gradually. At some
1 De Vutilite; contribution a la theorie des choix (Paris: Hermann et Cie, 1942), pp. 21-25.
critical point the utility level that can be attained for this income (and the
given price vector p) will be U. Clearly, this critical income level is a function
of U and p:
(7.5) m = mI(U,p)
which will be called the income solution of the indirect utility function. It
measures the minimum income which enables the consumer to attain the
utility level U when prices are p, and it is obtained from the indirect utility
function (7.1) by putting u= U and solving U=Uj{m, p) for m. This income
solution mt( ) plays a central role in index number theory.1
We substitute (7.5) into (7.1):
U = «/(ot7(17,p),p)
and differentiate both sides with respect to U and pt:
du, dm, du, dnij diij

\= — — 0 = ——+ —
dm dU dm dpt dpt
which gives in view of (7.2) and (7.3):
dnij 1 dnij
(7.6) i = 1, ...,n
51/ “ Jp]
Thus the income necessary to attain the utility level U in the price situation p
increases by (1 /X°)dU when the target utility level increases by dU, and it in¬
creases by q°dpi when the i‘h price goes up by dp,-. The relationship of the
latter effect and the income effect of price changes will be obvious. Note also
that all 77 + 1 derivatives of the income solution m,( ) are positive.
We now proceed to the third step, which involves the comparison of two
different price vectors. One vector, to be denoted by p0, consists of a set of
reference prices; these will be assumed to be constant throughout. The other
is the current price vector p, the elements of which are assumed to be constant
in this section but will be variable in the next. We ask the familiar question:
How much more expensive is the price situation p than the price situation p0?
1 It is interesting in this connection that maximizing the utility function u{ ) subject to

the budget constraint is equivalent to minimizing total expenditure m = p'q subject to the
constraint of a fixed utility level. The latter interpretation implies that we search for the
lowest income level which is consistent with given U and p, and it is therefore directly
related to (7.5). In fact, this interpretation will be our vehicle when we consider the allo¬
cation problem of the firm in Sections 8.4 through 8.6.
6.7 THE TRUE COST OF LIVING PRICE INDEX 211
The answer given by the economic theory of price index numbers is as

follows. Consider a certain utility level U; then the income needed to attain
that level is mx(U, p0) in one case and mfU, p) in the other. Let mfU, p)
exceed mfU, p0) by 10 per cent, say. This means that prices have changed
from the base situation (p0) to the current situation (p) in such a way that the
consumer needs an income increase of 10 per cent to be able to reach the
same utility level XJ. This required change in money income reflects a 10 per
cent increase in the cost of living. Accordingly, the ratio
mjjU, p)
(7.7) -PC(p|Po, U)
mi(U, Po)
is known as the true cost of living price index of the price situation p with
respect to the price situation p0 (the base situation) at the utility level U. Note
that the comparison of the two price vectors is based on the same utility level,
since U occurs both in the numerator and in the denominator of (7.7).
Hence, if we want to compare prices in the Midwest of the United States with
those in Egypt’s Nile valley, our measurement should not be based on the
different standards of living of a wealthy American farmer and a poor
fellah.1 We should raise the fellah to the utility level of his American counter¬
part and find out how much income is needed in Egypt to achieve this. That is
what (7.7) amounts to. If we proceed in a different manner without making a
utility adjustment, we face the same difficulties as those when using an iron
yardstick to compare a distance in the winter in the Midwest (outdoors) with
a distance in Egypt. The temperature during the experiment should be the
same in both places; similarly, the utility level should be the same in the base
situation and in the current situation. Note also that even if the comparison
is made at the same utility level U, as in (7.7), the result will in general
depend on the value of U. That is, the cost of living comparison may show
that a 10 per cent income increase is needed at a low level of utility but an
8 per cent increase at a higher level. [A difference of this kind may occur
when we reduce the American farmer to the fellah’s utility level instead of the
other way round.] We shall consider such variational problems in the next
section.
The true cost of living index can be illustrated graphically. We take the
two-commodity case and consider the three-dimensional Cartesian m, pu
1 This objection (although in less extreme form) can also be raised to the one-dimensional
scale analysis of Section 5.3. We shall return to this problem in the next chapter (Section
7.9).
p2.space which is given in the left-hand graph of Figure 6.1. Income is

measured along the vertical axis, the two prices along the horizontal axes.
Each point in this space corresponds with one price-income situation. It
follows from (7.1) that there is then also a well-defined utility level associated
with each point. In the left-hand diagram we have two points, A0 and A
representing (m0, p0) and (m, p), respectively, and B0 and B are their
projections in the horizontal price plane.
The right-hand diagram contains the vertical plane through A0, B0, A
and B. The curve through A0 is the locus of all points in this plane where the
utility level is the same as it is in A0, viz., equal to Uj(m0, p0) = U, say. [This
curve is the intersection of the vertical plane and the indifference surface
through A0 in the price-income space.] Our curve intersects the line AB in a
point P, so that PB is the income which is needed in the current situation to
attain the same utility level U as that of the base situation. Thus, the true
cost of living index evaluated at this particular utility level is equal to the
ratio PB/A0B0. The actual income of our consumer in the current situation
is AB, which is less than the required PB. Hence the utility which he enjoys
in the current situation is below that of the base situation.
6.8. A Comparison of Cost of Living and Marginal Price Index Numbers
We just made a comparison of the price situations in B and B0 which is

based on the utility level u,(m0, p0) that prevailed in the base situation. The
6.8 COST OF LIVING AND MARGINAL PRICE INDICES 213
obvious question is: What is the result if we choose any other utility level,
say Uj(m, p) of the current situation? [Answer: Draw the indifference curve
through A and conclude that the true cost of living index is now equal to the
ratio AB/SB0.] However, before going into this matter we shall first define
the true marginal price index. It will turn out later in this section [see
equation (8.7) below] that this index is related to the question just asked in
the sense that it plays an important role in the sensitivity of the true cost of
living index for changes in the utility level at which the latter index is
evaluated.
We imagine that the prices of the base situation are p0 as before, but that
income m0 is raised infinitesimally such that the utility level is now U+dU
rather than U. Hence we are on a higher indifference curve; see the curve
through Q and R in Figure 6.1. To attain the new utility level in the base
situation one needs an income of
QB0 = mj(U + dU, p0)
The original income m0 — A0B0 = mI(U, p0) leads to a utility level U. To

obtain the utility increment dU in the base situation one thus needs an in¬
come increase of
QA0 = m/(l/ + dU, p0) — mj(U, p0) =, p0) dU
which is the vertical distance between the two indifference curves in the base
situation. The corresponding distance in the current situation is
RP = mr(U + dU, p) - m,(U, p) = ~mI(U,v)dU
We then take the ratio of these two distances:
3 / .
-m,(U, p)
m dU
(8.1) PM(p|p0,D)= --
dumi(U’ Po)
which is known as the true marginal price index of the price situation p with
respect to the price situation p0 at the utility level U.1 The interpretation is as
1 This index was used by Ragnar Frisch as early as 1932, who called it the “marginal
living index.” See New Methods of Measuring Marginal Utility, Volume 3 of Beitrdge zur
Okonomischen Theorie (Tubingen: J. C. B. Mohr), pp. 74-82.
follows. If the consumer moves from A0 to Q, his income increases while

prices remain constant at the level p0. This increase induces the consumer to
buy another basket of goods and services whose utility level is higher. To
obtain the same utility increment in the current price situation p our con¬
sumer should move from P to R. In both cases additional amounts of income
are involved. If the additional amount which is needed in the current price
situation exceeds that of the base situation by 15 per cent, say, this means
that prices have changed by that much on the average at the margin. That is,
prices have then increased by 15 per cent on the average for the additional
expenses which are needed to obtain the same utility increment in the two
price situations. It will be clear that the true marginal price index shown in
Figure 6.1, the ratio RP/QA0, is evaluated at the utility level of the base
situation. [Our starting point was the level tq(m0, p0) prevailing in A0.] The
right-hand side of (8.1) shows that the index varies in principle with U even
when p and p0 are fixed, just as the true cost of living index.
We now proceed to such variational problems and start with the de¬
pendence of Pc and PM on the components of the current price vector p, to be
denoted by pu ...,pn. [Remember that the vector p0 of the base situation is
assumed to be constant throughout.] Considering Pc first, we conclude from
its definition (7.7) that p occurs only in the numerator mj(U, p) and from (7.6)
that the derivative of this numerator with respect to pt is qf. Going back to
the discussion around this equation we conclude further that this q? is the
optimal value of the zth quantity when prices are p and when income is such
that U is the maximum utility level that can be attained in this price situation.
We shall indicate this explicitly by writing the second equation of (7.6) in
the form1
(8-2 *) — m/(C,p) = g?(C7,p) i = 1,..., n
or in elasticity form
a [log (F/, p)] Piq°(U, p)

(8.3)
d (log p^ mi(U, p)
= ^°(I7, p) 1,...,n
Thus the elasticity with respect to the z'th current price of the income required
to attain the utility level U in the current price situation is equal to the value
1 Note that - q^m, p) is a demand function in income and prices, whereas q° = q°.(U, p)
is a function which expresses the same optimal quantity in the utility level and prices. For
their equivalence see footnote 1 on page 210.
share of the ith commodity in the following price-utility situation: the prices
are the same as those of the current situation and the income is such that the
maximum utility level that can be attained in this price situation is U, which
is the level at which Pc is evaluated. Given that the price vector is current p,
q?(U, p) and w° ( U, p)
are the optimal quantity and the corresponding value share, respectively, of
the ith commodity in the current situation if the utility level U of (8.2) and
(8.3) coincides with the maximum level that can be attained in the current
situation. For the moment, however, we prefer to keep U unspecified. We
shall make a particular choice at the end of this section.
We derive Pc from mfU, p) by dividing by mfU, p0), which is a constant
with respect to pt. The immediate conclusion is that the price elasticity of
Pc is identically the same as that of its numerator:
d [log-Pc(p|p0, Uj]
(8.4) wt°(C/,p) i = 1, ...,n
d (log
for the interpretation of which we refer to the end of the previous paragraph.
Note that the right-hand side is independent of the reference prices p0.
Next we consider the true marginal index in an analogous manner. Only
the numerator of the right-hand side of (8.1) depends on the current price
vector. The derivative of this numerator with respect to pt is
d2 8 V d d
m7([/,p) tnfU, p) qHu, p)
dpfiU dU _dpi dU
see (8.2). The expression behind the second equality sign measures the change
in the zth optimal quantity when the prices are fixed at the level p of the
current situation but when there is an increase in the utility level U at which
the index PM is evaluated. Let us define
(8.5) M = mj(U, p)
which is the income associated with the U just mentioned and the current
price vector. A fixed p implies that, when U increases, so does M. We can
therefore evaluate the U derivative of q® by means of the income derivative of
the ith demand equation:
d fd<h\ _d. m/(L,p)

qHu, p)
dU \dmJm=MdU
216 THE CONSUMER'S ALLOCATION PROBLEM
Hence the derivative of PM with respect to pt is
mi(U. p)
c .. . dpidJJ
cPi 8
m,(U.p0)
cU
oqA cU_ (cqi\

PV (plPo, u)
KSmJm=M 6 \cmjm=u
—- p0)
cU
or in elasticity form:
c[logFw(p|p0,t/)] 'c(p^h)'
(8.6)
c(log p;) dm Jm = M
Thus the elasticity of the true marginal price index with respect to the current
price of the /,h commodity is equal to the Ith marginal value share in precisely
the same situation as that of the share defined in (8.3): The prices p are those
of the current situation, the utility level U coincides with the level at which
PM is evaluated. This follows directly from the addition m = M on the right in
(8.6) : see (8.5).
The last variational problem to be considered is the dependence of Pc on U.
Going back to the definition (7.7) we note that U. contrary to p. occurs both
in the numerator and in the denominator. Consider the derivative of the
natural logarithm of PL:
c d
”h(U,V) —-7t7j(l/,Po)
(8.7) —t log P (p|Po, U) =-— -—-—
cl mj(l. p) nij (U, p0)
c
— mi(U,pQ)
cU cU p)
8 "h(U. Po)
—-m^L'.po)
cU
Tr-»»/(tf. Po)
CU
(P v - Pc)
mÛ, p)
We conclude that the derivative of log Pc with respect to U is equal to the

excess of PM over Pc multiplied by a ratio which is positive in view of (7.6).
If the utility level at which Pc is evaluated increases, the index will go up or

down depending on whether PM is above or below Pc. Hence it is indeed
true, as we alluded to in the first paragraph of this section, that the true
marginal index plays an important role in the problem of the sensitivity
°f P f°r changes in the utility level at which it is evaluated. If PM exceeds
P , an inciease in this level will raise the latter index. This is understandable,
for the price elasticity of Pc is a value share w° and that of PM is a marginal
share, see (8.4) and (8.6). If we raise the utility level, a larger proportion of
total expenditure will be spent on commodities whose marginal share exceeds
the oi dinary value share. Since PM depends more heavily on the prices of
these commodities than Pc, it stands to reason that Pc will increase if PM
exceeds Pc. Note further that the dependence of Pc on the utility level can also
be read from Figure 6.1. This index is equal to PB/A0B0 when we are on the
utility level of A0. It becomes RB/QB0 when the utility level is that of Q. The
latter ratio exceeds the former if RP/QA0 exceeds PB/A0B0; that is, if PM
evaluated at the utility level of A0 exceeds Pc evaluated at the same level.
Our final topic is the link between the true indices Pc and PM and the
indices whose log-changes are Dpt and Dp\ as defined in (5.3) and (5.5),
respectively. This is particularly simple for the marginal index. Our Dp't
corresponds to the true marginal index in the special case in which this index
is independent of the utility level at which it is evaluated, the marginal value
shares p,- being constants.1 The case of the cost of living index is more
subtle.2 Consider Figure 6.2, which is similar to Figure 6.1 but which deals
specifically with the situations in two successive periods. The left-hand
vertical line refers to the price vector pr_! of the previous period, the right-
hand line to pt of the present period. The incomes are mt_1 and m„ re¬
spectively, so that the corresponding utility levels are
(8.8) Ut-1=MI(mt-1,vt-i) ut = uI(m„ p()
We conclude from Figure 6.2 that the consumer was in B in the previous
period and that he is presently in H. The cost of living index which compares
pf with pf_i can be evaluated at U„ at Ut_u or at any other utility level. We
shall consider in particular the intermediate utility level U*, which is defined
as the following level between U, and Ut^1: We multiply the income corre-
1 It will be noted that this special assumption makes the derivative of PM with respect to
U of no particular interest. This explains why we considered only the derivative of Pc.
2 The exposition which follows is based on the work of T. Kloek. See Indexcijfers: enige
methodologische aspecten. Chapter 6 (unpublished Ph. D. dissertation, Rotterdam 1966).
Pm Pt
Fig. 6.2. Illustration of the intermediate utility level U*t
sponding to the lower utility level (mt^1 in the figure) by some k> 1 and
divide the income of the higher level by k in such a way that we reach the
same utility level U* in both situations. This means that the ratios AC/AB
and EH/EG should be equal, or in algebraic form
(8 9) Pr—l) ^ mi(Uf Pr) or Pr-!) _ ™,
It will be noted that in the case of constant prices, the income

required to attain U* is equal to the geometric mean of mr_ j and mt. The two
vertical lines of Figure 6.2 coincide in that case.
The special form of the true cost of living index is thus
(8.10) PC(P,\Vt-uUt*)
The right-hand numerator indicates that the consumer moves in the present
period from H to G, the denominator indicates that he moved in the previous
period from B to C. The latter walk amounts to a logarithmic income change
which is equal to
pf_i) Pr-l)
log = log
Pr_l) mt- 1
6.9 MONOTONIC TRANSFORMATIONS OF THE UTILITY FUNCTION 219
while prices remain at the level pr_x. Going back to (5.13) we conclude that
the ith value share is then changed from wi>t_1 to
*
! 5
(8.11) wfW.p,-,) * + (ft - log
™t-i
In the same way we find for the present period that the share is changed from
wit to
mr(Ut*, p()
(8.12) w? (Ut*,Vt) ~ wit + 0i - w*) log
m.
The sum is
(8T3) w° (U*,p() + w,°(l/*,p(_i) « wft +
in view of (8.9). We now apply (8.3) and lemma (5.8) to
log Pc (Pr|Pf — U*) = log (I/*, p() - log m* (17*, p(_x)
and find that
wi (Ut ,P() + Wj (U*,pt-i) Wit + W; t_i

(8.14) DPi, £>Pn
i=l ;=i
is a quadratic approximation to the log-change in the true cost of living

index (8.10), where “quadratic” is to be interpreted in the sense that the error
is of the third order in the price log-changes. The « signs in (8.11) through
(8.14) are immaterial, because they too refer to third-order errors in price and
income log-changes; see (5.13) and (5.9). Since the right-hand side of (8.14)
is simply Dpt as defined in (5.3), our conclusion is that Dpt is an approxi¬
mation of the log-change in the true cost of living index for pf with pf_! as
base evaluated at the intermediate utility level U*, and that the error of the
approximation is of the third order of smallness in the price log-changes
Dplt, ..., Dpnt and the income log-change Dmt.
6.9. Monotonic Transformations of the Utility Function
In the last several decades there has been a tendency to replace the utility
function by a preference ordering. That is, if andq2 are arbitrary batches
of commodities, it is assumed that the consumer is able to state whether he
prefers qt to q2, or q2 to q1? or whether he is indifferent between them;
further, if there is a third batch q3 and if the consumer prefers qx to q2 and
also q2 to q3, it is assumed that he prefers qx to q3 (similarly for indifference).
220
Such a preference ordering can be represented by a utility function u( ), but

also by any monotonically increasing function of u, say F{u). The “ordinal
viewpoint implies that only those properties of the utility function are ac¬
cepted as valid which are invariant under such a transformation. The
“cardinal” viewpoint does not impose this condition.
It is easily seen that the theory described in this chapter is not ordinal in
this sense. For example, preference independence implies
d1 2u
(9.1) t—7 — 0 fora11 l*J
dqfiq j
If u is replaced by F(u), where F( ) is an increasing function, so that

F' = dF/du>0 under the assumption of differentiability, the marginal utility
of the i'th commodity becomes
dF(u) dF du du
dqt du dqt dqt
and its derivative with respect to the /th quantity:
d2F d2u du du
-= F'-+ F-
dq^qj dqfiqj dqt dqj
where F" = d2F/du2. We must conclude that even if (9.1) is true, the new
cross-derivative does not vanish except when F" = 0, in which case we confine
ourselves to linear transformations. Therefore, a zero value of a cross-
derivative is not a property which is invariant under monotonic transfor¬
mations of the utility function.
Nevertheless, our approach is only quasi-cardinal, since we can interpret
preference independence as follows: The consumer’s preference ordering is
such that it can be represented by a class of utility functions which contains a
subclass of the additive form (4.13). This subclass is merely chosen to simplify
the demand equations; it minimizes the number of coefficients to be esti¬
mated. The interpretation is analogous for block-independent preferences.
It is important to realize that one always needs some procedure to reduce the
number of coefficients. This is not really achieved by the ordinal theory,
which merely states that real income and all relative prices affect the demand
1 For more detailed considerations see G. Debreu, “Representation of a Preference

Ordering by a Numerical Function,” Chapter XI of Decision Processes, edited by R. M.
Thrall, C. H. Coombs and R. L. Davis (New York-London: John Wiley and Sons, Inc.,
and Chapman and Hall, Ltd., 1954).
6.9 MONOTONIC TRANSFORMATIONS OF THE UTILITY FUNCTION 221
for each commodity. It gives no indication whatever that the margarine price
is more important than the sugar price in the demand equation for butter.
The present approach does give some indication, viz., in terms of the Hessian
matrix of the utility function - however intuitive that indication may be.
Also, the ordinal theory does not specify what kind of price index should be
used to obtain relative prices. The present approach states clearly that this
should be the marginal price index.
The idea of monotonic transformations of the utility function does have
some clarifying value with respect to the concepts of preference independence
and block-independence. Let us take the exponent of the utility function
(4.13):
(9.2) eu(q) = f\ eUi(qi)

i= 1
This is a monotonic transformation. If there is no over-saturation (i.e., if

du/dqi<0 is excluded for any nonnegative values of the q’s) the left-hand
function is a nondecreasing function of its arguments. On the assumption
that u( ) has finite upper and lower bounds we can then apply a linear
transformation to e"(q) such that it takes the zero value when all q's vanish
and the unit value when they are all infinitely large. It is easily seen that
eu(q) can then be regarded as a cumulated distribution function in n dimen¬
sions. In the special case (9.2) this distribution is characterized by stochastic
independence, which is an additional justification of the term preference
independence. The block-independence case corresponds to stochastic vector
independence, the vectors consisting of the q's of each set Sg, g = 1, ..., G.1
This stochastic interpretation implies that utility is regarded as a proba¬
bility. When we return from (9.2) to (4.13), we simply take the logarithm of
the probability. Hence we may say that the additive forms (4.13) and (4.15)
correspond to the informational variant of the utility-probability concept.
1 The idea of utility as a probability can also be applied to the utility function of wealth
described by H. Markowitz, “The Utility of Wealth,” The Journal of Political Economy,
Vol. 60 (1952), pp. 151-158. If the lower limit of the utility curve in his Figure 5 is defined
to be zero, and the upper limit one, this curve describes a cumulated distribution function
corresponding to a bimodal density function, the local minimum of the latter being the
point of present wealth. Efforts are presently being made by B. M. S. van Praag of the
Econometric Institute to apply measure theory concepts to utility theory, which is a
related approach.
6. A. Proof of the Lemma on Local Quadratic Approximation
To prove lemma (5.8) we start by considering the case in which /( )

depends on one argument only. We indicate successive derivatives by/'( ),
/"( ), ... and assume that they exist up to any required order. A Taylor
expansion applied to the first-order derivative gives
(A.l) f'{x) = f’(y) + (x- y)/"(y) +i(x- yff'"{y) + ...
We now apply the Taylor expansion to the function itself and use (A.l):
/ 0) - / (y) = (x - y) f 00 + i(x - y)2 f"(y) + £(x - y)3/"'(y) + .••

= i (* - y)/'(y) + Hx - y)[/'(y) + 0 - y)/"0)]
+ iO - y)3/"'(y) + •••
= Ux ~ y)U'(y) + f'(x) - Hx - y)2/'"(y) - •••]
+ i(x - y)3/'"(y) +
= i(x-y)[f'(x) + f'(y)] -ti(x - y)3/"'(y) + ...
This proves the lemma for the one-argument case. The second term on the
last line is the leading term of the approximation error.
In the more general case - where the argument is a real-valued vector -
we write
(A-2) / (y + z) = / (y) + z7y + iz'Fz + R3
where fy is the gradient of/( ) in y, F the Hessian matrix in y, and R3 a

polynomial in z whose leading terms are of the third degree. The specification
(A.2) thus implies that, if/( ) is not a polynomial, it is replaced by a Taylor
expansion around y. In particular, when y + z = x, we have
(A.3) / (x) - / (y) = (x - y)'fy + }(x - y)' F(x - y) + R*
where R* is the same function of x-y as R3 is of z. Next consider the

gradient/x of/( ) in x:
<A-4) /x = /y + F(x —y) + R*
222
6.B THE TRUE INDEX OF REAL INCOME 223
where R? is a vector of polynomials in x — y whose leading terms are

quadratic. On combining (A.3) and (A.4) we obtain
/(*)-/ (y) = I (* - y )'(/„ + /, - R?) + R|

= i (* - y)' (/„ + /,)+ [R? - i (* - y)' R*‘]
It is immediately seen that the term in brackets on the second line is of the
third order in the elements of x - y. This term is the error of the approximation.
6.B. The True Index of Real Income1
This section deals with the theory of the true real income index, which is
the quantity analogue of the cost of living index theory of prices. The latter
theory is of more direct relevance to the demand equations, but the former is
not without relevance either, so that it is worthwhile to pay some attention to
it at this place.
In both theories the basic concept is mfU, p), the minimum income which
is needed to attain the utility level U at prices p. Consider then a utility U0 of
a base situation and another utility U of a current situation. It is essential now
that we apply the same price vector p to both situations. Suppose that
mfU, p) exceeds m/(C70, p) by 20 per cent, say. This means that the consumer
is so much more well-to-do in the current situation than he was in the base
situation that he can afford to spend 20 per cent more at the constant
prices p. Accordingly, the ratio
mi(U, p)
(B.l) Q(u\u0,v) mi(U0, p)
is defined as the true index of real income for the utility level U with U0 as base
evaluated at the prices p. On comparing (B.l) with (7.7) we conclude that
both Q and Pc have the form of the ratio of one mf ) to another mf ). An
important difference is that in the case of Pc we have the same U in the
numerator and the denominator, which is the utility level at which this index
is evaluated, whereas in the case of Q we have the same p in the numerator
and the denominator, which is the price vector at which the real income in¬
dex is evaluated. Actually, the questions asked are entirely different. P(
answers the following question: How many more dollars do I need in the
current situation compared with the base situation to enjoy a given standard
of living? The question answered by Q is: How many more dollars are
needed to have the standard of living of a Midwestern farmer compared with
1 The analysis described in this section is based on the work of T. Kloek, loc, cit.
that of a Southern farmer when the prices to be paid for the goods and
services are those which prevail in the Midwest? (Or those prevailing in the
South or in Arizona or anywhere else.)
It is nevertheless true, despite these differences, that the two theories are
intimately related. The theory of the true real income index can be illustrated
with the same type of charts as those which were used in Sections 6.7 and 6.8.
Take, for example, the situation described at the end of Section 6.7. It was
stated there that the consumer’s utility level in the current situation is below
that of the base situation. He is currently in A (see Figure 6.1) and should
have been in P in order to have the original standard of living. If we identify
the utility level in P with U0 and that in A with U, the true real income index
as defined in (B.l) is equal to the ratio AB/PB when evaluated at current
prices. Note that real income comparisons are always made on one vertical
price line [constant p in the sense of the same p in the numerator and the
denominator of (B.l)], whereas cost of living comparisons are always made
on one indifference surface (constant U).
We proceed to consider variational problems similar to those which we
analyzed for the true cost of living index in Section 6.8. The derivative of Q
with respect to U is simple:
d
mi(U, p)
d dU
(B.2) Q(u\u0,v)
dU p)
which is positive in view of (7.6). The price derivatives will be given in

elasticity form. Consider the logarithm of Q:
logg(C|(7o,p) = log - log mj([70,p)
and apply (8.3) with the following result:
^ [log Q(u\u0, P)]

(B.3) ~=Wi(U,-p) - w?(l/0,p) i = 1,n
d (log Pi)
The elasticity of the true real income index evaluated at prices px, ...,pn with
respect to pt is thus equal to the difference of the values taken by the z'th value
share in two situations: the first corresponds to the current utility level U, the
second to U0 of the base situation, and in both we have the same price
vector p. Suppose then that the pricePi of a luxury good increases; suppose
further that U> U0. Given that the consumer is currently more well-to-do
than he was in the base situation, he will now spend a larger proportion of
6.B THE TRUE INDEX OF REAL INCOME 225
his income on the /th commodity. The right-hand side of (B.3) is therefore
positive, so that this equation states that the true real income index for U
with U0 as base increases when /?,- increases. This is understandable, because
a higher price of a luxury implies that the consumer needs a larger income to
attain the given utility level U> U0.
We shall now show that the volume index whose log-change is Dqt as de¬
fined in equation (4.4) of Section 5.4 is a local quadratic approximation to
the log-change in the true real income index which compares periods t and
t— 1. The index is to be evaluated at prices which are geometric means of
those in t— 1 and t. This price vector, to be written pf with elements
(B.4) p* = yjPitt-iPit i = 1, n
will be called the “intermediate price vector” for reasons of symmetry with
the intermediate utility level U*.
Our true real income index is thus
Pt*)
(B.5) Q(Ut\Ut.u p?) =
i,p?)
Consider the right-hand numerator. We shall approximate its logarithm

from the observed value in the current period, using (8.3) and lemma (5.8) and
Pit
log — = — jDpit i = 1,..., n
Pit
The result is
n
(B.6) logmI(Ut,p?) « logm/(L/r,pf) + ^ "+ ^ $Dpit)

/= i
wu + wi> (^>P*)
= log mt — Dpit
i= 1
In the same way we obtain for the logarithm of the denominator:
n
(B.7) log mJ(E/,_1,p(*) « logm(_x +

l
i= 1
Wtt-1 +
4
(Ut~'i,p*)
Dpit
The « signs in (B.6) and (B.7) refer to third-order errors in the price log-
changes.
The logarithm of Q as defined in (B.5) is then found by subtracting (B.7)
from (B.6):
(B.8) log Q(Ut\Ut_utf) x Dm,
Wit + Wi, t - 1 + ( Uf P( ) + Wi ( Uf - 1 ’ Pt ) n ^
-- " L>Pit
i= 1
The value shares wf(Ut, pf) and wf (x, p?) are not observable. We shall
approximate them linearly from wit and wi>(_ 1; respectively, so that their sum
becomes
(B.9) +
dw°(Ut, pr) _ gw^l/t-uPt-J

wit + Wi>t_1
5 (log pj) dQogPj) .
Dp,
jt
j= 1
wit + wu_1
Given that the approximation is linear, the error implied by the first ~ sign
is of the second order in the price log-changes. The expression in square
brackets on the second line is of the first order in income and price log-
changes, because it is the difference between two derivatives of the same
function evaluated at argument values which differ to that order. This
expression is multiplied by Dpjt, so that the third line of (B.9) differs from the
first by an error of the second order in income and price log-changes.
We now combine (B.8) and (B.9). Since the sum of the two value shares
considered in (B.9) is multiplied by Dpu in (B.8), the error just mentioned
becomes now of the third order. The result is
Wit + Wi>t- 1
(B.10) log <2(t/,|[/,_!, pf*) k, Dm, Dpit
i= 1
= Dmt - Dqt — 5t k Dqt
Our conclusion is that both the log-change Dmt in real income and the log-
change Dq, in the volume index are local quadratic approximations to the
log-change in the true real income index for U, with Ut_l as base at the inter¬
mediate price vector p/'. The difference between Dqt and Dmt is the allo¬
cation discrepancy 5,, which is of the third order and therefore irrelevant to
our degree of approximation. In fact, we shall make a shift from Dmt to Dq,
as an explanatory variable in the demand equations. The precise reasons will
be explained in the beginning of the next chapter.
CHAPTER 7
EMPIRICAL IMPLICATIONS OF THE
ALLOCATION MODEL OF THE CONSUMER
7.1. A Reformation of the Demand Equations; A Marginal Utility Shock Model

The objective of the present chapter is to analyze the empirical implications
of the demand model of Chapter 6 by means of some of the data of Chapter 5.
The realism which is needed to carry out this task implies that we should add
error terms to the demand equations (5.1) of Section 6.5. We shall therefore
write them as follows:
n
(1.1) w*Dqit = niDjht + £ VijDp'j, + vit i =

j= i
where vh is a random disturbance with certain statistical properties that will
be discussed more fully in this section. Now it will turn out that these
properties have the following implication:
(1.2) £ v„ = 0
i= 1
which is not satisfied by the disturbances of (1.1). We know this from the
discussion under (2) of Section 6.5. We found there that, when the n demand
equations are added, we obtain Dqt on the left and Dmt on the right. If we
have disturbances as in (1.1), the result of the summation is
n n
Dqt = Dmt + £ vit = Dmt - Dpt + £ vit

i= 1 i= 1
which implies that the sum of the disturbances is equal to the allocation dis¬
crepancy 5t, see equation (4.5) of Section 5.4. Although we found in that
section that these discrepancies are very small, they are not exactly zero as
(1.2) requires. However, there is a very simple way of achieving this goal. We
replace the log-change in real income by the log-change in the volume index,
so that the demand equations take the following form:
n
(1.3) w*Dqit = ^Dq, + £ vijDPjt + vit i=l,...,n

7=1
It is easily verified that condition (1.2) is now satisfied. Equation (1.3) is the
227
228 EMPIRICAL IMPLICATIONS OF THE ALLOCATION MODEL
demand equation which we shall use throughout this chapter. We shall

nevertheless continue to speak about real income even when the volume index
and its log-change Dq, are really used.1
We proceed to develop a model which specifies, in some detail, all moments
of the first and second order of the disturbances vit.2 To do so satisfactorily
we should go back to the two cornerstones of consumer demand theory: the
utility function and the budget constraint. Now the constraint is formulated
in terms of income and prices, which are assumed to be exogenous. This
excludes the idea of randomness. We must therefore concentrate our efforts
on the utility function. Let us assume that it is quadratic:
(1.4) u(q) = a'q + iq'Uq
so that the marginal utilities are linear functions of the quantities:
du
(1.5) —- = a + Uq
oq
where the Hessian matrix U consists of fixed elements. The exposition is

simplified considerably if we make the quadratic assumption. It is not really
restrictive, because the quadratic approximation suffices when we consider
only first-order effects in the neighborhood of the consumer’s optimal point.
The n proportionalities of marginal utilities and prices and the budget
constraint can be written in the following partitioned form:
u p~ q — a
_p 0_ _-2_ m
which implies that the demand equations are
0.6) q° = - U“‘a + D77jU-(<J*‘p)(U',p)'a + -Tz^rir‘p

P u P p U p
This is easily verified by using the inverse of the bordered Hessian matrix as
given in Section 6.2.
The reader who went through Section 6.B of the Appendix to the previous chapter will
realize that this is only a matter of terminology. Both Dqt and Dirk are accurate to the
second order with respect to the log-change in the true real income index evaluated at
prices which are geometric means of those in t — 1 and t.
“ An earlier attempt was made by H. Theil and H. Neudecker in “Substitution, Com¬
plementarity, and the Residual Variation around Engel Curves,” Review of Economic
Studies Vol. 25 (1957), pp. 114-123. The models suggested in that paper are related to the
one which is proposed here, but they are more complicated and less elegant.
7.1 A REFORMULATION OF THE DEMAND EQUATIONS 229
Suppose now that the marginal utilities are subject to additive random
shocks. That is, during the transition from t-1 to t a random vector Aa is
added to the right-hand side of (1.5) which indicates a stochastic variability
of the utility of the last (marginal) unit of each commodity. We imagine that
this variability is due to the behavior of certain neglected determining
factors. The effect on the quantities bought is
0-7) Aq = - U-'fl — —^y-pp'U-1) Aa

V PU *p J
see (1.6). We conclude that the quantities bought are now also random
variables, which is precisely what is needed to interpret the disturbances
vit of the demand equations (1.3). We recall that in the infinitesimal case the
dependent variable of the ith demand equation is
Pi
w4 (log qt) dqt
m
Hence, if we want the vector v of disturbances of all n demand equations

which are caused by the marginal utility shocks, we should premultiply both
sides of (1.7) by (1/m)P, where P is the nxn diagonal price matrix. So we
obtain
v= PU -11
m /■pakPpp'u")4a
«?2° 1
-PU_1P I - ll PU_1P P~1 Aa
m V2° m #2°
PU_,P i
m
where use is made of Pi = p, i being a column vector of n unit elements.

Furthermore, #2° stands for the expected value of the marginal utility of in¬
come. The random shocks of the marginal utilities of the n commodities
imply that not only the optimal quantities but also the associated value 2° of
the marginal utility of income is random. The use of <?2° rather than 2° is
preferred in the above equation in order to concentrate all randomness in the
last vector Aa. This advantage becomes clearer when we introduce the
matrix N of all price coefficients in all n demand equations. On the basis of
equation (4.9) of Section 6.4 we define
(1.8) N= [Vy] =-PU_1P

m
from which we immediately conclude that the use of S’X0 instead of 2° pre¬
vents the v’s from becoming stochastic. We can then write
0.9) ' = -(N-s,Nu'N)/"A'
This result can be simplified even further by the introduction of a column

vector p of marginal value shares and the use of
(1.10) i'Ni = (j) Ni = #
see equations (4.11) and (4.10) of Section 6.4. We then find
(1.11) v = —(N —(/>pp')^-0P_1Aa
or in scalar form:
" x Ad:
(1.12) vt = - X (vu - (pudij)-o i = U •••, n
j= 1 Pj6A
where Acij is the shock of the jth marginal utility.

The interpretation of (1.12) is straightforward. The disturbance vt of the
ith demand equation is a weighted sum of the ratios AajKpj^X0), which are
the marginal utility shocks measured as fractions of the corresponding
“equilibrium values” (pj multiplied by the expectation of A0) of these
marginal utilities. The weights are, apart from sign, the coefficients vtj of the
specific substitution effects and — of the general substitution effects,
see equation (4.6) of Section 6.4. In other words, each marginal utility
shock Actj has a specific and a general substitution effect on the demand for
each commodity; the weighted sum of all these effects is the disturbance of the
demand equation.
We shall now make some simple statistical assumptions on the marginal
utility shocks. The first is
(1.13) *?(Aa) = 0 implying <fv = 0
The disturbances are thus all supposed to have zero mean.1 * The second is the
crucial assumption:
(1.14) <f[Aa(Aa)'] = kU
1 The simplest extension amounts to assuming that there is one marginal utility shock Aaj
which has a nonzero mean. Note that this implies that all disturbances have a nonzero
mean, not only the;111, in a manner which is determined by the coefficients of (1.12).
Note also that such nonzero means represent time trends, because the demand equations
have a first-difference character.
7.1 A REFORMULATION OF THE DEMAND EQUATIONS 231
where k takes some negative value. This condition implies, as far as the
variances ,
, d2u
fi{Aa)2 = k—2 i = 1.n
dqt
are concerned, that when the ith marginal utility is very stable in the sense that
changes in qt have little effect on its value, it is also stable in the sense that its
shocks have a small variance. This is not at all unreasonable. Regarding the
covariances: -
d2u
£ (A ci/A a ) = k - —— i #j
dqidqj
these are all supposed to vanish when there is preference independence. Hence
this form of independence implies a uniform zero value of all correlations of
marginal utility shocks. Similarly, when there is block-independence, all
correlations are assumed to vanish which deal with marginal utilities corre¬
sponding to different blocks. When the cross-derivative does not vanish, the
covariance has the opposite sign (due to k<0). Suppose for example that
du/dqi is a decreasing function of qj, so that Aat and Acij are positively corre¬
lated according to (1.14). This is not unreasonable either. When the ith
marginal utility decreases with increasing qj, this can be interpreted in the
sense that the two commodities satisfy similar wants. The associated positive
correlation implies that, on the average, deviations of the same sign from the
expected marginal utilities are concentrated on such related wants. For the
case of a positive cross-derivative the picture is exactly the opposite.
If condition (1.14) is satisfied, the covariance matrix of the disturbance
vector v as defined in (1.9) takes a very simple form. We postmultiply v by its
own transpose, which gives
N —— Nu'N )P-1 (Aa)(Aa)'P_1 ( N Nu'N

(<fA0)2' TNi i'Ni
If we then take the expectation to obtain we find kU for the term in

the middle. Hence:
k
1 Nu'NjP_1UP_1( N-^ Nu'N
^ = («rlN ,'N.
1 1
—I N - Nu'N N"1 N- Nu'N
m£?r i'Ni" i'Ni"
which can be simplified to

k
(1.15) £{yy') - Nu'N
m<oA° i'Ni
EMPIRICAL IMPLICATIONS OF THE ALLOCATION MODEL
232
It is easily seen that the sum of the n disturbances is identically equal to zero,
i'v = 0, in accordance with (1.2). This follows from
k t, — Nu'N11 = 0
d’(i'v)2 = «f (i'vv'i) =
mS’X0 i'Ni /
Our results will take a more elegant form if we write k as follows:
, mS 'j1° 2 dX
(1.16) k = a2—— = <r m—- --
(j) d (log m)
The negative sign of k is ensured by the positive sign of a2, m and £1°
combined with the negative sign of 4> and the X derivative. After adding a
subscript t to v in order to indicate that this disturbance vector refers to the
transition from t-l to t, we thus obtain:
(UT) '(V0-^(N - l4N“'N) “
The specifications (1.13) and (1.17) complete the demand model (1.3) as far
as the moments of the first and the second order are concerned. We conclude
that the variances and covariances are proportional to the corresponding
total substitution effects v^-0/q/iy of the price changes. Hence the q’s and v’s
play a double role in the demand model. They are the coefficients of the
systematic parts of the equations and thus determine the expected value (the
first moment) of each dependent variable, given the changes in income and
prices. They also play a major role in the matrix of second moments.1 The
1 It is of some interest to point to the similarity of this covariance specification with the
covariance matrix of estimates that emerges in maximum likelihood theory. The latter
matrix is the negative inverse of the matrix of second-order derivatives of the logarithmic
likelihood function. Suppose now that we regard utility as a probability and then take its
logarithm (in accordance with the remarks made at the end of Section 6.9); the resulting
function t/( ) can then be compared with a log-likelihood function and the matrix P_1UP_1
contains its second-order derivatives with respect to the n expenditures, the typical element
of that matrix being
82u
8(piqi)d(piqd
Now (a2/^)N is a negative multiple of the inverse of P_1UP see (1.8), and can thus -
in a way - be regarded as the analogue of a maximum likelihood covariance matrix. [The
subtraction of mi' from (l/</>)N in (1.17) merely serves to ensure that the disturbances add
up to zero identically.]
7.2 AGGREGATION AND DISAGGREGATION 233
division by </> — i'Ni in (1.17) merely serves to normalize the specific substi¬
tution matrix N such that the sum of all elements is 1. The only additional
parameter is a2, which is the basic variance measure of the complete set of
all demand equations. It is equal to the expected value of a quadratic form
whose vector is v, and whose matrix is the inverse of the normalized matrix
just mentioned, apart from the factor n-1:
(1.18) £ [v' (</>N 1) vr] = a2 (n - 1)
The proof is as follows:
The factor n — 1 in (1.18) should be regarded as the number of linearly

independent random variables, see (1.2).
In the special case of preference independence the covariance pattern is
simplified as follows:
var vit = o2Hi{l - p,)

(1.19)
cov (vit, vjt) = - a2ninj i #j
Hence, when preferences are independent, the marginal utility shock model
implies variances and covariances of disturbances which are proportional to
those of random sampling from a multinomial urn whose proportions are
the marginal value shares.
7.2. Aggregation and Disaggregation of Demand Equations; Partitioning of

Demand Analysis
We shall now consider the aggregation of demand equations for individual

commodities to equations for sets of commodities.1 The obvious dependent
1 The aggregation problem of the demand equations (1.1) has been considered by A. P.
Barten and S. J. Turnovsky in “Some Aspects of the Aggregation Problem for Composite
Demand Equations,” Report 6415 of the Econometric Institute of the Netherlands School
of Economics (1964). Their results correspond with those described here except for the
covariance specification of the disturbances, which they derived from an earlier report by
A. P. Barten and H. Theil. The present specification is preferred because it has a direct
utility interpretation (see Section 7.1). Also, its implications for the conditional equations
[see equations (2.10) and (2.11) below] are much simpler than those of its predecessor.
variable is
(2.1) W*DQgt= £ wftDqit

ieSg
see equation (6.3) of Section 5.6. We shall confine ourselves to the case in
which the sets S1} ..., SG are block-independent.
The right-hand side of (2.1) indicates that we should add all demand
equations (1.3) that belong to Sg. This gives
(2.2) W*DQgt = MgDqt + £ £ VijDp'jt + £ vit

1 e Sg J e Sg ie Sg
where
(2.3) Mg= Y Vi 9 = 1, —,G

ieSg
is the marginal value share of Sg. We evaluate the price term as follows,
making use of the symmetry of the v’s and of equation (4.17) of Section 6.4:
DPj, = <t> Vj(DpJt - Dp',)

jeSg
= «,( Y Jj DPj. - Dp) = 4>m,dp;,

jsSg
where
(2.4) DP'„ = DP',-Dp', DP’, = V ~Dpu g = l,...,G
ieSg 9
Hence DPgt is the log-change in the partial marginal price index of Sg and
DPgt deals with that price index relative to the overall marginal price index
(whose log-change is Dp't).
We conclude that (2.2) can be written as
(2.5) W*DQm = Mfiq, + <PM,DP’, + V„
where
(Z6> K, = I v„
ieSg
It is immediately seen that, as far as the systematic part of the equation is

concerned, (2.5) is the analogue of
w*Dqit = PiDqt + (j)piDpit + vit

7.2 AGGREGATION AND DISAGGREGATION 235
which is the demand equation for the ith commodity in the case of preference
independence (v;i = vfj-= 0 for /#y). The next question is whether the
residual part has second moments similar to (1.19). For the variance we have
a
(2.7) = } } * 0’itvjt)=— (Vij-(pHiHj)
ieSg jeSg ieSgjeSg
a2 V
^ = a2Mg( 1 - Mg)
ieS„
and for the covariance of Vgt and Vht, g#/?:
<2.8) *(v„v„)=£ ^ *(vuvj,)=a-~ y y (- </wo)

i eSg je Sh i e Sg j S Sh
- °2MgMh
Clearly, these results are completely analogous to (1.19). We conclude that in

the case of block-independence a simple summation of the demand equations
of each block leads to a smaller system of G demand equations which has all
properties of preference independence. It will also be noted that the M’s
defined in (2.3) add up to 1 and are all positive, just as the /r’sin the preference
independence case. The second property follows from
1
(2.9) Vij > 0
i 6 Sg j G Sg
the appropriate submatrix of the v’s being negative definite.

We now proceed to the problem of disaggregation. We have seen that
aggregation in the block-independence case leads to a smaller number (G) of
preference-independent demand equations. The question then arises whether
it is possible to formulate demand equations for individual commodities
within each Sg in terms of set concepts only; that is, in terms of the amount
spent on that set and the relative prices of that set. If this question can be
answered in the affirmative, demand analysis can be partitioned in the block-
independence case. The first problem is then how income is to be allocated
to the G sets, given Dqt and the set price indices DP'gt, see (2.5). The second
is how the amount allocated to Sg should be further allocated to the individual
commodities in Sg, given DQgt and the relative price changes of the com¬
modities in Sg.
To answer this question we consider the income term of the /th demand
equation and express it in the various terms of (2.5):
MiDMt = MgDch
- ^ (W*DQ„ - 4>M,DP;, - V„)
The ;'th demand equation can then be written as follows:
Vi
(2.10) w*Dqit = W*DQgt + £ vi}Dp'jt - ^DP; + vit - y Vgt
Mg i<=S„
Mi
= v'
M,
Kdq„ + E »«(DPjl - DP■;,) + v„ -
M,
v„
j e Sg
The first term on the right is the conditional marginal value share qJMg
multiplied by the sum of the quantity components of the value share changes
of all commodities in Sg. The conditional marginal shares are the extensions
of the /Ts which we need in the present type of equation; their sum (ie Sg) is 1.
The expression W*tDQgt is the generalization of Dq, in (1.3); for if we have
the set of all n commodities instead of Sg, Wgt becomes 1 and DQgt becomes
Dqv The price term is also in accordance with that of (1.3); the deflation is
now in terms of the partial marginal index (DPgt), not the overall marginal
index (Dp[). The sum of the disturbances is identically zero:
Mg
I M„
Kr = 0
see (2.6). This is in accordance with (1.2). The covariance of two such
disturbances (i,jeSg) is
S
Mi_ Ml
Mg Mg
Mj
~ <?(Vi,Vjt) + ~~{ <?Vg2t <?(Vgtvit)
IVlg Mg
If we use (2.7) and
(Vij - (frpi/ij) = a2/q( 1 - Mg)

i
j 6 Sg
7.3 A NUMERICAL ILLUSTRATION 237
we find after some rearrangements:
(2.11)
if A*/\
0 Mg Mj
The conditional shares pjMg and Pj/Mg are multiplied by cpMg, not by (p as
one might have expected. This can be clarified by pointing to (2.9), which
shows that the sum of all v’s within Sg is equal to (f>Mg, not to cp. It can be
proved along similar lines that the covariance of the disturbances of two
demand equations of the form (2.10) is zero if they correspond to different
blocks (;ieSg,jeSh, g¥^h).
We have thus shown that if there is block-independence, we obtain demand
equations of the preference independence type by summation and conditional
demand equations for the commodities in each block. The latter equations
have a form similar to (1.3). Conditional equations of different blocks can be
regarded as “independent” in the sense that the right-hand variables have
nothing in common. We have W*DQgt for the quantity index in one block,
W*tDQht in the other block, and g^h. Similarly, the price variables are
Dpjt — DPg't, jeSg in the first block, Dpkt — DP,'t, keSh in the second. Finally,
there is zero correlation among the disturbances whenever they correspond
to different blocks.
It is in principle possible to formulate a complete hierarchy by splitting up
blocks into sub-blocks. This will not be pursued here. From now on we shall
confine ourselves to the “upper layer” of such a hierarchy, for which prefer¬
ence independence is postulated. Actually, this assumption is, in a sense, not
restrictive at all, because we can always transform prices and quantities such
that it is true. We shall make no use of this result in the text but refer for its
proof to the Appendix of this chapter (Section 7. A).
7.3. A Numerical Illustration; The Information Value of Demand Equations
The theory will now be illustrated by means of the Food-Vice-Durables-

Remainder example of Section 5.6. To simplify the notation we shall proceed
as if these are four individual commodities. That is, we shall write wit for their
value shares, not Wgt; similarly Dpit for the price log-change and IIt for the
variance of these log-changes, not DPgt and TI0t as we did in equations (6.3)
and (6.7) of Section 5.6. The results of the previous section provide a theo-
retical justification of this procedure; notational convenience is the practical

justification.
We shall work with four demand equations of the preference independence
type:
(3.1) w*Dqit = HiDq, + vaDp'it + vit i= = 4)
and specify the marginal value shares as follows:1
Food:
Vice: q2 = .1
(3.2)
Durables: /r3 = .4
Remainder: q4 = .3
It was mentioned in Section 5.6 that the value shares were on the average
about .3 for Food, .1 for Vice, .25 for Durables, and .35 for Remainder.
Flencethe specification (3.2) implies income elasticities p;/w; of the order of
.7 for Food, 1.0 for Vice, 1.6 for Durables, and .9 for Remainder.
The income flexibility <p will be put equal to —.4. Hence the marginal
utility of income is postulated to decrease by 24 per cent when income in¬
creases by 1 per cent. This income elasticity exceeds all elasticities of the four
commodity groups in absolute value. We conclude from the discussion under
(4) of Section 6.5 that an increase in the relative price of any of these com¬
modities will raise its value share (provided that Dp,zzDp't). Furthermore,
using v;i = </>/<;, we conclude that the price coefficients take the following
values:
Food: vu = — .08
(33) vice: v22~ ~ -04
Durables: v33=—.16
Remainder: v44 = —.12
The numerical specification (3.2)-(3.3) plus the data of Chapter 5 enable us
1 The statistical estimation of the coefficients of the demand equations does not belong
to the objectives of this book. This subject was considered by A. P. Barten, Theorie en
empirie van een voUedig stelsel van vraagvergelijkingen, unpublished Ph. D. dissertation,
Rotterdam 1966. [An article in English is forthcoming.] The numerical specification
(3.2)-(3.3) and also (4.5) in the next section is based on discussions with Barten, which
were in turn based on his “Evidence on Slutsky Conditions for Demand Equations,”
Report 6504 of the Econometric Institute of the Netherlands School of Economics
(1965).
to compute the values taken by the disturbances v^.1 These values, together
with those taken by the left-hand variables of the demand equations, are
presented in Table 7.1. It shows, for example, that the value share of Food
increased by .0314 in 1921-22 as far as the volume component is concerned,
that income and price changes accounted for an increase of only
.0314 —.0113 = .0201, so that an increase of .0113 (the disturbance) is left
unexplained. An obvious question is whether the disturbances have variances
and covariances that are in accordance with the theoretical pattern (1.19).
To answer this question we compute the estimates of these variances and
covariances, which are of the form
1
i,j = 1, ...,«(= 4)
T
t
where T is the number of observations. There are 18 observations in the

prewar period 1921-22 through 1938-39, for which the estimated covariance
matrix is
~ 31.7 - 1.3 - 8.1 -22.3
- 1.3 5.7 - 5.3 .9
(3.4)
- 8.1 - 5.3 30.5 - 17.1
- 22.3 .9 - 17.1 38.5
For the 15 postwar observations 1948-49 through 1962-63 we have the

following matrix:
26.3 1.5 - 15.5 - 12.3
1.5 3.9 - 7.5 2.1
- 15.5 - 7.5 63.9 - 40.9
- 12.3 2.1 - 40.9 51.0
and for the 33 prewar and postwar observations combined:
29.3 - .1 - 11.5 - 17.8'

- .1 4.9 - 6.3 1.5
- 11.5 - 6.3 45.7 - 27.9
- 17.8 1.5 - 27.9 44.2
i The price index Dp' which is used to deflate the prices in the four demand equations is
obtained from the four separate Dpu by weighting them in accordance with the A s of (3.2).
[We shall come back to Dp't in Section 7.5.] It follows from (2.4) that strictly speaking we
should have used four marginal price indices, which could have been obtained from the
price indices of the 16 smaller commodity groups of Section 5.4. This refinement has not
been made, although it is in principle possible. In fact, an attempt in this direction was
made by Barten and Turnovsky, loc. cit.
TABLE 7.1
DEPENDENT VARIABLES AND DISTURBANCES OF THE DEMAND EQUATIONS
FOR FOUR COMMODITY GROUPS
Food Vice Durables Remainder
w*uDqu Vlt w'2tDqzt V2t w*3tDq3t V3t wAtDqAt VAt
Prewar
1921-22 3.14 1.13 -.02 -.64 4.17 .58 .35 -1.07
1922-23 .05 .44 -.32 .02 -.94 -.93 -.61 .46
1923-24 -.37 -.22 .20 .26 -.11 .15 -.31 -.18
1924—25 -1.84 -1.39 -.14 -.08 -.38 .08 1.24 1.38
1925-26 1.50 .71 .81 .44 1.07 -.54 -.15 -.61
1926-27 .79 .12 -.02 -.22 1.28 .42 .21 -.31
1927-28 .63 .12 .34 .02 .58 -.57 1.21 .43
1928-29 -.36 -.61 .19 -.13 .60 .11 .96 .63
1929-30 .67 -.23 .18 -.05 1.47 -.56 1.57 .85
1930-31 .90 .36 -.24 -.09 -.91 -1.00 .36 .74
1931-32 .66 .43 -.60 -.38 -.14 -.13 -.95 .08
1932-33 -1.04 -.60 -.22 -.15 .54 .35 -.03 .39
1933-34 -.74 .04 -.34 -.03 -1.73 -.66 -.19 .65
1934-35 .06 .26 .02 .05 -.52 -.06 -.65 — .25
1935-36 -.40 -.64 .14 -.06 2.17 .39 .99 .30
1936-37 -.19 .18 .10 .05 -.52 .26 -.46 -.49
1937-38 .07 .36 .27 .26 -1.52 -.93 .36 .30
1938-39 1.30 .01 .40 -.18 2.24 .69 1.24 -.51
War transition
1939-48 -5.30 -3.94 -.35 .81 -6.61 1.02 3.40 2.11
Postwar
1948-49 1.27 .60 .20 -.02 3.67 1.80 -1.11 -2.38
1949-50 .56 .61 .07 -.21 .51 .23 .04 -.63
1950-51 .56 .85 -.18 .02 -2.82 -1.20 -.35 .33
1951-52 .16 .81 .09 .30 -.65 -1.51 -.54 .40
1952-53 1.55 .56 .49 .02 1.39 -.66 1.46 .08
1953-54 1.28 .24 .38 -.19 3.08 .56 .60 -.61
1954-55 .36 -.75 .27 — .32 3.07 .59 1.90 .48
1955-56 .89 -.20 .80 .07 3.31 .38 1.51 -.25
1956-57 -.69 -.51 .30 .48 -.03 -.28 -.15 .31
1957-58 .69 .56 -.15 .03 -.85 -.88 -.37 .29
1958- 59 .44 -.06 .38 .05 1.17 -.12 .93 .12
1959- 60 .90 -.42 .41 -.22 2.77 .53 1.41 10
1960-61 1.12 .32 .60 .14 1.97 .18 .51 -.63
1961-62 .65 -.01 .35 -.06 1.57 -.07 1.11 14
1962-63 .89 -.10 .48 -.10 2.71 .37 1.24 -.17
Averages over time
Prewar .27 .03 .04 -.05 .41 -.13 .28 15
Postwar .71 .!7 | .30 -.00 1.39 -.00 .55 -.16
Note. All figures are to be multiplied by 10“2.
The theoretical pattern (1.19) is completely determined as soon as tr2 * is

specified, given the /Fs of (3.2). We take
(3-7) a2 = 2 x 10~4
which implies the following theoretical covariance matrix:
32 -4 - 16 - 12“
-4 18 - 8 — 6
(3.8)
- 16 - 8 48 - 24
- 12 - 6 - 24 42
The matrices (3.6) and (3.8) are fairly close to each other. We should take into
consideration that only one parameter (a2) has been adjusted. Also, economic
theory must be expected to be less powerful with respect to second moments
than it is with respect to first moments. Furthermore, there is the problem of
observational errors, which have been disregarded completely but which do
have a disturbing influence when the values of the disturbances vit are
measured. The largest difference between (3.6) and (3.8) is provided by the
variance of the demand equation for Vice, for which the estimate is far below
the theoretical value. No attempt will be made to explain this feature in terms
of the Calvinist tradition of the country.
Another important question is to what extent the systematic part of the
demand equations succeeds in “explaining” the behavior of the dependent
variables. An obvious measure is „ ,
Lvn
(3'9) l~R? = v/*n V
2.(wuD1ti)
where Rt is the multiple correlation coefficient of the zth demand equation. In

the prewar period the ratio (3.9) takes the values .26 (Food), .56 (Vice),
.13 (Durables), and .61 (Remainder); for the postwar period the values are
.34, .25, .12 and .48, respectively. [The numerators of these ratios are the
diagonal elements of (3.4) and (3.5), not those of (3.8).]
A more interesting answer to the same question is obtained when we recog¬
nize that our demand equations deal essentially with changes in value shares.1
1 The analysis described in the remainder of this section and in Section 7.4 is largely
based on H. Theil and R. H. Mnookin, “The Information Value of Demand Equations
and Predictions,” The Journal of Political Economy, Vol. 74 (1966), pp. 34-45. The
numerical outcomes are slightly different, which is partly due to the fact that the demand
equations of that paper are formulated in Dun instead of Dqt, partly to differences in the
numerical values of the quantity log-changes.
242 EMPIRICAL IMPLICATIONS OF TLIE ALLOCATION MODEL
Since these shares are positive and add up to 1, we may decide to measure
the fit of the demand equations by means of information concepts. Let us
indicate any prediction of wit by vi>i(; it is assumed that these predictions are
also positive and add up to 1. Then the information inaccuracy
(3-10) It=iwtt log^"

i=i wu
is the obvious measure for the merits of wit, i— 1, Three sets of forecasts
will be considered. The first is no-change extrapolation:
(3.n) wit = wi>t_1
Demand theory claims that it can achieve more than (3.11) does, because it
makes certain statements on the change Awit. Going back to equation (5.13)
of Section 6.5 and applying the formulation (1.3), we conclude that our
demand equations imply1 *
(3.12) Awit = (/;; - w*)Dqt + w*Dpit + £ vijDp'jt + vu

j= i
If we knew all coefficients on the right and also the values of all varia¬
bles including the disturbance vit, we would be able to obtain a perfect
prediction of Awit and hence, given wu,_u also of wit. It is not realistic, how¬
ever, to assume that the disturbances are predicted perfectly. Let us suppose
that the demand analyst predicts that they will coincide with their conditional
expectations (given the data of year t—1), which are zero under the assump¬
tion that the distuibances are not autocorrelated." If it is assumed, further-
1 Note that in (3.12) Dmt is replaced by Dqt twice: once in the demand equation [see (1.3)]
and once in the price-quantity-income decomposition of the value share change. Note
further that the ~ sign of equation (5.13) is replaced here by =, which is necessary to
obtain definite predictions of Awit and hence of wu.
The Von Neumann ratios of the prewar disturbances are 2.08 (Food), 1.87 (Vice), 2.59
(Durables), and 2.06 (Remainder). For the postwar period they are 1.06, 1.94, 1.01, and
1.10, respectively. The prewar figures present no evidence of nonzero correlation; the
postwar figures suggest some positive autocorrelation. In what follows we shall proceed
under the assumption that two disturbances of the same or of different demand equations
are uncorrelated when they belong to different pairs of successive years. Note that the
significance of the observed means of the disturbances over time can be tested easily under
his condition. The standard deviations of the ffs vary between 4 x 1CF3 and 7 x 1(D3
according to (3.8), so that those of an average of 18 or 15 observations vary between 1CF3
and 2 X 10- The observed means on the last two lines of Table 7.1 are therefore not
sign, cantly different from zero. Hence there is no reason to introduce a trend term in the
demand equations (see footnote 1 on page 230).
more, that perfect forecasts are available of Dq, and Dpu, ..., Dpnt, the
demand analyst’s prediction of Awit is equal to the right-hand side of (3.12)
excluding vit. Hence his forecast of wit is
(3.13) wit = wit — vit
which is the second type of forecast to be considered.

The first line of Table 7.2 deals with the extrapolation method (3.11). It
contains the average prewar and postwar values of It and the single I, value of
the war transition. The second line presents similar data for the demand
analyst’s forecast (3.13). [The third line will be discussed in the next section.]
TABLE 7.2
INFORMATION INACCURACIES OF ALTERNATIVE DEMAND PREDICTIONS
Prewar Postwar War

Forecast wu
average average transition
Four commodity groups

Extrapolation (3.11) 396 556 6082
Demand forecast (3.13) 202 259 3797
Food
Extrapolation (3.11) 121 148 1155

Vice
Extrapolation (3.11) 26 45 2019

Durables
Extrapolation (3.11) 244 377 3007

Remainder
Extrapolation (3.11) 161 204 1831

Note. The unit of measurement is 10 6 nit.

The figures indicate that on the average, both in the prewar and in the post¬
war period, knowledge of the coefficients of the demand equations plus
knowledge of all income and price changes reduces the information in¬
accuracy by about 50 per cent. The absolute figures of the war transition are
larger for understandable reasons. The percentage reduction of its I, ob¬
tained by (3.13) is a little smaller (less than 40) than it was on the average in
the prewar and postwar periods.
The table contains also similar results for individual commodity groups.
They are obtained by concentrating on one value share wit and its comple¬
ment 1 —wit, which amounts to combining all commodity groups other than
the /th. Since 1 — wit is the forecast of 1 — wit, the resulting information in¬
accuracy is
(3.14)
The figures of Table 7.2 indicate that the demand analyst obtains similar
improvements over the no-change extrapolations at the individual com¬
modity level, at least in the prewar and postwar periods. Prewar Vice is a
minor exception (see lines 7 and 8). The result is very bad for Food during
the war transition; that commodity group is primarily responsible for the
rather small percentage reduction that was mentioned at the end of the
previous paragraph.
We add the following comments:
(1) It will be observed that the inaccuracy values for individual com¬
modities are all smaller than the corresponding values for all commodities
combined (first three lines of the table). This is due to
(3.15) 4 < I,
which is proved as follows:
7.4 DEMAND PREDICTIONS 245
Hence Iit It is equal to a fraction 1 — wit of a conditional information in¬

accuracy, the condition being that the ith commodity is disregarded. We have
ht= It when wit= 1. If we have wit< 1, the strict inequality sign holds in (3.15)
except when
wit wit
--- — for each j / i
1 wit
- 1 - wit
in which case Iit — It. This special case implies that for each commodity jî
there is perfect prediction in the following conditional sense: Divide the
expenditure on the jth commodity by total expenditure excluding the expenses
on the /th commodity; this ratio, for each jî, is to be predicted perfectly in
order that Iit = It rather than Iit < It.
(2) It will be noticed that w* occurs twice in the right-hand side of (3.12).
The prediction procedure (3.13) is based on the assumption that w* is known,
which is not really correct, because w* is the average of wi>t_ t and wit, and wit
is the object of prediction. This can be corrected by replacing w* by
which will normally have hardly any effect in time series analysis. If it does
make any difference (in cross-section analysis, say), one can proceed in the
following iterative manner. First, apply the replacement just mentioned in
(3.12), which leads to a forecast w'it of wit. Then replace w* in (3.12) by the
average of w'it and vpijf_l5 which leads to a new forecast w-'r, and so on. The
convergence will probably be very rapid.
7.4. The Information Value of Demand Predictions; The Impact of the Random
Variability of Coefficients and Disturbances
An important point to be considered is that an “ordinary” demand analyst

must be expected to predict below the level of (3.13), since his income and
price forecasts are not perfect. Demand equations are useful for the purpose
of unconditional prediction only to the extent that they are supplemented by
satisfactory predictions of income and price changes.1 It would be interesting
to see what would happen to the information inaccuracy when official fore¬
casts rather than observed values are substituted. This is not possible, how¬
ever, since no such forecasts are available. The following experiment is
nevertheless instructive. One may argue that an accurate prediction of such
microeconomic variables as relative price changes is more difficult than that
of the change in real income. So let us suppose that Dq, is predicted perfectly,
that all relative prices are supposed not to change, and that the disturbances
i -phe case of conditional prediction, given income and price changes, is of course different.
vit are predicted to coincide with their conditional expectations (zero). Hence
the forecast of Awit is equal to the first term on the right of (3.12), so that
(4.1) wit = wiJ_1 + (fit-wft)Dqt
This means that the value share of a luxury (necessity) is predicted to increase
(decrease) when Dqt goes up, irrespective of the - unknown - development
of relative prices.
The information inaccuracy J, as defined in (3.10) can also be computed for
the forecasts (4.1). The third line of Table 7.2 contains prewar and postwar
averages, which are about halfway between the averages of the extrapolation
inaccuracies and the demand forecast inaccuracies of the previous section.
Hence we do lose when nothing is known about relative price changes, as
could be expected, but we do not lose the complete gain over no-change
extrapolation. The latter statement does not apply to the war transition,
which is primarily due to Durables.
A second point which we should consider is that one ordinarily does not
know the true values of the coefficients of the demand equations. We usually
have a set of point estimates and an estimated covariance matrix. The impli¬
cations of the estimation procedure can also be evaluated along informational
lines, although the logarithmic criterion is difficult to adjust to the quadratic
estimation criterion which is implied by the use of variances and covariances.
We can, however, expand the natural logarithm of wit/wit according to powers
of the ratio of wit — wit to wit with the following familiar result:
n
(4.2)
i— 1
This approximation is very accurate even for no-change extrapolations,

because the year-to-year changes in value shares are so small. As a next step,
then, we can take the expectation:1
n
1 y ^ (wit - wity
(4.3) sit
2L wit
1 Note that we disregard the random nature of the right-hand denominator (wit). This
is of minor importance, however, because the random component of wu, given Wi,t-1, is the
disturbance vu of the demand equation; and the standard deviations of the v's are all less
than .01, see (3.4) through (3.8).
which will now be evaluated under the assumption of perfect income and
price predictions. We write fa and vu for the point estimates of and vu,
respectively. The forecast of the itb value share is then
n
Wit = 1 + (A - wtt)Dqt + w*Dpit + X ViftP'jt
j= i
where vit is again predicted to vanish. If we compare this forecast with the
observed wit = wi>t_l +Awit [where Awit is as specified in (3.12)] we find the
following prediction error:
wit ~ wit = (A - fi^Dq, + (vu - vu) Dp'it - vit
where use is made of vij = vij — 0 for i^j. We then square both sides and
take the expectation. On the assumption that A and vu are unbiased and that
they are uncorrelated with the disturbances vit, we obtain
(4.4) £ (wit - wit)2 = (Dqt)2 var fa + {Dpitf var vu

+ 2DqtDpit cov (A, vu) + varvit
The variance of vit will be specified numerically according to the ap¬
propriate diagonal element of the adjusted covariance matrix (3.8). For the
covariance matrix of the v’s we shall take1
V22 V33 v44

4 2 4 3 Vn
2 9 4 4 V22
(4.5) V = 10“4 x
4 4 16 8 V33
3 4 8 16 V44
The diagonal elements of V determine the standard errors. If we interpret the

v’s of (3.3) as v’s in the present context, we have the following estimates and
standard errors:
Food: v11 = -.08 (.02)
Vice: v22 = — -04 (.03)
Durables: v33= — .16 (.04)
Remainder: v44= — .12 (.04)
Note that this specification implies that the second price coefficient does not
differ significantly from zero. Note also that 0, being the sum of all four v’s,
1 These numerical values have a similar approximation interpretation as the /Ts and v’s
of (3.2)—(3.3); see footnote 1 on page 238.
has a variance which is equal to the sum of all elements of V. Hence.
(4.7) var </> = 95 x 10
so that the standard error of 0=— .4 is almost .1. Note finally that the
sampling distribution of
is fully determined by the distribution of the v’s. This is pursued in the

Appendix of this chapter, (Section 7.B), where the variance of /2; and the
covariance of ju, and vH are derived by means of a large-sample approxima¬
tion. In particular, the following standard errors of the marginal value shares
are implied:
Food: /)1 = .2(. 04)
Vice: /x2 = . 1 (.06)
(4.8)
Durables: /t3 = .4(.06)
Remainder: /54 = .3(.06)
The numerical specifications listed in the previous paragraph supply us

with the variances and covariances which we need in the right-hand side of
(4.4). If we then divide by wit and sum over i in accordance with (4.3), we
obtain - approximately - the expected information inaccuracy SI, which is
due to the stochastic character of the coefficient estimates and to the random
variation of the disturbances vit. Finally, we average over time, which gives:
T n
(4.9)
r= i i= 1
T n
t= 1 £= 1
T n
t= 1 i= 1
T n
t =1 i= 1
TABLE 7.3
DECOMPOSITION OF THE EXPECTED VALUE OF AVERAGE INFORMATION INACCURACIES
Description of component Prewar Postwar
Four commodity groups
Total expected inaccuracy 350 362

Due to disturbances 315 296
Due to coefficient estimates 35 66
due to variances of income coefficients 31 57
due to variances of price coefficients 8 6
due to covariances -4 3
Food

due to variance of income coefficient 3 7
due to variance of price coefficient 1 1
due to covariance 1 -0
Vice

due to covariance -3 5
Durables

due to covariance 1 2
Remainder

due to covariance -3 -3

The first three terms on the right represent jointly the effect of the random
variation of the coefficient estimates on the expectation of the average in¬
accuracy I. The fourth term represents the effect of the disturbances of all
demand equations. Each of the first three terms deals with one aspect of the
random variation of the coefficient estimates: the first with the variances of
the income coefficients, the second with the variances of the price coefficients,
and the third with the covariance of /q and vu in each equation. Note that
covariances of coefficients and of disturbances of different demand equations
do not enter into the picture.
The decomposition (4.9) is shown in the first six lines of Table 7.3.
The results indicate that about 80 to 90 per cent of the expected infor¬
mation inaccuracy is due to the disturbance variances, both in the prewar
and in the postwar period. This suggests that our limited knowledge of
the demand function coefficients is not very serious compared with our
inability to predict disturbance values. The contributions of the variances
of the marginal value shares are much larger than those of the variances
of the price coefficients. This is to be ascribed to the larger importance
of real income changes compared with the variation in relative prices.
The covariance contributions are close to zero. They are not necessarily
positive.
One can derive similar results for individual commodity groups. The
analogue of (4.2) is
1 wlt
( - W,)2 1 (1 - wit - 1 + wit)2 _ 1 (wit - w;t)2
2 wit + 2 1 ~wit 2 wif(l — wit)
After proceeding along familiar lines we finally obtain
(4.10) ■ l , L\~ var»„.y (Dptf

2T Li vvlf(1 - wit) IT Lw„(l-wif)
t=i i=i
+ cov (fih fit) y DqtDp'it 1 y* var vit

T U wii(l - wit) 2T [j wit( 1 — wit)
t-i t= i
We conclude that the one-commodity values STt depend only on the variances
and the covariance of the two coefficients of the fth demand equation and
on the variance of the disturbances of that equation. The numerical results are
7.5 MARGINAL PRICE INDEX AND MARGINAL UTILITY OF INCOME 251
shown in Table 7.3. They indicate that the one-commodity decompositions

are largely similar to that of all commodities combined.1
7.5. The Marginal Price Index and the Marginal Utility of Income; Standard
Errors of Their Log-changes
The log-changes Dpt = Zw*Dpit and Dp't = EpiDpit of the cost of living price
index and the marginal index are compared in the first two columns of
Table 7.4. One observes that they are normally very close to each other. The
discrepancies are all less than 2 per cent with the exception of the war
transition; in 29 cases out of 33 the discrepancy is less than 1 per cent. The
third and fourth columns contain the log-changes in the marginal utility of
income, both in real and in money terms:
(5.1) Dl't = 1 * * * Dq, DXt = -- Dqt - Dp'

It'is
simply a multiple 2\ of Dq„ apart from sign, it only reflects the changes in the
volume level in a straightforward manner. The war transition values of DX[
and DXt are interesting. There was a substantial decrease in volume and
hence a large positive value of DX't; but the price increase Dp\ was much
larger, so that the marginal utility of income in money terms decreased
considerably. The postwar values of DXt are mostly negative and rather large
in absolute value, which is due to the combined influence of volume increases
and price increases.
The figures in parentheses in Table 7.4 are standard errors. They are
derived under the assumptions of the previous section; that is, the four p’s
which are used for Dp\ are only estimates (/t’s) and the same applies to </>
in (5.1). The sampling error of Dp\ is then
n
£ (A - Ah) °Pit
i= 1
1 The disturbance parts of the decompositions (4.9) and (4.10) can be regarded as the
approximate expectations of the average information inaccuracies of the demand forecast
(3.13). Hence the corresponding figures of Tables 7.2 and 7.3 should be approximately
equal. A comparison shows that this is indeed the case with Vice as the major exception:
the figures of Table 7.3 are 109 (prewar) and 98 (postwar), those of Table 7.2 are 34 and 22,
respectively, all in 10~6 nit. This is of course due to the fact that the Vice variance of (3.8),
which is used in Table 7.3, is considerably larger than the estimated variances in (3.4) and
(3.5). It also leads to larger figures on the second line of Table 7.3 compared with those
of Table 7.2, which deal with the four commodity groups simultaneously.
252 EMPIRICAL IMPLICATIONS OL THE ALLOCATION MODEL
WO r-ooooONW7ofOco r- on o OO h- NO
CO coonnoononcnoj-o WO NO OO co wq cn <n
CN —* CN co co of co* cn* cn co*
ON CN00co00<0nCNCN©©00ON00 *—< of co
of nONo\O^ÔM\0'TiTtoN^ O —■ o
CN of O co rf no’ wo t—’ cn ’ r-’ wo* cn o
NO
T 7 7 1 1 77
O wo cn o oo no —. (— CO WO NO 4f rf
of h co >0 IN M
s: i/o* CN .
CN co co co —’ co’ cn cn* co’
1*^
Q t VO Qn rf co no *—i ON M on O r-
© q ON co CN co © fN| \0 CO wq CN co CO
CN
£
d cn no’ cn cn co* Tt no* T-H r-’ o ON o'
CN <o
O 7 I
LOG-CHANGES OF PRICE INDICES AND OF THE MARGINAL UTILITY OF INCOME
I On
On ^^°^5200coin«ONOo«co7o
77nhOco(Nrico70 <n — -—1 —
O' 'sO M M NO — IN — On Tt — TT of
NO cn o on r- (NMcoONMOhiCON. 4 nO
of Nt OO 6 H
oo —i CN
<N on co oo r-' ON CO OO O WO (N NO
CN WO © Ov O of r-
h Tl OO IC) q ON NO CN r-- ON
r-’ of oo o
r- CN CN CN
2
oo ON O CN co of l/o NO t"- 00 ON 0 <N
of WO wo
t I I, I
WN NO NO sO NO
> M
TABLE 7.4
On oo
of
On
of
4 OO ON
wo wo WO
CN CO 4 O n S ct
NO NO NO ^ fli
onononononononononononon
ON ON ON ON ON ON -° >
Q. cj
/■-s __
(N on ti D ici co '—o
of Of ON NO (N
TCOCO'OONCOhO of O OO NO r- wo of <N
— __
-—■ s—•- 'w s—✓
- -4 ’ J
I ^
J^Xj-fNE'lONOCN. - IN O of —1 wo wo NO O
O) h M CO ON OO M o co CN ON co wo ON
oo On fN fN ’t h ON NO CN Of
l
WO no oo f oo oo wo fT' r7rosvosC7'C7^-7r'00
NO
NO OO ON 00 OO O
wq rq © vo on ON h
of r- 04 OO CN CO fN
oo io i-7 CO
CN
7"
NO o ON wo
£ of — i CN* oo’ wo’ NO co ON f CN
(N r-7 cn rsj CN
X.
I
r7r^ôvov°fô'ooo
r-wqq<Nro — Oroorof VDfN„„nTt(N(N
'O OO NO (N CO • ON , VO co
— (N Tt — °c ^ ^ O (N (N °° S
h
q 3 S o
t q 9 in
^ on
^ NO
I I* *0 NO ON Tj- T ' CO D7 1/7 04 <
CN
I
ON NO NO ON IT)
—i cn cn no r- WO
o •t | M- '■O 06 rn _I m 'T M" <N
of
7 1
N n 10 \o
7ji77°777777SSS ct> ' J, i i 1 i 1I 1
r: M ^ i- t, o n 00 o o a gp
r-
a
7.6 THE COVARIANCE OF PRICE AND QUANTITY CHANGES 253
On squaring and taking the expectation we obtain the sampling variance of Dp',:
(5-2) var Dp't= £ £ DpitDpjt cov (fih fij)

»=ij=i
after which the standard error is found by computing the square root. The
result (5.2) indicates that the covariance matrix of the marginal value shares
is needed. This matrix is derived in the Appendix of this chapter (Section 7.B)
by means of a large-sample approximation. In the same way we obtain the
variance of Dl't by multiplying (Dq,)2 by the variance of the reciprocal of 0,
which turns out to be .371. Hence the standard error of Dl't is a fraction
V-371 = .61 of the absolute value of Dq,. For the standard error of DX, we
need the covariances of the (Ts and 0, which are also derived in the Appendix.
The standard errors of the log-changes in the marginal price index are
mostly small. Nevertheless, they are such that the significance of the differ¬
ence Dp't — Dpt is destroyed in a great many cases. The standard errors of the
log-changes in the marginal utility of income are much larger, which is due to
the influence of 0 and its relatively large sampling error; but the point
estimates of Dl't and DX, are also mostly larger in absolute value than those
of Dp't. Needless to say, these standard errors reflect only the variability of
the /i’s and of 0; they do not reflect any errors in the price observations,
which are in practice an additional source of uncertainty.
7.6. The Covariance of the Price and Quantity Log-changes; Statistical Testing
of the Income Flexibility
We return to the second moments of the price and quantity log-changes

that were computed in Section 5.5, in particular to the covariance:
n
(6.1) T, = X wn (Dpi, - Dp,){Dqit - Dqt)

i= 1
n
= I (Dpu ~ Dp,)w*,Dqit
i= 1
The second line shows that Tt can be regarded as a weighted sum of the left-
hand variables of the demand equation (1.3). We substitute the right-hand
side under the assumption of preference independence:
n
r, = £ (Dpit - Dpt) (piDqt + vuDpit + vit)

i= 1
The term containing Dq, can be simplified as follows:

n
Dq, Z Pi (Dpi, ~ Dp,) = {Dp, - Dp,) Dq,

i= 1
For the price term we have, using =
<j> t Hi(Dpu - Dp,)(Dpit - Dp') = 0 X Vi(DPit ~ °Pt)2

i= 1 i= 1
which can be written as <^77', where
(6.2) n't = X Pi(DPit ~ DPtf

i= 1
is the logarithmic variance of the price changes with the marginal value shares
as weights. We conclude:
n
(6.3) T, = (Dp; - Dpt) Dqt + + I (.Dpit - Dp,) vit

i= 1
The covariance of the logarithmic price and quantity changes can thus be
written as the sum of three terms. The second is always negative except when
all prices change proportionally (77; = 0), in which case it vanishes. That term
is the substitution effect of price changes on rt: If some prices increase more
than others, this leads to changes in the quantities consumed in such a way
that there is a negative correlation between the price and quantity log-changes.
However, this conclusion may have to be modified when real income changes.
This effect is represented by the first term on the right of (6.3). Suppose that
the prices of luxuries increase more (or decrease less) than those of necessities,
so that Dp\>Dp,. Then the consumption of necessities is stimulated at the
expense of luxuries when real income remains unchanged. But suppose also
Dqt> 0 ; then there may be an increase in the consumption of luxuries in spite
of the price changes, which contributes to a positive rather than a negative
sign of rt. This is shown by the first right-hand term of (6.3).
The last term in (6.3) is a weighted sum of the disturbances of the demand
equations, which has zero expectation under condition (1.13). Using the
covariance specification (1.19) we can derive its variance as follows:
£ £ (DPn ~ DPt)(DPjt ~ Dp,)S(vitVjt)

i=lj=l
n
£ Pi{DPit - Dpt)2 - £ £ PiPj{Dpit - Dp,)(Dpjt - Dp,)

Li= 1 i=lj=l
n
= cr £ PiiPPit - Dp,)2 - (Dp; - Dp,)2
.i = 1
= fT2 £ Pi(DPi, - Dp',)2

i= 1
£ a>
r-~ ro i—i 1—1
O £ ON <N OO ro i-H r- o
__,
00
-a o OO o 1 CO CO OO NO o O r- (N
£ Q, r-’ 1 <o r- 05 CO (N •
| 1
1 1 1
P4 o i
o
£
O
VO r- 05 (N 05 r- ON oo ro 50 Nt to 50
OO oo T-H T—1 CN 05 NO t/5 to (N r- (N 50 »o
3 £
•— !—i oo 1—! CN <N NO 1 cn | T—( (N CO 1 CN*
+-> (D co 1 1 1
CN 1 1 1 1 1 1 1 1 1
X) t
to 53
m R £
Q
<D A
DECOMPOSITION OF THE COVARIANCE OF PRICE AND QUANTITY LOG-CHANGES
£ £ X) ro CN ON OO NO OO CO 50 OO r- r- NO CN
o £
1 ON CO O o 00 r- r- ro NO o r- 05 50 05
O <lj on 1—t 1 CN 1—1 1 <N t-h CO
£ 50 1 1 1 l 1 1 1
HH 1 1 i i t
1
<D
r- CN ,—i r- i—H o r- r- 00 50 o to "rt- ro o 50
aj ON 50 CN o o CO CN ro (N CO CO 05 O r- T—1 o ro
y——i CO CN CN to 1 oo CN r- d ro’ CN
d CN 1 T— 1 1 1 l 1 1 1 1
> co 1 1 1 1 1 1
O 1
u
oo 05 O <N CO «o 50 r" oo 05 o <N ro u

<D
on on on to to »o to to to to 50 50 50 NO d b£)
T t 1 1 1 1 1 1 1 1 1 ! 1 £ d
05 OO On © CN CO to 50 r- OO On o IN
C/3
co N- ■*fr to to to to to to t/> »o to «o 50 50 50 a>
On 05 ON 05 05 05 05 05 ON 05 05 05 05 05 05 05 o >
1 1
t“H T—l l—t T—l T—' *“*
i—i i—i
T“l
Ph d
a tL)
£ g
oo on r—, T—1 r-< r- 50 NO to OO to Ol ro r- 05 CN d-
T3 o ON o ’—1 ro ro OO 05 i 50 to CO ro 05 (N CO CN
e & OO 50 50 ro 1 (N <N 1 ro* (N CN ro* T—l
£ £ 1 1 1 1 1 1 1
* o 1
o
£
O
05 CN ro y—4 «o 05 t"- CN r- OO o NO O 50 (N ro to
3 6 o q 00 o OO oo N; On *o (N nf q 05
.t- J-H ro’ »o oo CN ro*
-*-> <D 50 1 ro’ 1 1 1 (N 1* f CN
C/3 -*-* 1 1 1 1
Note. All figures are to be multiplied by 10~4,
-D 1 1 1 1 1 1 1 1 1 1 1
£ i- 1
m
£
x.
ft,
D
CN r- o o 50 oo VO ro ro r-> r- On <o ro
E £ CO 50 o CO CN NO ro o <N O 50 T" oo o 50 50 ro to T-1
O j-i CN ro to
1—1
O
£
HH
<D
^
J
7 ! 1
| |
1 l
0)
o
£ o <N to ro ro O oo O On ON i—( ro to oo 50 ON ■nh NO
o O oo 50 05 o y—i <N i—i r- NO O T-H <N 05 CN
S—i
© CN 50 1 1 1 i—i to o Nj* 1 <N T—t to’
> ro 1 1 i I l t“i ! ’— | I I
O l 1
u
to 50 r- CO 05 o (N ro to NO r- oo 05
CN CN <N (N CN <N CN <N CO ro ro ro rî ro ro ro ro ro bC
i 1 1 1 1 1 1 d
4- to 50 r-~ OO 05 O <N ro 4 to 50 r- oo L*
CN CN CN <N CN CN CO ro ro ro ro ro ro ro ro (U a>
05 05 ON ON 05 05 05 05 05 05 On ON 05 On 05 05 >
1—H T-H l-H i-^ d
Hence:
(6.4) S Z (°Pit - °p>)t = a2n't
which shows that the variance of the random component of Tt is a multiple

a2 of precisely the same price variance 77) as that in the second term on the
right of (6.3). We conclude that this random component vanishes identically
when all prices change proportionally. As a matter of fact, all three terms in
the right-hand side of (6.3) vanish in that case, which is in accordance with
rt = 0 when Dpit — Dpt, i— 1, ..., zz.
The covariance and its three components are shown in Table 7.5 for the
four-commodity example. The decomposition is based on 4>= — .4 and the
fi’s of (3.2). The results indicate that the income term (Dp't — Dpt)Dqt is of
minor importance in most of the prewar years, which is at least partly due to
the fact that Dqt was usually close to zero in that period. It is of greater
importance in the postwar period and it even exceeds the substitution term in
absolute value in 8 out of 15 observations. The figures show also that the
disturbances of the demand equations play an important role in the co-
variance, particularly when the price variance 77] is large [which is in ac¬
cordance with (6.4)]. There are, as a whole, six positive covariances rt; four
of them owe their sign to the demand equation disturbances and would have
been negative if these disturbances had been zero. If we divide the random
components by their theoretical standard deviations [i.e., by cr^/77], see (6.4),
where a is specified according to (3.7)], they should theoretically have unit
variance. The observed standard deviation of these standardized random
components is 1.07 in the prewar period and .93 in the postwar period, which
is a satisfactory agreement.
It is quite interesting to consider (6.3) as a regression equation in which
n't acts as an independent variable. Let us assume that the marginal value
shares are known, so that Dp\ and 77] can be computed; then we can use the
specification (6.3)-(6.4) to test any hypothesis about (f>, e.g. the hypothesis
0 = - • 4. The procedure is quite simple: Correct the covariance Tt for changes
in real income by subtracting the first term on the right of (6.3) from both
sides - this term does not contain any unknown parameters - and apply
weighted least squares based on (6.4):
T
A
(6.5)
l ji;Dr,-(Dj>;-pR)j)g,]
?n:
»=i
Z [rt-(Dp't-Dpt)Dqt] £ Ct
t= i__ _ t= i
t r
z n't z
t=i i=i
where
(6.6) Ct = Tt- (Dp, - Dpt)Dqt
is the price-quantity covariance corrected for the income effect of the price
changes. We conclude that the weighted least-squares estimator of </> is
simply the sum of these corrected covariances divided by the sum of the price
variances, the latter being marginally weighted. On comparing (6.5) with
(6.3) we find that the sampling error </>* — <f> is equal to the sum over t of the
residual component of r, divided by the sum of Wt. Hence the error has zero
mean and its variance is
) ( n )
IP
t=i t'= i b
Z (DPit DPt)vit ^ I z (Dpjt' - Dpt)vjt, |
Z
t= i
(DPu ~ DPt)(DPjt ~ Dpty(vitvjt)
which in view of (6.4) can be simplified as follows:
(6.7) var 0' =—-
Z
t=i
n’t
Note that we made use of the assumption that vit and vjY are uncorrelated
whenever t^t'.
If we apply (6.5) and (6.7) to the 18 prewar observations [with o2 as
specified in (3.7)], we obtain a </> estimate of - .56 with a standard error of .12.
The estimate based on the 15 postwar observations is —.23 (with standard
error .15), which is much closer to zero. The difference .33 is not significant,
however, since its standard error is as large as .19. [The square of this standard
error is computed as the sum of the variances of the prewar and postwar
estimates.] The outcomes are fairly sensitive to changes in the set of obser-
vations; if we delete the first observation, the prewar estimate changes from
-.56 to -.43. For all 33 prewar and postwar observations as a whole the
estimate is — .42 with a standard error of .09. This is in close agreement with
the —.4 value which has been used since Section 7.3 and also with the
variance (4.7).
7.7. Moments of Log-changes Based on Marginal Value Shares; The Luxury-

Necessity Index
In the previous section we suddenly met TTt, the variance of the price log-
changes which is obtained when we use the marginal value shares ju; instead
of the “ordinary” shares as weights. This is the direct second-moment
extension of the first moment Dp\ which we already met in the previous
chapter when formulating the demand equations. The question arises what
can be said about such “marginal” moments in case we apply them to
quantities as well. This leads to
n
(7.1) Dq't = X ^Dqit

1=1
(7.2) K\ = X MDqit-Dq')2
i= 1
n
(7.3) r; = X Pi(DPit - Dp't)(Dqit - Dq't)

i= i
The second moments Tl't, Kf T't are shown in Table 7.6 for the four-com¬
modity example. They are presented together with their counterparts
nt, K„ T, which are based on the weights w*. We conclude from the figures
that corresponding moments are usually fairly close to each other. The
“marginal” covariance T't is mostly negative, just as its “ordinary” counter¬
part rt.
Consumer demand theory says nothing about 77', but it does have some¬
thing to say about all moments in which quantities play a role. We shall
consider this in particular for the first moment Dq't, which will be written as a
deviation from the “ordinary” first moment Dq,\
n
(7.4) Dq't — Dqt = X (P-i ~ w*)Dqit

i= I
The difference Dcft — Dqt = D(q't/q,) will be called the log-change in the
luxury-necessity index. The interpretation is as follows. We observe that in
the right-hand side of (7.4) the log-change Dqit is multiplied by a positive
7.7
VARIANCES AND COVARIANCES OF PRICE AND QUANTITY LOG-CHANGES WEIGHTED BY VALUE SHARES AND BY MARGINAL VALUE SHARES
THE LUXURY-NECESSITY INDEX

259
coefficient /q-w* when the ith commodity is a luxury, by a negative co¬

efficient when it is a necessity. Hence, when the quantity log-changes of
luxuries are more positive (or less negative) than those of necessities, both
sides of (7.4) will be positive; in the case of the opposite movement the sign
will be negative. In the former case there is a change in favor of commodities
whose quality is regarded as better from the consumer’s standpoint (because
he allocates a larger fraction of his income to them when he can afford to
spend more), in the latter case the change is in favor of commodities which
are less desirable. Therefore, we can regard the luxury-necessity index as a
measure for the quality of the consumer’s basket.
We proceed to express the log-change of this index in terms of income and
prices. For this purpose we substitute the demand equations into (7.4):
Dq't ~ Dqt= ) ~ w*tDqit - Dqt

i= 1
hi
* (l*iDqt + <PHiDplt + vit) - Dq,
i= 1
hi hi
* 1 )Dqt + <l>) %Dp'it +
Wit W: w
i=l i= 1 i= 1
where use is made of the preference independence implication vi; = </>/q. The
coefficient of Dqt can be regarded as the weighted variance of the income
elasticities /q/w* with the value shares vv* as weights:
hi hi hi hi
vv it l * jt
W,
* = ; w;,it I . * - 1 - 1
' W it LU WJty \Wit J Lt wit
i= l j= 1 i=l j=l
the weighted mean of the income elasticities (with the w* as weights) being
unity. We conclude
(7.5) Dq, - Dq, = Dq, ) vv*

Vlt (l ^^ - 1
\ vv,*,
i= 1
n
hi \ hi
, ,* + / -*V«
■i W;r
i= 1
L wu
i= 1
The first term on the right indicates that the luxury-necessity index goes up
7.7 THE LUXURY-NECESSITY INDEX 261
when real income increases - which is as it should be. Given Dq„ the increase
of the index is larger when the variance of the income elasticities is larger.
This is also understandable: When all income elasticities are 1, an increase in
real income has no effect on the index because /q — w* = 0 for each i; when
the income elasticities of some commodities are much larger than 1 and those
of certain other commodities much smaller, a real income increase raises
ceteris paribus the quantities of the former substantially relative to those of
the latter, and hence raises both sides of (7.4) as well. The ceteris paribus
clause is violated when there are changes in relative prices. This is represented
by the second term of (7.5), which is the substitution effect of price changes on
the luxury-necessity index. It is equal to a multiple (p of the weighted co-
variance of the price log-changes and the income elasticities, the marginal
value shares being used as weights. Given the negative sign of (p, we conclude
that when the prices of luxuries go up relative to those of necessities, the
index will decrease - which is also as it should be.
The random component of Dq't—Dqt has zero mean under assumption
(1.13). We use (1.19) to find its variance:
We conclude that the variance of the random component of Dq, — Dqt is

equal to a multiple a2 of a weighted variance of the income elasticities. The
weights are now the marginal value shares nh not the w* which we had in the
variance by which Dqt is multiplied in (7.5). It is interesting to observe that
TjU^/vv* is both the first moment ol the income elasticities when the s are
taken as weights and also the second moment when the wit are taken instead.
If we take the n’s of (3.2) and w’s equal to .3 (Food), .1 (Vice), .25 (Durables),
r-
co
»n^cor— coinOoocoONOOc©<N»n^H co
— Ooo — NOroinoJOvoOco — Ord O
I 7 I
O tîncocovoaNr-inoNt^
— OcnoNOm — (Nr-d —
n oo n
3 £ oo © o —
CO
JO I
P
GO
<3
s
<D in os © <N<N'^-rniorîooooo^« (N
B E CO m —< <N t'-r^SDVOOOmi/Ir'-ir'lr}- co
o C
O V £
I I
c ~ I
I I
z ,—s /'—*s ,,—s ,—N ,—s ,_, ___

O r- 1 ’—i ■3* ON oo oo o (N ON <N o — •n CO
H o ON *■—< in i—( O in «n (N CO CO
CO Q (N
o |
W w w w w w w w W w
I OO VO NO m NO T“H NO o 00 oo OO r- NO
o Q
r- r- o CO CO CO ' (M NO CO ON in CO t— of-
u
w co’
Q i 7 '
1
oo i .
r~ Q On
oJ-
o
in in
(N
in
CO
in in
in
in
VO
in
r*^
»n
oo
m
ON
«n
o
VO NO
CN
VO NO
t W)
- 5 ON oo On O <N CO 4- «n vo r- oo ON o
1 cj
W x of- in in «n in in m in m «n in VO vo NO CO
ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON O >
w
Q
T—< ” i—l CL, d
Z
I—I
a B o TT ON m o r^* OO OO CO On »n vo
X/lizjb o a © (JO (N CO «n CO «n o (N CO r- o CO o r-
T3 i i
W
U
w
C I I
1 T 1
2
i C2
>-
D
X
'sf r—f *n t-*" on no vo xj' t— r+ in »— no oo on
P (Nin©p<Np©Oco©co«n*— o ? S S K
£ X)
P
C/3
Q
£
<u
B £
<N on ©oooN«n<N<Nc4©inoooo(Ninooo
<N <N C1^rrltr1^0fN«'0(NhfSM'-i
8 5
P ■+-* I I
. . " . . 9 ^ ^ ^ ^ ^ ^ '—< in ’— co
I
*n O (N Tt O O O ON O Tj* o —' (N (N '— oo -i (N
^ NO i—i (N CO NO O co in o *—i ^ ro co on o) >n n
(N co ON <N CO in NO r-
(N (N <N CO CO CCN CO CO CO CO u P
I I I 1 1 1 1
CO
d d bo
<N
(N (N (N (NJ (N (N NO OO
<N <N
T-H <N
co
CO
CO
4
C<1
»n
CO
vo
1 1
OO
CO CO
ON ON ON On ON ON ON ON On ON ON ON ON ON ON ON ON Lh >
' ' T“H r—l cl d
7.8 EQUIVALENT INCOME CHANGES 263
and .35 (Remainder), we find Zfif/w* = 1.1305. Hence the variance of the in¬
come elasticities with the w*’s as weights is .1305 and the corresponding
standard deviation is .36. If we use the /.i’s as weights as in (7.6), the variance
becomes .1554 and the standard deviation .39.
The log-changes Dq[-Dqt are shown in the first column of Table 7.7. The
figures in parentheses are standard errors similar to those of Table 7.4 for
Dp\. They serve to indicate the effect of the random variation of the estimates
of the marginal value shares. The Dqit are taken as fixed for this purpose, so
that the formula used is simply
(7.7) var(D^; - Dqt) = var Dqt = £ £ DqitDqjt cov(&, fa)

i=lj=1
The upper left figures thus indicate that the luxury-necessity index went up by
almost \\ per cent in 1921-22 and that the standard error of this increase
percentage is slightly larger than 1 per cent. It is seen that on the whole these
standard errors are not particularly small compared with the point estimates.
The other columns of Table 7.7 present the decomposition (7.5). One
observes a substantial decrease of the index during the war transition, which
is mainly due to the increase in the prices of luxuries relative to those of
necessities; the decrease in real income was of secondary importance. After
the war the changes in relative prices were in opposite direction, in most of
the years at least, so that the substitution effect on the index was almost always
positive. The income changes were more important, however; about two
thirds of the postwar increase of the index is due to the increase in real in¬
come.1 The prewar figures are distributed rather evenly around zero. This
holds in particular for the substitution terms; their absolute values are less
than the corresponding values of the random component in 13 years out
of 18. If we standardize these random components by dividing them by their
theoretical standard deviations according to (7.6), we obtain 18 prewar and
15 postwar standardized residuals whose standard deviations are .79 and 1.15,
respectively.
7.8. Equivalent Income Changes

Other interesting results on first and second moments can be obtained
when we answer the following question: Take, for a particular commodity,
1 The average annual increase of the index in the postwar period is .0045 or slightly less
than i per cent, and the standard error of this estimate is .0019. Hence the figure is margin¬
ally significant. [This standard error is obtained by interpreting the Dq"s of (7.7) as the
postwar averages.]
the quantity component w*Dqit of the value share change as given; how large
should the change in income be to produce this given value w*Dqit if we
imagine that prices remain constant and the disturbances are zero? The
answer is simply
wf,Dqit = Dqit
(8.1)
Pi HilK
which will be called the equivalent income change of the ith commodity during
the transition from t— 1 to t. It is simply the quantity log-change Dqit di¬
vided by the income elasticity of the same commodity and the same period.
If the actual income change were equal to (8.1), and if all prices were indeed
constant and all disturbances zero, the quantity component of the /th value
share change would coincide with the value which is actually observed. Note
that there are n different equivalent income changes, one for each commodity.
Note also that (8.1) is derived directly from (1.3) and that it is therefore in¬
dependent of the special assumption of preference independence. In fact, the
derivations which now follow are not confined to the preference independence
case, at least not until equation (8.5) below.
We shall show that the equivalent income changes play a natural role in the
theory of first and second moments of the log-changes. For this purpose we
write the demand equations in vector form:
w*,Dq i, DPu
* *
= (Dq,)\i + N
_w*tDqnt _ _DP'nt_
where N and p are defined in (1.8) and (1.10). We then premultiply both sides
by cf)N_I and use Ni = </>p (or equivalently i = 0N_1p):
w*uDqlt DPu
\ *
<£N-1 = (Dq,) i + (/> + (j)N 'v,
_w*,Dqnl _ _DP'm_
Next we premultiply by the transpose of the vector of left-hand variables, to

be written d( for notational convenience:
(8.2)
The result is
(8.3) d(' (</>N ^d, = (Dqt)2 + 0 £ w*Dp'itDqit + d;(0N_1)v,

i= I
7.8 EQUIVALENT INCOME CHANGES 265
or if we subtract (Dqt)2 = dju'd, from both sides:
(8.4) d^N-1 - u')dr = (/> £ w*DpuDqit + d;(0N_1)vf

i= 1
Consider now the left-hand expression, to be written V„ under the special

assumption of preference independence (so that is a diagonal matrix
with l/ft on the ith diagonal place):
n
(8.5) Vt = d^N-1 - n')d, = V (™*Dqi‘)2 _ (Dqj2

L ^
i= 1
n
i— 1
n n
On comparing this result with (8.1) we conclude that Vt is the weighted

variance of the n equivalent income changes, the weights being the marginal
value shares. Also, the average of the equivalent income changes is the log-
change in the volume index, provided we use the same weights:
n
w*Dqu
(8.6) Dq, = Y Mi
i= 1 Mi
Our second conclusion is therefore that Dqt can be regarded as a weighted

average, both of the log-changes Dqit (with weights w*t) and of the equivalent
income changes (with weights ^,).
The third conclusion will be that C„ the price-quantity covariance cor¬
rected for the income effect of price changes, has a similar double inter¬
pretation. It was defined in equation (6.6) ir terms of T„ the covariance based
on the w* as weights, but there is also a second interpretation:
(/o8.7n\) v rn n ^w*Dq‘>
2. Vi(DPit - DPt)-
i=i M;
= Z w* (DPit - Dp,)Dqit - (dp', - Dp,)P>q, = ct

i= 1
Hence C, is also the weighted covariance, marginally weighted, of the

equivalent income changes and the price log-changes.
We now return to (8.4) and (8.5). It should be clear from these equations
that the variance Vt of the equivalent income changes is a measure of the
degree to which the demand for the individual commodities is affected by
factors other than the change in real income: V, vanishes when each log-
change Dqit is proportional to the corresponding income elasticity p-Jw*t,
but it is positive in all other cases. Hence Vt depends only on the substitu¬
tion effect of price changes and on the disturbances. This substitution effect
is the first right-hand term in (8.4), which can be evaluated as follows:
n
(8.8) </> X w*t (Dpit - Dp't) Dqit
= ^ Z w* (DPu ~ Dp,)Dqit (Dp't - Dpt)Dqt

i= i
= 0Cf
see (8.7). Hence the substitution component of the variance V, of the equiva¬
lent income changes is equal to the negative multiple (f) of the covariance of
these changes and the log-changes in prices. To evaluate the disturbance term
of (8.4) we substitute the right-hand side of the demand equations for dr',
which gives
tf'ObT1)^ = (Dqt)n' (<pN~ 1)vt + </> £ Dpitvit + v;(0N_1)Vr

i= 1
n
= </> Z DPitvit + v('(0N_1)v,
where use is made of p (<£N 1)vf = i'vf = 0. On combining this result with
(8.4), (8.5), and (8.8) we find
(8-9) vt = (f>ct + 0 Z DPitvit + y't (0N~J)v,
We conclude that the variance Vt of the equivalent income changes consists

of three parts. The first is the covariance term which was discussed above.
The second is a linear function of the disturbances, the third a weighted sum
of squares of these disturbances with the reciprocals of the marginal value
shares as weights:
v;(0N->(=y ^
L Pi
i= 1
Additional simplifications are obtained when we use the fact that the linear
expression in the disturbances is, apart from the factor precisely the same
7.8
THE VARIANCE OF THE EQUIVALENT INCOME CHANGES AND ITS DECOMPOSITION

EQUIVALENT INCOME CHANGES
Note. All figures axe to be multiplied by 1CL4.

267
as the random component of the price-quantity covariance Tt. This follows

from equation (6.3), which can be written in the form
n
(8.10) c, = 4>n't + X (Dp* - DPt) vu

i= 1
= W + £ (Dpit - Dp')vit
i= 1
where use is made of
£ (Dpt - Dpt) vit = (Dp, - Dpi) £vit = 0

i= 1 »'= 1
On combining (8.9) and (8.10) we then find
(8.11) Vt = 2(j)Ct - 4>2n’t + v; (0N_1)vt
The various terms of (8.11) are shown in Table 7.8 for the four-commodity
example. The left-hand variable Vt is obviously nonnegative. The first term
on the right is positive in the normal case C(<0, the second is always non¬
positive. The sum of these two terms is usually positive, because 20C( —02/7'
is equal to <^Cr except for a disturbance combination which has zero ex¬
pectation. The last term of (8.11) is a positive definite quadratic form in v„
the expectation of which is o2(n — 1); see (1.18). If we use u2=2x 10-4 the
expectation is thus 6 x 10-4; the prewar and postwar averages of the corre¬
sponding sample values are 4.2 x 1CT4 and 5.0 x 10-4, which is slightly less.
The table shows that the disturbance component accounts for a substantial
proportion of Vr. This proportion is almost 50 per cent for the war transition;
it is even larger for the prewar and postwar average figures, particularly for
the latter. Hence the disturbances were more prominent as factors which
determine demand (corrected for the change in real income) than the changes
in relative prices, which is in accordance with the results of Section 7.4.
7.9. The Coal Miner Families Revisited: Cost of Living Comparisons at

Different Real Income Levels1
We return to the coal miner families of Section 5.2 and to the one¬
dimensional price and quantity scales which were presented in Table 5.4 of
Section 5.3. The price scale was not derived in accordance with the require¬
ments for the true cost of living index that were discussed in Sections 6.7
1 The analysis described in this section was carried out by the author in association with
T. Kloek of the Econometric Institute.
7.9 THE COAL MINER FAMILIES REVISITED 269
and 6.8, because no adjustments were made for the fact that families in
different regions enjoy different standards of living.1 The purpose of this
section is to make such adjustments and to find out to what extent there are
different purchasing power parities for different utility levels. The approach
is basically the same as that of Section 6.8 when we proved that Dp, is a local
quadratic approximation to the log-change in the true cost of living index
when this index is evaluated at the intermediate utility level U*. This level
turned out to be very convenient. Here, however, we shall not select a
convenient level but those utility levels which are actually observed for the
coal miners in the six regions.
Our starting point consists of the prices pia and pib (in vector form pfl
and p6), the incomes ma and mb, and the value shares wia and wib in two
regions. We are interested in the cost of living in the ath region with respect to
the bth as base evaluated at the utility level of the c‘h region:
PC(Pa\Vb,Uc)
where
(9.1) Uc — Uj (mc, pc) or mc=mI(Uc, pc)
We shall approximate the logarithm of this true cost of living index by means
of lemma (5.8) of Section 6.5, using the elasticity result (8.3) of Section 6.8:
(9.2) log Pc (pjpft, Uc) = log mI (Uc, pfl) - log mi (Uc, pb)
w-(Uc,\ia) + w-(Uc,-pb) Pia

log
Pib
i= 1
which shows that we should find the value shares corresponding with the
utility level Uc and the price vectors pa and pb. We thus imagine that, while
the prices of the ath region remain as they are, the income level of that
region’s coal miners is modified such that their standard of living is equal to
that of their colleagues in the cth region. Hence the income change is from
ma to mj(Uc, pfl) and the associated change in the ith value share is
W,° (Uc, pfl) - Wia - w° (Uc, pfl) - w,° (Ua, pj
[Mi - Wia] + [/'< - Wi(Uc,Va)] J mi(Uc,pa)

m„

which is based on lemma (5.8) and on dwjd(log = — wt. We now solve

for Wi(Uc, pa), which gives
mi(Uc,Va)
log
m„
(9.3) wHUc,Va) ~ Wia + - Wia)
mi(Uc, pa)
1 + i log
mn
We proceed in the same way to find w®(Uc, pb), so that the expression on the
second line of (9.2),
»,u(C,p.) + w?(l/c,p 6),

(9-4) n. (C) =
ab -log —
Pib
i= 1
is thus approximated in terms of the basic concepts: Uc, pfl, pb,ma,mb, wia
wib and Hi. We can apply this result in principle to any triple a, b, c— 1,..., N,
where N is the number of regions compared. Doing so, we shall usually find
that these binary index values do not satisfy the requirements of the one¬
dimensional scale. They should do so theoretically, because this is one of the
properties of the true cost of living index:
mj(£7c, p2) mf(E7c,p3) = irti(Uc, p3)

mi(Uc, Pi) mI(Uc, p2) mÛ,, px)
Therefore, we apply the procedure of the adjusted one-dimensional scale:

N
(9-5) *,(0 = 3 V n,b(Uc)

b= 1
which has evidently a better theoretical justification than it had when we
applied it to the unadjusted data in Section 5.3.
The actual application of the formulas (9.3)—(9.5) requires knowledge of
the marginal value shares /i;. The approach to be followed is quasi-Bayesian in
the sense that the uncertainty with respect to these parameters is handled by
means of a prior distribution. [This is only quasi-Bayesian, because we do
not draw a sample of observations to revise our prior ideas to posterior
convictions.] Specifically, there are n = 21 commodities and hence as many
unknown /Ts. Their prior means fa are given in the third column of Table 7.9.
They are derived on the basis of a luxury-necessity appraisal in terms of in¬
come elasticities (see the first column). These elasticities are of the form fiJWi,
hence they should be multiplied by value shares wu for which purpose we take
the average of wn, wi6 of the six regions. The individual yvia’s are given in
TABLE 7.9
PRIOR MEANS OF THE MARGINAL VALUE SHARES OF 21 COMMODITY GROUPS
Marginal
Income Average
Commodity group value share
elasticity value share
(fid
1. Bread, rice, spaghetti and other wheat

products .1 .071 .01
2. Meat, meat products and fish 1.1 .113 .13
3. Butter, margarine and other fats
and oils .6 .056 .03
4. Milk, cheese and eggs .6 .063 .04
5. Vegetables, potatoes and fruit .8 .062 .05
6. Other foods .8 .032 .03
7. Beverages (alcoholic and
nonalcoholic) .8 .064 .05
8. Tobacco and tobacco products .7 .029 .02
9. Men’s clothing 1.4 .036 .05
10. Women’s clothing 1.7 .031 .05
11. Other clothing (including children’s
clothing) 1.4 .039 .06
12. Footwear 1.3 .029 .04
13. Furniture and linens 1.5 .050 .08
14. Flousehold appliances 1.9 .035 .07
15. Earthenware, glass, etc., and house
repairs .9 .030 .03
16. Coal, gas and electricity .7 .059 .04
17. Elouse rent .7 .072 .05
18. Cleaning and repairs of clothing .5 .024 .01
19. Personal and medical care 1.0 .022 .02
20. Education and recreation 1.6 .044 .07
21. Private and public transportation 1.8 .038 .07
Table 5.2, the averages in the second column of Table 7.9. The prior means
are then found by multiplication of the appropriate elements of the first two
columns and by rounding off to multiples of .01 such that T/fi = 1.
Prior variances and covariances are specified according to the multinomial
pattern:
S (hi “ A)2 = hi (! ~ A)
S [(fi; - fV) (fij - fljjl = ~ ôhiij i /j
This is a very simple pattern. It implies in a sense that the prior uncertainty
with respect to all marginal value shares is uniform. This will not always be
realistic; if it is not, alternative although more complicated specifications are
possible, but such refinements will not be considered here. We must now
formulate a numerical specification of <r0, which is the basic uncertainty
measure of the prior convictions. We take
(9.7) = .03
On combining (9.6) and (9.7) we conclude that the prior standard deviation
of a Hi which corresponds to /q=. 1 (a high value according to Table 7.9) is
.009, which amounts to a coefficient of variation of 9 per cent. If we take
^. = .01 (the lowest value) we obtain a prior standard deviation of .003 and
hence a coefficient of variation of 30 per cent. Therefore, when an additional
amount of income is expected to lead to a comparatively small addition to the
expenditures on some commodity group, the uncertainty with respect to this
addition is small in an absolute sense but large in a relative sense, which
seems quite reasonable. We complete the specification of the prior distribution
by stating that it is multinormal.
The logarithmic price index numbers na(Uc) defined in (9.5) will now be
presented numerically on the basis of the prior means Given that na(Uc)
depends linearly on the /t’s via (9.4) and (9.3), these numerical results will be
prior means na(Uc), a— 1, ..., N, of the multinormal distribution of the
N logarithmic price indices corresponding to the utility level Uc, c= 1, ..., N.
We conclude from (9.3) that we still need the logarithm of the ratio of
mi(Uc, pfl) to ma. This logarithm will be approximated by
N
b= 1
which is the logarithmic volume index of the c<h and ath region according to
the one-dimensional scale of Table 5.4, Kab being the binary index defined in
equation (1.10) of Section 5.I.1
1 Note that the ratio of mi(Uc, p«) to ma = nii(Ua, pa) is, in the language of Section 6.B
of the Appendix to Chapter 6, the true index of real income of the cth region with the ath as
base evaluated at the price vector of the latter region. The approximation of its logarithm
by means of (9.8) neglects the special role of this vector and it involves TV — 2 other regions
than the ath and cth. Note, however, that the leading term in the expression (9.3) for the
adjusted value share is Wj0, which is independent of the difference of the standards of
living of the two regions, so that the adjustment is a second-order effect. A small error
in the adjustment leads therefore to a third-order error in the adjusted value shares.
/
O
E
W)
o
<D
-d
<D
u.
£
O
Lh
Dh
o
PRIOR MEANS AND STANDARD DEVIATIONS OF LOGARITHMIC PURCHASING POWER PARITIES
<D
-d
d
<D
d
CD
5-
d
CX
d
C/5
<D
s-
d
CJ)
<d
.d
H
-O
<D
d
£
CD
-O
o0)
d
(D
S-H
(S ^
TS jd
> a
ij H
a g
2 uc
-§ £
t3 Sc
60 «
° E
<D J5
H ° >
I
The results are given in logarithmic form in Table 7.10. The first column
contains the one-dimensional price scale of Table 5.4 (the case of the un¬
adjusted value shares). The next six columns deal with the price scales
corresponding to the utility levels of the six regions arranged in increasing
order. The figures in parentheses are standard deviations; they will be dis¬
cussed in the next paragraph. The last column contains the marginal price
index whose logarithmic binary form is
(9.9) n'ab = £ ^ log Pia

i— 1 Pib
The figures of that column refer to prior means nab, obtained by replacing /q
in (9.9) by /q. We conclude from the table that the “raw” scale of Table 5.4
is fairly accurate. When the value shares are adjusted to utility levels ranging
from that of Italy to that of Belgium, only small deviations (all less than
1 per cent) are observed. When we move from left to right in the table, so that
the utility level increases, the cost of living index changes in the direction of
the marginal index. This is in accordance with equation (8.7) of Section 6.8.
It is also seen that the marginal index may deviate rather substantially from
the cost of living index. For example, take France and Italy, for which the
logarithmic difference of the marginal index is .1094 — ( — .0453) = .1547. The
corresponding cost of living index difference is .0971 at the Italian utility
level and .1061 at the French level.
The figures discussed in the previous paragraph are all prior means. The
corresponding standard deviations, based on the specification (9.6)-(9.7), can
be derived straightforwardly, given that na(Uc) as defined in (9.5) is linear in
the /Ts. It appears that these standard deviations (given in parentheses) are
generally small. This is due to the fact that the adjustments of the observed
value shares wia by means of the uncertain marginal shares are of moderate
size. We do observe a tendency of the standard deviations to increase on
either end of the utility scale, which is in accordance with the larger ad¬
justments made.1 The standard deviations of the marginal index are much
larger, because their weights are completely determined by the uncertain //’s.
Given the better theoretical justification of the one-dimensional scale after
the adjustment of the value shares to the various utility levels, one should
1 One might perhaps expect that the standard deviations on the diagonal should be zero,
because these correspond to the case c = a in (9.3) and hence do not require any adjustment
by means of the uncertain /ds. However, each column of Table 7.10 is a one-dimensional
scale normalized such that the figures add up to zero. This implies that the adjustments
for the other elements of the column have their effect on the diagonal element.
expect that it fits the binary comparisons better than in the case of the raw
price scale of Table 5.4. The last line of Table 7.10 shows that this is true for
all utility levels except that of Italy. The table shows also that the mean
square residuals from the one-dimensional scale decline when we move from
left to right. This is due to the fact that, when the utility level goes up, the
cost of living index moves in the direction of the marginal index, which
satisfies the one-dimensional scale exactly because its weights depend only on
the commodity index i, see (9.9). Hence the better fit for higher utility levels
is more or less self-evident. It is therefore encouraging to observe that the fit
for the German level is also better than that of the raw scale in spite of the
fact that at that level the adjusted share w°(Uc, pfl), c = Germany, for four of
the six regions is computed by giving a negative weight to the corresponding
marginal value share.
7.A. Breese and Chead: The Principal Components of Consumer Preferences
In this section we shall show that demand equations of the type
w*tDqit = HiDqt + X vij ( DPj, X PkDPkt) + vit

k= 1
can always be written in preference independence form by means of a suitable

linear transformation. The condition is that the matrix N= [v;j] is symmetric
and negative definite. To prove this statement we start by writing the n
demand equations in vector form:
(A.l) d, = (Dqt) p + N (I — ip) nt + vt
= -^-Ni + N(I- / -u'NL, + v(

iNi V iNi /
where
w*uDqu Dpu
(A.2) n, = •
dr = •
_w*Dqnt _ _Dpnt_
Since N is symmetric and negative definite, it can be written as the product

of an orthogonal matrix, a diagonal matrix and the transpose (or inverse) of
the orthogonal matrix:
(A.3) N = R'ER R' = R_1
The diagonal elements of E are the latent roots of N, which are all negative
because of the negative definiteness of N. On substituting N = R'ER into
(A.l) we then obtain
d' - .'R>ERlR ERl + R ER(I - rR ERll, R ER )’t- + V'
276
7.A
BREESE AND CHEAD 277
and after premultiplication by R:
Dqt
<A 4) Rdl = .RER,™' + ER 1 - l'K7ElnU R ER )"■ + Rv'
Dqt
l7R7ER|ERi + E 1 - l R ERlR" R E K + R>-

The form of (A.4) suggests that it can be considered as a set of demand
equations of the preference independence type, because the price term has
the diagonal matrix E before the parentheses in the same place as that of N
in (A.l). However, this is a matter which should be verified step by step. We
notice, first of all, that each of the n equations of (A.4) deals, in principle,
with all n commodities simultaneously. This follows from the fact that R is
not necessarily diagonal. The left-hand variables are of the form
X rij(w*tDqjt) ••• X rnj(w*tDqjt)

J=1 J=i
where R= [r(J], The second line of (A.4) shows that the price log-changes are
weighted conformably, because the elements of Rrcr are
n n
X rlJDPjt ••• X rnjDpjt

j= 1 j=1
Hence (A.4) is an equation system in transformed prices and quantities or

equivalently a system of demand equations for n transformed commodities.
The left-hand variable of the ith equation is a weighted sum of the quantity
components of all value share changes, the weights being rn, ..., rin. The
right-hand variables of that equation are, apart from Dqt, the weighted price
log-change Ir^Dp^ with precisely the same weights rn, ...,rin plus n — 1
other weighted price log-changes ZrkjDpjt, kî- The former price log-change
plays a special role in that equation because of the diagonal structure of E
in (A.4) before the parentheses; the latter log-changes act as deflators as we
shall see below. We conclude that (A.4) is derived from (A.l) by combining
the n separate commodities to n linear transforms of these commodities.
Instead of one demand equation for bread and one for cheese we obtain
equations for “breese” and for “chead.”
Another point of interest is that the sum operator in (A.l) is replaced by a
weighted-sum operator in (A.4). That is, wherever we had i in (A.l), we
have Ri in (A.4). Take, for example, the income flexibility 0 = i'Ni. Since the
role of N is taken over by E, we now have
(Ri)'E(Ri) = i'R'ERi = i'Ni = </>
so that (A.4) can also be written as
(A. 5) Rd( = ERi + E ^1 — * Ru'R'E^j Rn, + Rvr

<P
Dq t
ERi + E - 0 Q ERij Q ERi Rnt + Rvr
0
One conclusion is that the income flexibility is invariant under the transfor¬
mation implied by the premultiplication of the demand equations by R. The
marginal value shares are not invariant. The first term on the right of (A.5)
shows that the new vector is
(A.6) pR = ^ ERi = R Q R'ERi j = R Q Ni j - Rp
Hence the new pR is obtained from the old p by premultiplying by R. This is

not surprising, given that the new equations are obtained from the same
transformation. The sum of the new marginal shares is not equal to 1, but the
weighted sum is:
(A.7) (Ri) Mr = I'R'Rfi = i'n = 1
Furthermore, the equivalent of Ni = $n is
(A.8) ERi =
which follows directly from (A.6).

On combining (A.5) and (A.6) we find that the transformed demand
equations can be written as
(A-9) Rdt = (Dq,)nR + (E - (j)\iR\i'R) Rrt, + Rv,
which is fully analogous to the original form
d; = (£<7r)h + (N - 4>n\i')nt + v(
Consider then the price term in (A.9):
ERr( - (0hr)OrRO = ERtc, - (ERi)(p'R'Rjq)

= ER [nt - (n'jq)i]
7.A BREESE AND CHEAD 279
Hence the price term of the ith equation is
(A-10) e< ] HkDpkt = rijDP'jt

k= 1 j= 1
where st is the ith diagonal element of E. We conclude that there is indeed

only one relative price term in each equation, that this term deals with the
deflated price of the same transformed commodity as that of the left-hand
variable, and that the deflation itself is invariant under the transformation.
The last statement is based on the fact that we have precisely the same Dp'jt
in (A. 10) as in the original demand equations before the transformation.
Finally, we should consider the transformed disturbance vector Rvf in
(A.9). Its expectation is obviously zero under assumption (1.13). Its covari¬
ance matrix under assumption (1.17) is
(A. 11)
This is, apart from the negative scalar multiplier <t2/</>, precisely the same as
the coefficient matrix of the price term in (A.9), and the matrix itself has the
form of the difference between a diagonal matrix and a matrix of unit rank.
Therefore, here too we have the analogue of preference independence. Note
that the sum of the transformed disturbances is not zero, but that the weighted
sum is:
(A.12) (Ri) Rvt = TR'Rv, = Tvt = 0
Essentially, the procedure described here amounts to diagonalizing the

matrix N of price coefficients. Since N is equal to the inverse of the Hessian
matrix U apart from pre- and postmultiplication by a diagonal matrix, this
diagonalization amounts, in turn, to a transformation of the n quantities
such that the utility function after transformation can be written in additive
form. Specifically, the marginal utility of each transformed commodity is then
independent of the “quantities” of all other transformed commodities. We
may also say that the objective of the transformation is to describe utility as a
function of n unrelated basic wants rather than of n commodities. This is

particularly clear when a comparison is made with the principal components
technique in statistics. The principal components of a set of variables form
another set of variables which are uncorrelated. Hence their covariance
matrix is diagonal, and the technique thus amounts to diagonalizing the
covariance matrix of the original variables.
7.B. Sampling Variances and Covariances of Coefficient Estimates
The variances and covariances of the coefficients of the demand equations

will be derived here in the following four steps:
(1) Our starting point is the covariance matrix V of the four v’s as given
in (4.5). Since 0 is equal to the sum of these v’s, its variance is equal to the
sum of all elements of V, see (4.7). The covariance of 0 and $u is
cov(v;,.,0) = £ cov (vi;, $jj)

j=i
so that the vector of these covariances is
V11 V22 V33 V44
(B.l) 10-4 [13 19 32 31]
(2) The variances and covariances involving p’s are derived by means of a
large-sample approximation. We have gt = v,,/0 and hence
-> ^ dg-i dvî d(j)

va 0
We interpret the differentials as sampling errors around the true values,
square both sides of the equation and take the expectation. This gives
(R ^ var A _ var Ai , var 0 ^ cov (t>„, 0)

' * ' 2 2 ' » 1 ^-
vii 0- vi;(/)
apart from terms of higher order of smallness. Next we substitute the

appropriate diagonal element of V for the variance of t>.., the value (4.7) for
the variance of 0, the appropriate element of the row (B.l) for the covariance
oft),,- and 0, and the estimates fih v>„, 0 for the true gu vu, 0 values which
appear in the denominators. This gives the large sample approximation of
the variance of
7.B SAMPLING VARIANCES AND COVARIANCES 281
The covariance of & and vu is also needed in (4.9) and (4.10). To find its
value we multiply both sides of (B.2) by dvu and then take the expectation,
which gives
co v (/*„$„) var vi; co \{yih$)
(B.4) -=-
Hi v„ <f>
(apart from higher order terms). The further derivation is analogous to that
of the variance of
(3) The variances and covariances described above suffice for the compu¬
tation of the expressions (4.9) and (4.10). We need more of them for the
standard errors of Section 7.5, so we shall present a complete survey. The
covariance of Ht and fij is found by multiplying dfii/Hi as given in (B.2) by a
similar expression for dHjhip which leads to
cov(/h, Hj) = covfij.y var $

MiM; Wjj 4>A
cov(v;i,0) COV (Vjj,4))
V«0 vn<t>
After applying the familiar substitutions we arrive at the following covariance

matrix of the fCs:
Ml ^2 Ha
16.25 - 7.50 .00 - 8.75~ hi
- 7.50 38.44 - 18.75 — 12.19
(B.5) 10-4 x
.00 - 18.75 35.00 - 16.25 Ms
- 8.75 - 12.19 - 16.25 37.19_ M4
To find the covariances of the h s and v’s we multiply both sides of (B.2)
by dvjj, which gives
cov (/),., Vjj) cov (9ti, tjj) cov (*jj> $)
Mi Vii <f>
with the following numerical result:
Mi M2 M3 Ha
r- 3.50 - 1.75 3.00 2.25" Vu

4.50 - 17.75 9.00 4.25 V22
(B.6) 10 4 x
6.00 - 2.00 - 8.00 4.00 V33
8.00 -2.25 11.00 - 16.75_ V44

Finally, we have the covariances of the p’s and 0. We multiply both sides
of (B.2) by d(p and conclude:
COV (Pi, 0) COV (v(i, 0) var 0
Pi V/i 0
which leads to
Pi P2 P3 Pa-
(B.7) 10-4 [15.00 - 23.75
- 15.00 - 6.25]
(4) The standard error of Dp\ is now found directly by means of (5.2) and
(B.5). The same applies to the standard error of Dq't-Dqt, see (7.7). To find
the variance of DX't we need the variance of the reciprocal of 0, see (5.1).
We have
dtp
T2
and hence
var 0
(B.8)
for which we substitute 95 x 10 4/(-.4)4 = .371. [The conclusion is that the

estimate of the income elasticity of the marginal utility of income is -2\ and
that its standard error is .61.]
For the variance of DX, we consider the sampling error
^JDq< ~ (fit ~ DP«
0 — 4> 4
-2— D(h ~ Z (P,- - Pi) DPi
i= 1
After squaring and taking the expectation we find

4 4
(Dq,)2
(B.9) var DX, — var 0+ ) ) DpitDpjt cov (&, p,.)
0
i=lj=l
4
2Dqt r
+ ) DPit COV (p;, 0)
7.C. Summary of the Main Symbols Used in Chapters 5, 6 and 7
This section contains a brief summary of the most important symbols

introduced in Chapters 5 through 7. As a general rule only those symbols are
7.C SUMMARY OF SYMBOLS 283
mentioned which are used in more than one section. The order is alphabetical
(first Latin, then Greek), but those variables which are defined only in terms
of log-changes are listed by their main symbol, not under D.
Ct is the covariance of the price and quantity log-changes, based on the

value shares w* as weights, corrected for the income effect of price changes:
Ct = rt- (Dp't - Dp,)Dq,

n
= Z w* (Dpt, - Dp,){Dqit - Dq,) - (Dp't - Dp,)Dq,

i= 1
It is also the covariance of the price log-changes and the equivalent income
changes, based on the marginal value shares as weights:
c,= -Dq)
*=1 \ J
dr is the column vector of the left-hand variables of the n demand equations
during the transition from period t — 1 to period t:
= O uDqlt ... w*Dqnt]
These demand equations have the following form:

n
w*,Dqit = qtDqt + £ vyD^ + vit

j=i
[In Chapter 6 the disturbance vit is deleted and Dqt is replaced by Dm,.]
D is the log-change operator (all logs in Chapters 5-7 are natural loga¬
rithms) :
xt
Dxt = A (logxf) = log --
xt- i
where xt_x and xt are successive values taken by some variable x.

m is income or total expenditure; mt is its value in period t, ma and mb are
its values in the atb and the bth region.
Dm, is the log-change in real income:
Dm, = Dm, — Dp,
m = mI(U, p) is the income solution of the indirect utility function, which

measures the minimum income necessary to attain the utility level U in the
price situation p.
pi is the price of the ith commodity; pn is its value in period t, p,a and p^
are its values in the ath and the bih region.
Dpt is the log-change in the cost of living price index:
n
Dp, = I w*Dpit
i= 1
Dp\ is the log-change in the marginal price index:

n
Dp', = Z biDPi,
i= 1
Dpit is the log-change in the i,h price deflated by the cost of living price
index:
Dpi, = Dpit - Dpt
Dpit is the log-change in the ilh price deflated by the marginal price index:
Dpi, = Dpit - Dp't
/>c(p|p0, U) is the true cost of living index of the price vector p with the
vector p0 as base evaluated at the utility level U:
p)
Pc(p|p0,[/) =
mÛ, p0)
PM(p|p0, t/) is the true marginal price index of the price vector p with the
vector p0 as base evaluated at the utility level U:
— m7(C/,p)
, cl)
PM( PlPo,^)= —
djj mi(U, p0)
pf is the intermediate price vector, its elements being equal to the geometric
means of the corresponding prices in periods t — 1 and t:
Pit = yjPi, t — 1 Pit
P is the diagonal n xn matrix of prices px, pn.

DPgt is the log-change in the partial (cost of living) price index of a set Sg
of commodities:
ieS,9
DPgt is the log-change in the partial marginal price index of a set Sg of

commodities: ^
DP’
M„ DPit Mg= X Pi
ieS„
ieS„
DPg, is the log-change in the partial marginal price index of a set Sg de¬
flated by the overall marginal price index:
DP'gt = DPgt - Dp't
qt is the quantity bought of the ith commodity; qit is its value in period t,
qia and qib are its values in the ath and the bth region.
Dqt is the log-change in the volume index:
n
Dqt = X w*Dqit
i= 1
[Dqt is also the average, marginally weighted, of the equivalent income

changes w^DqJpi.]
Dq't is the log-change in the volume index based on the marginal value
shares as weights: n
Dq't = X PiD(ht
Dq’ — Dq, is the log-change in the luxury-necessity index:
Dq[ — Dqt = X 0; - K)D<lit

i= 1
Q(U\U0, p) is the true real income index of the utility level U with the
level U0 as base evaluated at the price vector p:
m,(U, p)
Q(u\u0,v)
mj(U0,p)
q^ — qi{rn, p) is the ilh demand equation. It expresses the optimal quantity

in income and prices. In vector form: q°^q(m, p).
qf = q°(U, p) is the function which expresses the ith optimal quantity in the
utility level and prices. It answers the following question: If prices are p and
if income is such that U is the maximum utility level that can be attained in
this price situation, what is the optimal quantity of the ith commodity?
qm is the col umn vector of derivatives of the n demand functions with respect
to income.
Qp is the n x n matrix of derivatives of the n demand functions with respect
to the n prices, its (i,j)th element being dqjdpj.
DQgt is the log-change of the partial volume index of a set Sg of com¬

modities:
ieSg ieSg
u = u(q) is the consumer’s utility function.

«= p) is the indirect utility function, which describes the maximum
utility level that can be attained when income is m and prices are p.
U = [d2u/dqidqj] is the Hessian matrix of the utility function.
U* is the intermediate utility level between periods t— 1 and t, defined
implicitly by
mi(Ut*,yt-i) =_mr_
mt-i pt)
where (mt_l5 p(_x) is the income-price situation of period t — 1 and (m(, pt)
that of t.
vit is the disturbance of the ith demand equation during the transition from
period / — 1 to period t. [For this equation see under dr]
vt is the column vector of the disturbances in the n demand equations:
< = Olf ••• vnt]
Vt is the weighted variance of the equivalent income changes:

n
i— 1
is the value share of the /'(h commodity:
Ptfi
= -
m
Its value is wit in period t, wia and wib in the ath region and in the 6th, re¬
spectively.
w* is the average of the ith value share in the periods t- 1 and t:
* wi,t-i + wit
wit --
2
\viab is the average of the ith value share in the a'h and the 6th region:
vf°(£7, p) is the value share of the ith commodity when prices are p and
when income is such that U is the maximum utility level that can be attained
in this price situation.
w*Dqit is the quantity component of the change in the ith value share:
Awit « w*Dpit + w*Dqit - w*Dmt
and it is also the left-hand variable of the ith demand equation. [For this
equation see under d,.]
wffDqiJpi is the equivalent income change of the ith commodity during the
transition from period t — 1 to period t.
Wgt is the value share in period t of a set Sg of commodities:
Wgt= ieSg
Y, *tt
W*t is the average of the value shares in periods t— 1 and t of a set Sg of
commodities:
W* = Wg’t~l + Wgt - y wit*

fgt
ieSa
rt is the covariance of the price and quantity log-changes, based on the

value shares w* as weights:
r, = X w* (Dpu - Dpt)(Dqit - Dq,)

t=i
[For rot and rg„ g = l, ..., G, see page 163.]

r; is the covariance of the price and quantity log-changes, based on the
marginal value shares as weights:
r, = t H,(DPil - Dp',)(Dq„ - Dq',)

i= 1
St and dab are the allocation discrepancies of, respectively, the value shares
in two successive periods t-1 and t and those of two different regions
(the ath and the 6th):
s \ + Wit , WH \ u,*n...
d, = ) —1— -log - - = ; witDwit
Wi,t-1
i= 1 i=1
Win + W:ib W;
&ab — log
2 w ib
i =1
Kab is the logarithmic difference of the volume levels of two regions

(the ath and the bth):
n
Wjg + Wjb
âb
2 Qib
K, is the variance of the quantity log-changes, based on the value shares w*

as weights:
Kt= t K(Dqit-Dqt)2
i= 1
[For K0t and Kgt, g = 1, G, see page 163.]

is the variance of the quantity log-changes, based on the marginal value
shares as weights:
K; = t Vi(D<ht ~ Dq't)2
i= 1
X is the marginal utility of income:

duj
dm
X° = X(m, p) is the optimal value of the marginal utility of income for given
income m and prices p.
Xm is the derivative of the marginal utility of income with respect to income.
Xp is the column vector of derivatives of the marginal utility of income with
respect to the prices pu ..., pn.
DXt is the log-change in the marginal utility of income from period t— 1 to
period t.
Dl't is the log-change in the marginal utility of income in real terms from
period t— 1 to period t:
Dl't = DXt + Dp't
Pi is the marginal value share of the ith commodity:
p is the column vector of marginal value shares pu ..., pn.

pjw* is the income elasticity of the z'th commodity during the transition
from period t— 1 to period t.
vu is the coefficient of the jih deflated price Dp'jt in the demand equation for
the ith commodity. [For this equation see under d,.]
N is the n x n matrix whose (i, j)th element is v(- •.

nab is the logarithmic difference of the cost of living price indices of two
regions (the ath and the bth):
n
V
Xab= ) ---log —
+ Wlb , Pia
L 2 Pib
i=1
is the logarithmic difference of the marginal price indices of two regions

(the ath and the bth):
, v
Kab = L Pi 1°S —
i Pi°
i= 1 Pib
nt is the variance of the price log-changes, based on the value shares w*

as weights:
n,= t
i— 1
<{Dpit-Dpt)2
[For not and IJgt, g = 1, G, see page 163.]

n't is the variance of the price log-changes, based on the marginal value
shares as weights:
n\ = t Pi (pPu - DP'tf
i= 1
a2 is the basic variance measure of the covariance matrix of the disturbances

of the n demand equations:
1
&(vfvO = <j N - pp' J
(f) is the income flexibility, which is the reciprocal of the income elasticity
of the marginal utility of income:
1 _ d (log X)
cf) d (log m)
CHAPTER 8
INDUSTRIAL CONCENTRATION
AND THE ALLOCATION PROBLEM OF THE FIRM
8.1. The Measurement of Industrial Concentration
The objective of the present chapter is the extension of some of the concepts
which were introduced for the consumer in Chapters 4 through 7 to the
theory of the firm. The subject of this section and the next two sections is the
measurement of industrial concentration.1 This is a topic which is closely
related to the measurement of income inequality: The more income is
concentrated in the hands of a few households, the more inequality there is;
and the larger the shares of the largest firms in their branch of industry, the
more concentration there is in this industry. It stands to reason that similar
informational concepts will be useful for both areas. It will turn out that this
is indeed the case, but that there are nevertheless interesting differences, here
as well as in other extensions of the theory of the consumer household to that
of the firm.
We write A for the number of firms in a certain industry. It is assumed that
there is prior agreement as to the way in which their shares in the total are
measured; for example, market shares (in volume or value terms), shares in
total output (physical or dollar output), or numbers of employees per firm as
fractions of the total number of employees in the industry. We shall indicate
these shares by yh i= 1, ..., N, and require that they be nonnegative and add
up to 1:
N
(!•!) X A=1 yÔ i= 1 ,...,N

i= 1
The entropy of the distribution whose probabilities are these /s is
(L2) H(y) = X Li l°g -

i=i T;
1 The exposition is based on H. Theil, M. Scholes and P. Uribe, “An Informational

Approach to the Measurement of Industrial Concentration,” Report 6512 of the Center
for Mathematical Studies in Business and Economics, The University of Chicago (1965).
290
8.1 MEASURING INDUSTRIAL CONCENTRATION 291
which can be regarded as an inverse measure of concentration. If one share is

1 and all others are 0 (yt = 1 for some i, yj — 0 for each j# /) we have H (y) = 0
(the minimum entropy value) and the maximum degree of concentration. If
all shares are equal and hence equal to l/N, we have 7/(;;)=: log N, which is
the maximum entropy value given N and also the minimum degree of
concentration given N. When the number of firms increases while all remain
of the same size, the entropy log ./V increases too. This is in accordance with
the decreasing degree of concentration: When there are 200 firms of equal
size which jointly account for total output, there is much less concentration
than when the industry consists of only two firms of equal size, and certainly
less concentration than in the case of an industry which consists of only one
firm. In the last case we have H (y) = 0 and a maximum of concentration.
The arguments presented in the previous paragraph suggest that the
entropy is an appropriate inverse concentration measure: The higher the
entropy, the lower the degree of concentration and vice versa. We should be
careful, however, for we were close to a similar conclusion in Section 4.1
when considering income inequality; and we had to reject that conclusion
because of its undesirable aggregation implications. So let us consider the
aggregation problem here by combining firms to G sets of firms. The share of
set Sg is then
(1.3) Yg=lyt 9 = U-,G

ieSg
To make things specific we consider 21 makes of automobiles in the United

States which are combined to the following 4 sets (the figures in parentheses
are average percentage shares of the total number of cars produced in the
period 1936—1964):1
S1: All General Motors cars (47.2)

1. Chevrolet (25.14)
2. Pontiac (6.76)
3. Oldsmobile (6.22)
4. Buick (7.11)
5. Cadillac (1.98)
S2'. All Chrysler cars (18.0)

6. Plymouth (9.18)
7. Dodge (5.11)
1 The production figures are taken from the latest issues of Ward's Automotive Yearbook
(Detroit, Michigan; annual publication).
292 INDUSTRIAL CONCENTRATION
8. De Soto (1.37)
9. Chrysler and Imperial (2.31)
S3 \ All Ford cars (25.7)

10. Ford and Edsel (21.13)
11. Mercury (3.94)
12. Lincoln and Continental (.60)
S4\ All Independents (9.1)

13. Crosley (.06)
14. Hudson (1.15)
15. Kaiser (.59)
16. Nash (1.29)
17. Packard (.93)
18. Rambler (2.28)
19. Studebaker (2.42)
20. Willys (.38)
21. All others (.05)
The disaggregation of the entropy H(y) is of the familiar type:
(1.4) H(y) = H0(y)+ £ YgHg(y)

9= 1
where
(1-5) H0(y)= ^ Yg\o

9= 1 Yg
is the between-set entropy and
(1.6) Hg(y)= VLi * g

log
y '<■
g = l,...,G
i e Sg
is the entropy within S . We now approach the crucial problem. The between-
set entropy was rejected in Section 4.1 as a measure of between-set income
equality because it reaches its maximum value when the income shares Yg of
all sets are equal, independent of the number of individuals in the sets; in¬
stead, we preferred the excess of the maximum entropy value over the actual
value as a measure of income inequality, because this measure compares the
set incomes on a per capita basis. The question is therefore: Do we have to
make a similar shift from H (y) to log N—H(y) when measuring industrial
concentration? Answer: No. When making a between-set comparison be¬
tween General Motors, Chrysler, Ford, and Independents, we are indeed
8.1 MEASURING INDUSTRIAL CONCENTRATION 293
interested in their four set shares Yl5 Y2, Y3, YA. We are not interested in
per capita set shares, which in this case amounts to shares per make:
î/5, Y2/4, Y3/3, YJ9, the divisors being the numbers of makes in the
various sets. If General Motors decided to replace its Cadillac make by two
makes, Cadillac General and Cadillac Admiral, it would have six makes
rather than five, but we would still be interested in the share Yx of General
Motors; we would certainly not consider this alteration of such importance
that we would shift our attention from Yt/5 to Yl/6. This situation differs
radically from that of the income comparisons. If Yg stands for the income
share of some set Sg of individuals and Ng for the number of individuals
in Sg, the crucial concept in the between-set income comparison is per capita
income, which amounts to Yg/Ng (apart from a multiplicative factor inde¬
pendent of g), not to Yg. This leads - as we saw in Section 4.1 - to an in¬
formational measure of inequality rather than equality.1
These considerations induce us to accept (1.4) as a decomposition of the
“inverse concentration.” The numerical results for the 21 makes of auto¬
mobiles and their 4 sets in the period 1936-1964 (excluding two war years)
are presented in Table 8.1. The first column shows that the entropy over the
21 makes varied between 3 and 3.8 bits. [We shall use logarithms to the base
2 in this chapter.] The maximum entropy value is log 21=4.39 bits; hence,
given the number of makes, the observed values are of the order of 70 to 85
per cent of the maximum.2 The figures of the late forties and early fifties are
larger than those of the prewar years, thus indicating that there was less
concentration. Soon thereafter, however, there was an increase in concen¬
tration (partly due to the relative decline of Chrysler and Independents) such
that the entropy was approximately constant around a level of a little over 3
bits from 1958 onward. The second column contains the between-set entropy
H0(y) and the third the total within-set entropy IYgHg{y). The former exceeds
the latter throughout the period, but the difference is rather small for the
years around 1955. The between-set entropy varies between 1.6 and 1.9 bits; the
maximum is log 4 = 2 bits, hence the observed values are about 80 to 95 per
cent of this maximum. This percentage range is higher than the corresponding
range for the individual makes, so that we may say that in a relative sense
1 For a discussion of the Herfindahl concentration index (as a measure of inequality) see
the Appendix of this chapter.
2 Note, however, that there was no single year in which all 21 makes were actually pro¬
duced. The number of makes produced in each year varied from 13 to 19. If we compare
the entropy values with the attainable maximum in each year, the percentages are obviously
larger and fluctuate between about 75 and 90.
TABLE 8.1
THE ENTROPY OF PASSENGER CAR PRODUCTION IN THE UNITED STATES,
BY MAKES AND SETS OF MAKES, 1936-1964
Within separate sets

Total Between Within
Year General Chrysler Ford Inde¬
entropy sets sets
Motors pendents
1936 3.21 1.85 1.36 1.62 1.53 .18 2.40

37 3.38 1.88 1.50 1.82 1.69 .21 2.46
38 3.31 1.83 1.48 1.78 1.48 .49 2.36
39 3.48 1.86 1.62 1.87 1.59 .75 2.33
1940 3.43 1.80 1.63 1.84 1.66 .73 2.26
41 3.44 1.78 1.66 1.89 1.76 .59 2.25
42 3.43 1.87 1.56 1.92 1.65 .46 2.12
1945 2.97 1.74 1.22 1.90 1.80 .49 1.74
46 3.64 1.92 1.72 1.95 1.80 .81 2.27
47 3.70 1.91 1.79 1.95 1.78 .87 2.58
48 3.78 1.92 1.86 1.94 1.79 1.04 2.58
49 3.60 1.87 1.73 1.91 1.71 .89 2.44
1950 3.55 1.83 1.72 1.92 1.76 .90 2.42
51 3.59 1.87 1.72 1.94 1.72 .88 2.41
52 3.62 1.88 1.73 1.97 1.73 .90 2.46
53 3.42 1.79 1.63 1.87 1.70 .91 2.30
54 3.14 1.59 1.55 1.94 1.67 .75 2.17
55 3.27 1.65 1.62 1.99 1.67 .84 2.28
56 3.12 1.58 1.54 1.88 1.72 .79 1.80
57 3.12 1.66 1.46 1.88 1.67 .74 1.29
58 3.01 1.66 1.35 1.80 1.48 .63 .78
59 3.06 1.76 1.30 1.83 1.58 .56 .85
1960 3.10 1.75 1.35 1.77 1.48 .79 .68
61 3.04 1.68 1.36 1.77 1.48 .82 .67
62 3.00 1.62 1.38 1.78 1.50 .79 .64
63 3.04 1.65 1.38 1.80 1.41 .73 .54
64 3.03 1.65 1.38 1.86 1.43 .73 .01
there is more concentration among the 21 makes than among the 4 sets.1
The last four columns of the table contain the individual within-set
entropies Hg(y). They show a striking difference between the members of the
Big Three: In each year the within-Ford entropy is much less than the within-
1 It should be noted, however, that there is hardly any difference between the two per¬
centages in the years after 1960 if we consider that the number of makes is only 13 (see
the previous footnote). Both for the 13 makes and for the 4 sets the entropy is then slightly
above 80 per cent of the corresponding maximum.
8.2 CONCENTRATION IN DIFFERENT GEOGRAPHIC AREAS 295
Chrysler entropy, which in turn is less than the within-General Motors

entropy. This, of course, reflects the differences in the degree to which the
main make dominates within its set. The behavior of the entropy within the
Independents is dramatic. Until 1955, with the exception only of 1945, it
exceeded all other within-set entropies, year after year. But it declined
rapidly after 1955 until a nearly-zero level, leaving Rambler almost alone in
the Independents set.1 This decline is one of the factors that caused the
decrease in the total entropy H (y) during the last decade.
8.2. Concentration in Different Geographic Areas: New Car Registrations in

the United States, 1959-1964
We shall now go somewhat deeper into the automobile concentration

problem by taking the regional distribution into account. We shall leave the
production side and go to the demand side by concentrating on the new
passenger car registrations. The geographic unit is the state: each of the
states of the United States excluding Alaska and Hawaii plus the District of
Columbia, hence 49 states as a whole. We shall also aggregate these states to
the eight regions (sets of adjacent states) which we considered in Section 4.3
when analyzing regional income inequality. The period to be considered is
1959-1964. During this period most of the Independents played a very minor
role or no role at all. We shall therefore retain the sets Su S2, S3 as they were
defined in the previous section but modify S4 such that it consists of four
subgroups: Rambler, Studebaker, Other domestic cars, and Foreign cars.
The last category was not included in the analysis of the previous section; we
want to include it here because we are now working on the demand side. To
stress the difference in content we shall describe SA as “Other cars” rather
than “Independents.”2
1 The following account summarizes the history briefly. Hudson and Nash merged into
the American Motors Co. in 1954; Rambler, previously a model of Hudson and Nash,
emerged as a make in 1955; Hudson and Nash both disappeared as makes in 1958.
Packard and Studebaker also merged in 1954. The last Packard cars were produced in
1958, the last Studebakers (in the United States) in 1964. Kaiser and Willys disappeared
from the market in 1956.
2 The registrations data are also taken from the annual issues of Ward's Automotive Year¬
book. In 1959 and 1960 the miscellaneous domestic cars and the foreign cars were combined
into one group. They were separated by means of Ward’s table “Imported car sales pene¬
tration by states and geographical areas,” which gives the percentage (for each state) of
the total number of new car registrations accounted for by foreign cars. Whenever this led
to negative numbers for miscellaneous domestic cars (due to the rounding off of the per¬
centages to two decimal places), zero values were used instead.
As a whole, therefore, we have 49 states, 8 regions, one nation, 16 makes and

4 sets of makes. The notation is as follows. We write yik for the number of cars
of make i newly registered in state k, measured as a fraction of the national
total of all new car registrations (Alaska and Hawaii excluded). We write
49 16
(2.i) ki. = Ek= 1

yik y.k =E i= 1
yik
for the share of make i and of state k, respectively, of the national total, and
similarly
(2.2) = ieS„
Z y* y,. = Z ieSo
y,
for the number of cars of set Sg newly registered in state k and in the nation,
respectively, again in both cases expressed as a fraction of the national total
of all new car registrations.
We start at the national level, for which the decomposition (1.4) takes the
following form:
1
(2.3) Ya. +
ieSa
The left-hand side is the total national entropy, the two terms on the right are
the national between-set entropy and the total national within-set entropy.
These are given in Table 8.2 for each of the years 1959-1964 in lines 5, 10 and
15, respectively. The four lines above each of these represent geographic
decompositions which will now be discussed. For this purpose we need
certain regional shares:
(2.4) y'ir = E
keCr
yik y'.r = E y.k
ksCr
K= E Ygk
keCr
These are, respectively, the number of new car registrations of make i in

region C,, all new car registrations in Cr, and the registrations of Sg in Cr
(r=l, ..., 8), each of them expressed as a fraction of the national total.
Consider then the national between-set entropy, which can be decomposed
as follows: ,
(2-5) Y„ S log ti-

v
y .r y'
1 gr
TABLE 8.2
THE ENTROPY AND ITS GEOGRAPHIC DECOMPOSITION OF NEW CAR REGISTRATIONS
IN THE UNITED STATES, 1959-1964
1959 1960 1961 1962 1963 1964
Total entropy
(1) Average of 49 states 31857 32315 31760 31230 31514 31644

(2) State heterogeneity within
regions 94 85 80 87 97 87
(3) Average of 8 regions 31951 32400 31840 31317 31611 31731
(4) Regional heterogeneity 202 156 110 94 115 116
(5) Total national entropy 32153 32556 31950 31411 31726 31847
Between-set entropy

regions 38 36 31 36 44 41
(8) Average of 8 regions 18333 18377 17683 16872 17181 17441
(10) National between-set
entropy 18452 18469 17750 16925 17251 17512
Total within-set entropy

regions 56 49 49 51 53 46
(13) Average of 8 regions 13618 14024 14157 14445 144-30 14290
(15) Total national within-set
entropy 13701 14087 14200 14486 14476 14335
Note. All entropy values are expressed in 1(D4 bit.
The first term on the right is a weighted average of the regional between-set
entropies (with weights equal to the regional shares y'r). The second is the
average of the information expectations of indirect messages whose prior
probabilities are the national set shares Yx_, ..., Y4 and whose posterior
probabilities are the regional set shares Y[r/y[r, ..., Y4r/y'r. The latter term is
a measure for the heterogeneity of the regions with respect to the shares of
General Motors, Chrysler, Ford, and Other cars. The former can be written
as follows:
8 4
(2.6)
8 4
r= 1 keCr g= 1
The first term on the right, which can be simplified to

49 4
(2.7)
k=1 0=1
is the average between-set entropy of all 49 states. The second, which can be
simplified in the same way:
8 4
(2.8)
r = 1 ke Cr g= 1
is the average of the information expectations of a number of indirect

messages. Their prior probabilities are the regional set shares Y'r/y'r, the
posterior probabilities are the state set shares Ygk/y_k. Obviously, this term
measures the heterogeneity of the states within each region with respect to
the shares of the four sets of makes. The numerical results are shown in
lines 6 through 10 of Table 8.2. They indicate that the contributions of the
two heterogeneity terms to the national between-set entropy are small,
particularly the term which deals with the state heterogeneity within the
regions.
It is quite instructive to review the results obtained in reverse order (from
the top to the bottom in the table). Our starting point is now (2.7), the average
between-set entropy of all states. We add (2.8), which measures the between-
set heterogeneity of all states within their respective regions. The result is the
left-hand side of (2.6), the average between-set entropy of the eight regions.
Next we add the second right-hand term of (2.5), which measures the between-
set heterogeneity of the regions, and the result is the national between-set
entropy. Note that entropy never decreases by geographic aggregation, or

equivalently that concentration never increases by this type of aggregation.
In fact, this is quite natural. For suppose that half of the states buy only
General Motors and the other half only Fords. Then there is zero between-
set entropy (maximum concentration) in each state, but positive between-set
entropy (less than maximum concentration) in the country as a whole. When
the areas which are combined have the same distribution - over sets of makes
in this case - the entropy remains unchanged after aggregation; but this is
exceptional and normally we observe an increase. This is in contrast to the
aggregation of the items considered, in this case: aggregation of makes to
sets of makes. Then entropy never increases, H (y) being replaced by
H0 (y)^H (y), the equality sign of which corresponds to the special case in
which all within-set entropies vanish (which means that there is effectively
only one make in each set).
The procedure for the within-set entropy is completely parallel. The national
figure is the second right-hand term of (2.3), which is decomposed as follows:
4
+
yJYg.
The first term on the right is the average within-set entropy of the regions,
the second measures the within-set differences between the regions. The
former can be decomposed in turn:
8 4
(2.10)
K
y'ir
r= 1 9=1
4
5?
yik
yuJYgk
+
y'irlYgr
Here the first right-hand term is the average within-set entropy of all states,
and the second measures the within-set differences of the states within each
region.
The numerical results for the total within-set entropy are shown in
Table 8.2, lines 11 through 15. The outcomes indicate that the contributions
of the heterogeneity terms to the national within-set entropy are small, just
as in the between-set case; but that, unlike the latter case, it is not true that
the state heterogeneity within regions is dominated by the regional hetero¬
geneity. The geographic decomposition of the total national entropy is
given in the upper part of the table; its derivation is completely analogous,
and each figure in the upper part is equal to the sum of the two cor¬
responding figures in the middle and the lower part. The outcomes show
that the difference between the mean of the total entropies of the 49
states and the total national entropy is less than 1 per cent in each of the
six years.
8.3. Further Details on the New Car Registrations
The most important difference between the data considered in Section 8.1
and the corresponding national data of Section 8.2 is the composition of the
set 54: Independents in Section 8.1 and Other cars in Section 8.2. This aspect
is pursued in Table 8.3, which is entirely comparable with Table 8.1 except
for the difference with respect to SA (and for the fact that it deals with new
car registrations, not with cars produced). A comparison of the two tables
shows that the total entropy is about 5 per cent higher in the present case,
TABLE 8.3
THE ENTROPY OF NEW CAR REGISTRATIONS IN THE UNITED STATES,
BY MAKES AND SETS OF MAKES, 1959-1964
Within separate sets

Year Total Between Within
entropy sets sets General Chrysler Ford Other
Motors cars
1959 3.22 1.85 1.37 1.83 1.58 .57 1.41

1960 3.26 1.85 1.41 1.76 1.51 .76 1.44
1961 3.20 1.78 1.42 1.78 1.47 .82 1.40
1962 3.14 1.69 1.45 1.78 1.51 .79 1.41
1963 3.17 1.73 1.45 1.82 1.44 .74 1.36
1964 3.18 1.75 1.43 1.85 1.45 .73 1.20
8.3 FURTHER DETAILS 301
that about the same percentage difference applies to the between-set entropy
and the total within-set entropy, and that there are striking contrasts between
the separate within-set entropies. There are virtually no differences between
the corresponding output and registration within-set entropy values of the Big
Three, but there is a sizable difference as to S4. The within-,S4 entropy of the
output of Independents declined from .85 to .01 bit in the period 1959-1964;
the within-^ entropy of the registrations of Other cars fluctuated around the
1.4 bits level from 1959 to 1963, and dropped slightly (to 1.2 bits) in 1964.
Table 8.4 gives a clear picture of the effect of the inclusion of Foreign cars.
First, it raises the small share of S4 substantially, thus contributing to a larger
between-set entropy. Second, it reduces the dominant position of Ramblers
within S4, thus contributing to a larger within-S^ entropy. Third, the larger
share of S4 implies that the larger within-S^ entropy obtains a larger weight
in the total entropy. All these factors contribute to a lower degree of concen¬
tration when the foreign cars are included.
The Other cars group is interesting in itself, quite apart from the com¬
parison with Independents. We refer to Table 8.5, which contains the geo¬
graphic decomposition of the four separate within-set entropies. In algebraic
form the decomposition is as follows (g = 1, ..., 4):
49
Average state entropy:

k=l ieSg
8
State heterogeneity within regions:
+
8
Average regional entropy:
Regional heterogeneity:
+
National entropy within Sg:
The numerical results of Table 8.5 show that the state and regional heteroge-
TABLE 8.4
PERCENTAGE SHARES OF S4, RAMBLER, AND FOREIGN CARS
Percentage share of ^4 in Rambler’s share within £4 Foreign car

Year share within Si
production registrations production registrations (registrations)
1959 9.93 18.43 72.31 32.66 54.80

1960 9.83 15.73 73.75 40.88 47.87
1961 8.18 14.14 82.52 44.84 45.54
1962 7.81 12.20 83.99 50.00 39.84
1963 7.18 11.62 87.60 48.45 43.55
1964 5.10 11.00 99.80 42.73 53.73
Note. Si = Independents for production, Other cars for new car registrations.
neities account for \ per cent or less of the national within-set entropy in the
case of General Motors and Chrysler, for 1 per cent or less in the case of
Ford, but for 2 to 3+ per cent in the case of Other cars. Evidently, the last
group is characterized by relatively large differences in composition (mainly
Rambler versus Foreign cars) in the various states and regions.
8.4. The Demand for Factors of Production: The Fundamental Matrix

Equation of Production Theory1
Our next topic is the extension of the consumer’s allocation model to a

demand theory of production factors. It was clearly demonstrated by
Samuelson that there is indeed a close relationship between the two demand
theories.2 Our objective is therefore to formulate a demand model for pro¬
duction factors which is similar to the model of Chapter 6.
The starting point is a linear cost function
(4.1) C=p'q (possibly plus fixed costs)
where p= [//,] and q = [qt] are /7-element column vectors of prices and quanti¬
ties, respectively, of factor inputs. The cost function is minimized subject to
1 The exposition of this and the next two sections is largely based on H. Theil and
C. B. Tilanus, “The Demand for Production Factors and the Price Sensitivity of Input-
Output Predictions,” International Economic Review, Vol. 5 (1964), pp. 258-272.
2 See P. A. Samuelson, Foundations of Economic Analysis (Cambridge: Harvard University
Press, 1947), Chapter IV.
8.4 THE DEMAND FOR FACTORS OF PRODUCTION 303
TABLE 8.5
WITHIN-SET ENTROPIES OF NEW CAR REGISTRATIONS IN THE UNITED STATES, 1959-1964
1959 1960 1961 1962 1963 1964
General Motors
Average of 49 states 18238 17530 17751 17736 18078 18459

State heterogeneity within regions 37 33 37 39 43 41
Average of 8 regions 18275 17563 17788 17776 18120 18500
Regional heterogeneity 42 37 25 27 36 40
National entropy within General Motors 18317 17600 17813 17803 18157 18540
Chrysler

National entropy within Chrysler 15781 15111 14740 15057 14405 14470
Ford

National entropy within Ford 5683 7616 8193 7931 7373 7296
Other cars

National entropy within Other cars 14114 14368 14014 14061 13591 12022
Note. All entropy values are expressed in 10~4 bit.
the constraint of a production function,
(4.2) x = p(q)
where x is the output produced. The question is: What are the input quanti¬
ties q° which minimize (4.1) subject to (4.2), given fixed positive values of
output x and prices p? By varying x and p we then have the factor demand
equations
(4.3) q° = qO,p)
which differ from the consumer demand equations to the extent that income
m is replaced by output x as one of the arguments. There are more differences
but also several similarities as we shall shortly see.
We proceed along familiar lines by constructing the Lagrangian expression
p'q - p[u(q) - x]
p being the Lagrangian multiplier. On differentiating the expression with

respect to the vector q and equating the result to zero we obtain
dv 1
(4.4) --P = 0
dq p
Hence the marginal product dv/dqi of each factor is proportional to the

corresponding price and 1/p is the proportionality constant. The multiplier
itself is nothing else than marginal cost (and therefore comparable with A,
the marginal utility of income in consumer demand theory). This statement
is verified by substituting the demand equations (4.3) into the production
function (4.2):
(4.5) x — u(q(x,p))
and then differentiating both sides with respect to x:
(4.6) i-f fLp.iy oqt ox p l_j ox p ax

i= 1 i= 1
where use is made of (4.4) and (4.1). It will be noted that the procedure of
substituting the demand equations into the production function is analogous
to the construction of the indirect utility function (see Section 6.7). There is,
of course, a difference to the extent that in the present case we have the same
variable (x) on both sides of the equality sign, so that the economic interpre¬
tation of (4.5) is less straightforward.
We conclude from (4.4) that p (marginal cost) and its reciprocal are both
positive when all marginal products are positive. It will be assumed that this
is indeed the case. It is also assumed that the production function v( ) is
twice differentiable and has symmetric cross-derivatives, just as the utility
function of consumer demand theory.
8.4 THE DEMAND FOR FACTORS OF PRODUCTION 305
We can write the result (4.6) in the equivalent form
(4-7) p'q* = P
where qx is the /7-element column vector whose ith element is the output
derivative dqjdx of the z'th demand equation. By differentiating (4.5) with
respect to the jth price we obtain
n n
i= 1 i = 1
or
(4.8) P Qp = 0
where Qp is the nxn matrix of price derivatives of the demand equations, its
(i,j)th element being dqjdpj. Further, by differentiating (4.4) with respect
to x and pj we find
n
d2v 8qk _ dP 1 i = 1,..., n

Pi ^
Sqt8qk dx OX
—i
^
d2v
^ 11
_ °p if i /y
1
1
1
j
Cl)
dqidqk dpj
^3
k= 1
if i=j
or equivalently
(4.9) Vq* - VPX 1 = 0
(4.10) vqp-p (P;1y = P~1 I

where V = [d2vldqidqj) is the Hessian matrix of the production function, pj1
the derivative of the reciprocal of marginal cost with respect to output,
and Pp 1 the column vector of derivatives of this reciprocal with respect to
the n prices.
We combine equations (4.7) through (4.10) in the following partitioned
form:
V p Qp o P~li
(4.11)
p' 0. p o
This is the fundamental matrix equation of factor demand theory. It is very

similar to the corresponding equation (2.5) of Section 6.2 of consumption
theory. The first left-hand matrix is the bordered Hessian matrix, here of the
production function, there of the utility function. It is postmultiplied by an
(n +1) x (n +1) matrix whose first n rows are the derivatives of the n demand
functions. The only difference with the equation of consumption theory is that
the first column contains output instead of income derivatives. The (n+l)st
row consists of the derivatives of the reciprocal of marginal cost, whereas the
corresponding row of the consumption equation deals with marginal utility
itself, not with its reciprocal. The most interesting difference is the zero sub¬
vector in the lower right-hand part of the matrix on the right, which is — q0'
in the consumption case. This difference is due to the fact that in factor
demand theory there is no such thing as the income effect of price changes.
Income is replaced by output, which is expressed in volume terms, not in
money units.1
8.5. Solving the Fundamental Matrix Equation in the Case of Constant

Returns to Scale; The Substitution Flexibility
When solving the fundamental matrix equation (4.11) it seems obvious to

continue the analogy with consumption theory. There is, however, a difficulty
in assuming the Hessian matrix V to be nonsingular. In fact, constant returns
to scale, to which case we shall confine ourselves from now on, involves a
singular V. This is shown as follows. When there are constant returns to
scale, the production function is homogeneous of the first degree, so that
n
dv
(5.1)
oqi
according to Euler’s theorem. We differentiate (5.1) with respect to q j’
which gives
dv dv " d2v
t f l qif a j = 1
dq j dqj i=1 dqfqj
and hence
(5.2) Vq = 0
Therefore, given that the f s of the optimal point are assumed to be positive,
V must indeed be singular. In fact, it will be assumed in the sequel that V is
1 In order not to overburden the notation, we do not attach the superscript 0 to p, its
reciprocal, and q to indicate that the values of the optimal point are to be taken. This
aspect should be clear from the context.
8.5 SOLVING THE FUNDAMENTAL MATRIX EQUATION 307
negative semi-definite (its rank being n — 1 as we shall shortly see). This is very
close to the assumption of a negative definite U in consumption theory. It
implies decreasing marginal products, to be compared with the decreasing
marginal utilities.
Next we show that the vector qx of output derivatives and q itself are equal
in the optimal point, apart from a scalar multiplier. The proof is as follows.
We premultiply both sides of (4.9) by q', which leads to (q'p)p* 1 = 0 in view
of (5.2). Since q'p>0 we conclude
(5.3) p;x = o
which means that p_1 and its reciprocal, marginal cost, are independent of
output. [Hence total cost C is not only linear in factor inputs, see (4.1), but it
is also a linear function of output.] On combining (5.3) with (4.9) we obtain
(5.4) Vqx = 0
which in view of (5.2) means either of two things: Either qx is equal to q apart
from a multiplicative scalar, or the rank of V is less than n — 1. The latter
possibility must be ruled out, however, since it would imply that the bordered
Hessian matrix in (4.11) is singular, whereas the right-hand matrix of that
equation has full rank. To show that the bordered Hessian matrix is singular
ifq and qx are not equal to each other (apart from a multiplicative scalar) we
construct the vector pq —Cqv, which is #0 under the condition stated, and
consider
V p pq - Cq," 0"
o
_p' o_ 0
1
see (4.1) and (4.7).

We have thus established that q* must be a multiple of q. The multiple
must be p/C, because this is the only value which ensures that the left-hand
column vector in (5.5) is a zero vector, so that the bordered Hessian matrix is
indeed nonsingular. Also, we have
dv
(5.6) C = q'p = pq — = px
fiq
in view of (4.1), (4.4) and (5.1). Hence the multiple p/C is simply 1/x. We
conclude
1
(5.7) q* = -q
x
from which it follows that the output elasticity of the demand for each input
is unity (and identically so). This result enables us to simplify the factor
demand equations by writing qjx as a function ofp1? p„. Our task is now
to find an appropriate form of this function, similar to equation (4.12) of
Section 6.4.
We premultiply both sides of (4.10) by the transpose of q*, which gives
- (q.xPKPp-1)' = P~\x
in view of Vq,=o. see (5.4). Since q(.p = p, see (4.7), we have
1 1
(5-8) pp =~ 2q* = ~ 2 q
P P x
where use has been made of (5.7). We conclude that the vectors p~1, qx and q
are all proportional to each other. Furthermore, on combining (5.8) and
(4.10) we obtain
11,1 1
(5.9) vqp = i —2Pq*= ~i —r pq
P P P P x
which is very close to equation (2.11) of Section 6.2. There is one term missing,
which should not surprise us in view of the remarks made at the end of the
previous section. The most important difference, however, is that we cannot
solve (5.9) for Qp by premultiplication by the inverse of V.
To solve this problem we introduce the dimensionless quantity
d2v
K —-
OPi^Pj
(5.10) Xij
dv dv
dqt dqj
which is the reciprocal of the elasticity of substitution between the ith and the
7th production factor.1 Consider also the share of each factor in total marginal
cost:
/r hi PiPi . .
(5-11) yi = ~ i = l,...,n
px
These shares are positive by assumption, and they add up to 1 in view of

(4.7) and (5.7):
n
(5.12) Y,
/=!
= 1 k; > 0 i = l, •••,«
1 See e.g. R. G. D. Allen, Mathematical Analysis for Economists (London: MacMillan

and Co., 1947), p. 343.
8.5 SOLVING THE FUNDAMENTAL MATRIX EQUATION 309
Next we consider a weighted average of all Xij s, i¥=j, the weights being
proportional to the sum of the shares of the two factors involved:
n n
(s is) Z= I (y. + y,)x,t

i=lj=i+i
which will be called the (general) substitution flexibility. It measures the

degree to which, on the average, production factors can be substituted for
each other, and it will be assumed to be positive:
(5.14) X > 0
This sign corresponds to the case in which an increase in the quantity of one
factor raises the marginal product of another factor: d2v/dqidqj>0, /#/•
The substitution flexibility combined with (4.8) will be our vehicle in
solving (5.9) with respect to Qp. For this purpose we define the following
nxn matrix Z, the function of which is to transform V (by the addition of a
matrix of unit rank) such that the result is negative definite, just as the
Hessian matrix of consumption theory:
1 X dv dv 1
(5.15) Z = -y - -V -
X x2 dq dq' x
[see (4.4)]. We observe that Z is indeed negative definite, because for any
«-element vector a#0 we have
a'Va / a'p\2
a'Za <0
x \px /
in view of (5.14) and the negative semi-definiteness of V. We now multiply

(5.9) by l/x:
1 1 1
x
VQp=
px
I-
GW pq
and (4.8) by [x/(px)2]p:

X r. n
7 xlPPQp^0
(px)
and subtract with the following result:

Since Z is negative definite, we can solve directly for Qp:
(5.16) QP z_1 - z-V

px (px)
which in scalar form amounts to
i*
dpj px (px)2 [j Z mj
k= 1
where zlJ is the (i, j)th element of Z \ The result in elasticity form is more
elegant:
ik
dQ°g g,)
(5.17)
5 (log Pj)
ykyj
QiQk
This is the solution of the fundamental matrix equation (4.11) for the price
effect on the quantities bought. We shall use it in the next section for the
specification of factor demand equations.
It should be noted that the zlj which occur in (5.17) are not unrestricted.
We have
1_
Zq = — P
px
in view of Vq = 0 and p'q = px, see (5.2) and (5.6). We premultiply by Z"1 and
find that q is equal to a negative multiple — x/(px) of Z_1p, or equivalently
that cji is equal to this multiple of the sum over k of zikpk. Dividing both sides
by qt we then obtain
n n
ik ik
1 - PkPk yk
q iPk QiQk
k= 1 k= 1
or equivalently
y zik i
(5.18) ) —yk = ~-
Lu QiQk 1
from which we conclude that (5.17) can be replaced by the simpler form
<?(logr/i) / ziJ 1\
(5.19)
5 (log Pj) yj\qi9j +1)
Hence the elasticity of the demand for the ith input with respect to the
8.6 THE SPECIAL C.E.S. CASE 311
/h price is equal to they'th share of total marginal cost multiplied by the sum of
two terms, one of which is symmetric in i and j and the other is independent
of i and j. In addition, if we weight the former terms with the shares yj, we
obtain the latter apart from sign, see (5.18). It will be clear that this restriction
is rather close to equation (2.15) of Section 6.2. The symmetry of the first
right-hand term in (5.19) is due to
(5.20) z'J = zJI ij = 1,..., n
which follows directly from the Z definition (5.15); it is to be compared with

restriction (2.14) of Section 6.2. In fact, it will turn out that the role of Z and
its inverse is largely similar to that of U and U-1.
8.6. The Special C.E.S. Case; The Problem of the Specification of the Factor
Demand Equations
We recall that in the case of constant returns to scale the demand equations
can be formulated such that qjx is described in terms of prices. Using
equation (5.17) for the price elasticities, we thus have
Ik
Z \ Z
(6.1) d ^log ~1 -— yj- / — ykyj d (log Pj)
L-Qidj U didk
j= 1 k= 1
n n
yL did;
~ yj[d(log - Ly voog^)]
j= 1 k=1
The second line shows that the demand equations are in terms of relative
prices. The logarithmic change in the deflator is obtained by weighting the
individual logarithmic price changes according to the share yk of each factor
in total marginal cost. The elasticity of the demand for the ith factor with
respect to the deflated price of the jth factor is
z1J
(6-2) e;. =-jj
didj
and we conclude from (5.18) that, in each demand equation, the sum of all
deflated-price elasticities is equal to the negative reciprocal of the substitution
flexibility:
(6.3) £ BtJ = - - i=l,...,n

7=1 X
This result is closely related to equation (4.10) of Section 6.4, which deals
with the following parameters of a consumer demand equation: the deflated-

price coefficients vip the income flexibility </>, and the marginal value share
Hi. If we divide both sides of this constraint by the value share wh it is
formulated in elasticities:
n
j=i
The sum of the deflated-price elasticities v^/wt is thus equal to the product
of the income flexibility (j) and the income elasticity /r;/w;. Returning to the
corresponding constraint (6.3) of factor demand theory, we notice that <f>
should be compared with — l/x and the income elasticity with 1, this unit
value being the output elasticity of the demand for the production factor in
the case of constant returns to scale.
An interesting special case is the one in which all Xij s, i#j, are equal and
hence equal to x■ It follows from (5.10) and (5.15) that Z is then diagonal, so
that all s with unequal subscripts vanish, see (6.2). Each demand equation
has then only one deflated price; therefore, this case of equal elasticities of
substitution should be compared with preference independence in consump¬
tion theory. Actually, the present case is still more special due to the as¬
sumption of constant returns to scale, because we have en = —1/x for each i,
see (6.3). As an example we consider the C.E.S. (constant elasticity of substi¬
tution) production function1
(6.4) x = £ diqrj 0. > 0, £ 0t = 1
for which we have
O0i \qj
d2v _(C + ^djOj/x2 y+1
dqt dqj x XqiqJ * #j
and hence, in view of (5.10), xiJ = c +1 for all pairs (i,j), i^j. We conclude
that in the C.E.S. case the elasticity of the demand for each factor with
1 See H. Uzawa, “Production Functions with Constant Elasticities of Substitution,”

Review of Economic Studies, Vol. 29 (1962), pp. 291-299; and also the earlier articles on
the C.E.S. function particularly K. J. Arrow, H. B. Chenery, B. S. Minhas and R. M.
Solow, “Capital-Labor Substitution and Economic Efficiency,” The Review of Economics
and Statistics, Vol. 43 (1961), pp. 225-250.
respect to its own deflated price is — l/(c+l). The demand equation in

infinitesimal form is then
n
1
(6.5) d d (log p,) - X V(logpfc)
c + 1 ft= 1
which is to be compared with equation (4.12) of Section 6.4 for the special
case of preference independence (v^ = 0 whenever /#/) and unitary income
elasticities (pjw—l for each i). The similarity is even increased when we
multiply both sides by the factor’s share of total marginal cost:
(6.6) ytd log

<li If
—— y^ (log ps) - yt X ykd(\ogpk)
c + 1 fc= 1
Note further that, when both sides of this equation are summed over i
(from 1 through n), the right-hand side vanishes, so that we may conclude:
n
(6.7) X
i= 1
y-id 0°g di) = d (log x)
The left-hand side is the infinitesimal variant of the change in an input volume
index (in logarithmic form), which is thus equal to the logarithmic output
change. It will be noticed that the weights used in (6.7) are precisely the same
as those which occur in the price index of (6.1), (6.5) and (6.6). It is therefore
possible to decompose the change in total marginal cost in a way similar to
that of total expenditure of consumption theory, which leads to allocation
discrepancies for the firm. This is rather straightforward and will not be
pursued here.1
The preceding paragraph deals with a special case. In general we cannot
expect all Xifs to be equal, because some factors can be substituted for each
other more easily or less easily than is true for certain other pairs. In that case
the demand equation in constant-elasticity form runs as follows [see (6.1) and
(6.2)]:
(6.8) d(log-) = X
eU rf0°SPj)~ X JV*0°gfik)
\ xj j= i L fc=1 J
1 We only note that the logarithmic change of the price index of (6.5) is equal to the
logarithmic change of marginal cost. This result follows easily from (5.3) and (5.8); it
bears a clear resemblance to equation (6.3) of Section 6.6 which contains the marginal
utility of income.
which is the generalization of the special C.E.S. demand function (6.5). When
using this equation as a base for estimation purposes (after appropriate
amendments for the transition to finite changes) we face the difficulty that in
each equation n price coefficients have to be estimated. This raises the question
whether it is possible to obtain simplifications similar to those of block-
independence in the consumer’s case. This is in principle possible under the
following conditions. Let Xij — X hold for most of the pairs (i,j), i#j. These
are the pairs of factors which have “normal” elasticities of substitution. For
the other pairs we have Xij^Xi hut the difference Xij~X Is not too far from
zero in the sense that
x x2 Gq dq'
is a matrix whose nonzero off-diagonal elements ztj are small in absolute

value compared with the geometric mean of the corresponding diagonal
elements, ^/(zuZjj). It has been shown by A. P. Barten that in that case
Z_1 is such that z,J xO holds for all those pairs (i, j) for which ztj is exactly
zero.1 On combining this result with (6.2) and (6.8) we conclude that the ith
demand equation will then contain only the deflated prices of those com¬
modities with which the ith has above-normal or below-normal elasticities of
substitution, the e,/s of the other deflated prices being negligible.
When summing the e^’s over j we obtain —1/x, see (6.3). This is a linear
restriction on the elasticities; when it is specified that these are fixed param¬
eters, such a restriction can be handled easily in empirical estimation. The
symmetry condition (5.20) is more difficult to handle, since it implies
yisij=yjEji. The /s vary over time (dependent on the development of factor
prices), which is incompatible with fixed values of the e’s. We can obtain an
improvement in this respect by multiplying both sides of (6.8) by yt:
n n
(6.9) y,d I v*. d (log pj) - £ ykd (log pk)
where
(6.10)
1 Barten’s analysis is in terms of the Hessian matrix of the utility function, but the present
case is completely parallel. See A. P. Barten, “Consumer Demand Functions under
Conditions of Almost Additive Preferences,” Econometrica, Vol. 32 (1964), pp. 1-38.
The form (6.9) corresponds with (6.6) in the C.E.S. case and with equation
(4.12) of Section 6.4 in the consumption case. The symmetry condition (5.20)
now amounts to v*. = v ; clearly, variable y’s cause no difficulty when the
v*’s rather than the e’s are assumed to be fixed. However, we now have a
difficulty with respect to (6.3), which takes the following form in the v*’s:
(6.11)
The left-hand side is fixed, but the right-hand side varies over time. Neither
(6.8) nor (6.9) solves the problems of the two constraints (5.18) and (5.20)
simultaneously, whereas the counterpart in consumption theory, equation
(4.12) of Section 6.4, does succeed in this respect with regard to all constraints
(2.13) —(2.15) of Section 6.2.
A well-known measure of industrial concentration is the Herfindahl index,

which is defined as the sum of the squares of all N shares:1
(A.i) c=^yf
1 = 1
The maximum value is 1, which corresponds to the case of complete concen¬

tration: yt= 1 for some /, y~0 for all jAi. The minimum given N is 1/N,
which is attained when all N shares are equal. This minimum approaches
zero when N increases indefinitely.
We now divide the industry into G subsets, S\, ..., SG, so that the index of
concentration among subsets is
(A. 2) Yg=£yt g = \,...,G

g— i 1 e$g
and the index for concentration within Sg:
(A.3) C9= £ (yilYg)2 g = \,...,G

ieSg
We combine (A.I) through (A.3) as follows:

G G
(yJY,f = c„ V c,
3=1 i e Sg 3=1
or equivalently
(A.4) C = C0 X WgCg
3=1
1 The index was suggested by O. C. Herfindahl in “Concentration in the Steel Industry,”

unpublished Ph. D. dissertation, Columbia University, 1950. See also G. Rosenbluth,
“Measures of Concentration,” pages 57-95 of Business Concentration and Price Policy,
Special Conference Series of the National Bureau of Economic Research (Princeton:
Princeton University Press, 1955). The measure is actually older; its square root was
proposed as a concentration index by A. O. Hirschman, National Power and the Structure
of Foreign Trade (Berkeley and Los Angeles: University of California Press, 1945),
Chapter VI.
316
APPENDIX 317
where
Y2 Y2
(A-5) wf = -f=V“ g = l,...,G
° I
l>= 1
We conclude that the index of concentration among all firms of the industry
is equal to the index of concentration among subsets multiplied by a weighted
average of the concentration indices within subsets. The weights are non¬
negative and add up to 1; they are proportional to the square of the relevant
subset share. Since all right-hand C’s in (A.4) are between zero and one,
overall concentration C is at most equal to concentration among subsets C0
and also at most equal to the average within-subset concentration EwgCg.
This is analogous to the entropy decomposition, because the total entropy
(;inverse concentration!) is at least equal to the between-set entropy and also
at least equal to the average (or total) within-set entropy. The entropy de¬
composition is additive, whereas (A.4) is multiplicative. This is not really
surprising in view of the logarithmic character of the entropy measure.
Minimum concentration given the number of firms (N) corresponds to
C=l/N and H(y) = log N. Maximum concentration corresponds to C=1
and H (y) = 0. These two extremes satisfy the relation H(y)= — log C, which
is indeed in accordance with a multiplicative C decomposition and an ad¬
ditive entropy decomposition.
It is a disadvantage of the decomposition (A.4) that the weights wg are
proportional to the squares of the subset shares. Suppose that the industry
consists of three subsets with the following shares:
Yi = -2 Y2 = -4 T3 = .4
so that
C0 = (.2 )2 + (A)2 + (A)2 = .36
Hence w1 = Yx/CQ=%. Next we suppose that the last two subsets are
combined (by a merger, say) so that their new share is .8. The new between-
set concentration is
C0 - (.2)2 + (,8)2 = .68
which is larger (as it should be). This implies that the weight vtq now be¬
comes (.2)2/(.68) = 11t, which is almost 50 per cent less than it was before the
combination of S2 and S3. Hence the weight wq of the within-S^ concen¬
tration is not invariant against such mergers. It is therefore impossible to
define “the contribution of the within-5'1 concentration to the total concen-
tration among all firms” in a manner which depends solely on the shares yt
of the firms in St. In the entropy case this is straightforward. The contri¬
bution is then
YiHi 0) = Z T/ Jog Yl Ti = £ yi
ieSi yi ieSl
which depends only on yh /eS’,.

CHAPTER 9
INPUT-OUTPUT ANALYSIS AND ITS AGGREGATION PROBLEMS
9.1. The Input-Output Technique: A Brief Summary
We considered some special aspects of input-output theory briefly in

Sections 2.6 and 2.7. We now return to this subject to consider it in more
detail, particularly with respect to aggregation. The present section contains
a brief summary of the standard input-output technique. It is followed by two
sections dealing with aggregation. From Section 9.4 onward the information
concept will play an important role in the determination of the appropriate
aggregation method.
Let us imagine that some country’s enterprise system is divided into m
industries. If we take Table 2.3 on page 40 as an example, we may have
Agriculture, forestry and fishing, Food manufacturing industries, Chemicals
and petroleum refineries, and so on. The flow of goods and services from
industry i to industry j is indicated by xtj, the total output of industry i by x{.
We have:
m
(1.1) *i= x xij + fi *' = !,•••>™

j= i
where j\ is the part of f s output which is not absorbed by any of the m indus¬
tries. This is known as the final demand for the products of industry which
consists of commodities delivered directly to consumers, to the government,
to foreign firms (exports), and also of investment goods. Since the inter¬
industry flows xtj are all regarded as current flows (not capital flows),
investment goods supplied by any sector should indeed be considered as
components of final demand.
The decomposition (1.1) describes the total output of each industry as the
sum of m+ 1 nonnegative components. It deals with the output side of the
transactions of each industry. There is also an input side, because the
amounts xXj, ..., xmJ refer to the inputs bought by industry j from the m
industries. In addition, there are the primary inputs: Imports, Wages and
Gross profits in the case of Section 2.6. Let us write m' for the number of
primary inputs and yhj for the amount of the hth primary input which is
319
320 INPUT-OUTPUT ANALYSIS
absorbed by industry j. Then there is another set of m decompositions, now

dealing with the input side of the various industries:
m m'
(1-2) Xj = £ xu + Z Vhj J = 1. •••»m

i= 1 h= 1
On the left we have the total input of industry j. As the notation indicates
[see (1.1)] this is equal to the total output of the same industry. We recall from
Section 2.6 that the input-output system is designed such that this holds for
each industry by definition.
We proceed to consider the ratios of the interindustry flows xtj and the
primary input flows yhj to the total output Xj of the purchasing industry j:
i,j = 1,..., m
(1.3)
h — 1,..., m
The former are the industry input coefficients (or secondary input coefficients),
the latter the primary input coefficients. These m(m + m') coefficients are the
shares whose stability was analyzed in Sections 2.6 and 2.7. If there is indeed
stability in the sense that each dollar output of industry j requires a fixed
amount of au dollars input supplied by i to j, we are able to make perfect
forecasts of total output by industries, given final demand by industries. This
prediction procedure will now be described. Of course, in the real world there
is no stability in the sense described above, so that the forecasts will generally
not be perfect.
We substitute the au definition (1.3) into (1.1), which gives
m
(1.4) xt = Z aijXj + fi i = 1, •••> m

7=1
In matrix form: x = Ax + f or equivalently
(1.5) (I-A)x = f
where x is the m-element column vector of total output by industries, f the

vector of final demand (of the same order), and A= [ay] the m x m matrix of
industry input coefficients. We solve (1.5) for x, which gives
(1.6) x = (I — A)-1f
This is the total output vector which corresponds to the final demand vector f,
given the industry input coefficient matrix A. Suppose now that we have at
our disposal an input-output table of year t and hence, in view of (1.3), the
9.1 THE INPUT-OUTPUT TECHNIQUE 321
coefficient matrix of that year, to be written Af. Suppose that we are inter¬
ested in total output by industries in some later year t + x, t>0, and that we
know the final demand vector fI+T of that year. Then we predict xt+T by
(I —Ar)-1f(+t. This is a conditional forecast, given the final demand vector
ft+T, based on the earlier coefficient matrix Ar. The forecast is obviously
perfect when it is indeed true that the industry input coefficients au do not
change from I to H-t. They do change, of course, which leads to nonzero
prediction errors.1
Going back to (1.1), we see that each industry’s total output consists of a
final demand component. Since final demand is taken as given in the pre¬
diction procedure, it is evidently more realistic to say that this procedure
predicts total output by industries insofar as it is not final demand. So we
subtract final demand from total output of sector i:
m
(1.7) zt = xt -ft= Y xu 1= 1. m
j=i
This is the total intermediate demand for the products of industry i. Writing z
for its m-elenient column vector, we have in view of (1.6) and (1.7):
(1.8) z = [(I- A)"1 — I]f = A(I — A)_1f
Given the final demand vector f(+T of year t + x and the industry input co¬
efficient matrix A( of year t, we can then compute a conditional forecast of
the vector zr+t of total intermediate demand by industries in year t + t. It is
derived from (1.8) simply by attaching the subscripts t and H-i to A and f,
respectively.
For primary input we proceed in the same manner. Consider the primary
coefficient bhj as defined in (1.3). [We recall that for each j the sum over h of
bhj plus the sum over i of equals 1; this follows directly from (1.2). In
fact, this is the decomposition whose stability was analyzed in Sections 2.6
and 2.7.] We have yhj = bhjXj by definition. Hence, if we define
m
(1.9) yh= H yhj h = 1,..., m'

j=1
which is the total demand by all industries of the hth primary input, we have
y = Bx where y is the m '-element primary demand vector and B= [bhj] the
m' xm matrix of primary input coefficients. On combining y = Bx with (1.6)
1 For an extensive analysis of such prediction errors see Applied Economic Forecasting,
Chapters 6 and 7.
we obtain
(1.10) y = B(I — A)-1f
which enables us to make a forecast of y(+t, total primary demand by cate¬

gories in year t + x, given final demand f( + T of that year and on the basis of the
industry input coefficient matrix A, and the primary input coefficient matrix
Bf of some earlier year t.
The results (1.8) and (1.10) are the main vehicles in input-output pre¬
diction. The former deals with the total output of each industry as far as it is
absorbed by the m industries, the latter with the total supply of each primary
input which is absorbed by all m industries. Both sets of forecasts are
conditional on the industries’ output absorbed by economic agents outside
the domestic enterprise system, i.e., on the final demand vector; and both are
derived by means of input coefficient matrices which are computed from
transaction flows of some earlier year.
9.2. The Aggregation Bias of Intermediate Demand Predictions1
The stability of input coefficients is an assumption which is open to doubt

even when our industries consist of one firm each. It is obviously impossible
to work at the level of the individual firm when the objective of the analysis
has a nationwide character (which is the case for input-output analysis). It is
therefore necessary to aggregate firms to sets of firms, the components of
which will in general have different input structures. This will have its effect
on the predictions of primary and intermediate demand. It should be stated
at the outset that there are really two effects. One is concerned with the
reduction in size of the input-output table which is the result of the aggre¬
gation. If we accept the prediction equations (1.8) and (1.10) at the level of
the individual firm, we must expect discrepancies when analogous aggregative
equations are used. The other effect has to do with changes in input co¬
efficients over time. It is conceivable that aggregation leads to increased
stability of these coefficients in spite of the discrepancies just mentioned. The
analysis of this possibility is empirical in character and will not be pursued
here. We shall confine ourselves to the discrepancies of the first effect, which
occur even when the prediction year is the same as the year of the input-
output table which is used in the prediction procedure.
1 The exposition of this section is partly based on H. Theil, “Linear Aggregation in

Input-Output Analysis,” Econometrica, Vol. 25 (1957), pp. 111-122. See also the article
by Malinvaud listed in the Bibliography.
9.2 AGGREGATION BIAS OF INTERMEDIATE DEMAND PREDICTIONS 323
To keep in line with the terminology of the previous section we take the
m industries as our starting point and imagine that they are aggregated to M
industry sets Ix, ..., IM in such a way that each industry belongs to exactly one
set. We introduce the M xm matrix
1 . . 1 0 . . 0 . . . . 0 . . 0
0 . . 0 1. . 1.. .. 0 . . 0
(2.1) E =
0 . . 0 0 . .0.. .. 1. . 1
Each column corresponds to an industry, each row to an industry set. The

industries are arranged such that those which belong to the first set are the
first ones, those which belong to the second set next, and so on; this involves,
of course, no loss of generality. There is a unit element in E whenever the
column corresponds to an industry which belongs to the set of the row, and a
zero element in all other cases. It is easily seen that
-Z*il
ieh
rz*iiieh
~Y.fr
ieh
(2.2) Ex = Ez = Ef =
z
_I6/m
*.•
_ _ie/M
Z —
Y n
—i e IM —
are the vectors of total output, of total intermediate demand and of final
demand, respectively, by industry sets.
We proceed to consider the industry input coefficient matrix at the aggre¬
gative level. Take the flow from Ix to J2. This is the double sum of xu over
ielx,jel2, which is to be divided by the total output of the buying industry
set /2, i.e., by the sum of Xj over jel2, in order that we obtain the corre¬
sponding input coefficient. Hence this coefficient is
Iel jeh
i
Z XU
(2.3) = Z Z auwj
Z
jel2 ie 11 j el2 kel2
tel ijel2
where Wj is the share of the jtb industry in the total output of its set /2. More
generally:
(2.4) Wi =■ ~l - ielg,g = 1,...,M

L xj
jelg
On comparing the result obtained for the input coefficient (2.3) of (7X, I2)
with (2.1) we conclude that the industry input coefficient matrix at the macro¬
level is the MxM matrix EAW', where W is the Mxm matrix
wi • • • wMl 0 ... 0 .... 0... O'

(2.5) W= 0 ... 0 ^nii +1 • • • wmi +,„2.... 0 ... 0
W is of precisely the same form as E except that the unit elements are re¬
placed by the output shares of the individual industries in their sets. In (2.5)
we indicated the number of industries in Ig by mg, g = i, M. Note further
that we have
(2.6) EW' = I (of order M x M)
and
0 ... 0
W2... 0
(2.7) WE = (of order m x m)
0 0 ...WM
where Wg is an mg x mg matrix consisting of identical columns:
. uy Wm i + l • • wmi + i
(2.8) Wi = w2 =
w
_TVmi • ■ _^mi +7712 *
.
*
W
fV m \ + 7712_
We can now proceed to our task of applying the input-output prediction

technique at the aggregative level. We know that after aggregation the
industry input coefficient matrix A is replaced by EAW', and the vectors
x, z, f by Ex, Ez and Ef, respectively. The new object of prediction is the
vector Ez of total intermediate demand by industry sets. If we apply formula
(1.8) subject to the appropriate changes, we find that
(2.9) [(I — EAW')"1 — I] Ef
is the vector corresponding to Ez when we apply the input-output technique

at the macrolevel. But it follows from (1.8) that Ez is actually equal to
Ez = E[(I — A)-1 — I] f
We subtract this expression from (2.9) and conclude that
(2.10) (I — EAW')_1Ef — E(I — A)_1f = Pf

with P defined as
(2.11) P = (I-EAW')-1E-E(I- A)-1
is the error committed by the aggregation procedure with respect to total

intermediate demand by industry sets. This error, Pf, is called the aggregation
bias of intermediate demand. The bias is a linear function of final demand by
(individual) industries and its coefficient matrix Pis determined by the micro-
economic industry input coefficient matrix A and by the output shares of the
individual industries within their respective sets.
The following algebraic manipulations enable us to obtain a more trans¬
parent expression for the crucial matrix P:
E = (I - EAW')-1 (I - EAW')E = (I - EAW')-1E(I - AW'E)
By postmultiplying by (I —AW'E)-1 we find
(2.12) (I - EAW')-1E = E(I - AW'E)-1
and hence, using (2.11),
(2.13) P = E[(I - AW'E)-1 - (I - A)-1]
On the right before brackets we have the “summation matrix” E. It is post-

multiplied by the difference between two matrices which are both of the type
(I —A)-1, but in one of them the A is postmultiplied by the block-diagonal
matrix W'E which is given in (2.7). Furthermore, by replacing the MxM
unit matrix in (2.11) by EW' in accordance with (2.6) we find a third ex¬
pression for P:
(2.14) P = [E(I — A)W']-1E —E(I — A)-1
On postmultiplying the right-hand side by (I —A)W' we obtain I —EW' =

1 — 1 = 0. Hence
(2.15) P (I — A) W' = 0
which amounts to M2 linear constraints on the Mm elements of P.

The aggregation bias Pf can be simplified considerably by means of the
following approximation. We return to (2.13) and expand the two inverses:
P = E[I + AW'E + (AW'E)2 + ... - (I + A + A2 + ...)]
We then take the first-order terms:
(2.16) Px =EA(W'E-I)
and call Pjf the first-order aggregation bias of intermediate demand. Note
that we have
(W'E - I) W' = W' (EW') - W' = W' - W' = 0
and hence
(2.17) PtW'-0
which is even simpler than the corresponding constraint (2.15) on P.

We conclude this section by interpreting the first-order bias and con¬
sidering some special cases. First, take the Mx M matrix EAW' which occurs
on the right of (2.16); its (g, h)th element is
(2-18) X Z
ielg je Ih
auwj = Agh say
Note that this is the (g, h)th macroeconomic input coefficient [see the dis¬
cussion around (2.3)]. Next we observe that EAW' is postmultiplied by Ef in
the bias Pjf and that the gth element of the vector EAW'Ef is
h=l
L a,h I L j s Ih
The gth element of EAf is

m M / \
Z Z
ielgj=l
Z Z Z
h=ljelh\ielg J
We subtract EAf from EAW'Ef to obtain Ptf [see (2.16)] and find that the
gth element of this vector, to be written (P^) is
(2.19) m=h(vuVi
h= 1 jelh \ ielg /
= I Z »jU>-Za,j)/j
h=ljelh \ ielg JWj
f,
= Z (z
h= 1 \ielh /
Z
jelh
wj(Agh-
\
z
iel
The last line implies that each of the M elements of the first-order bias Ptf
can be regarded as the weighted sum of M covariances of the form
(2.20) Z
jelh
WjUgh-
\ ielg
Z J Xj
with weights equal to the corresponding industry outputs Zxt (sum over
iS-4). We can show as follows that (2.20) is indeed a covariance. Take for
any (g, h) combination the differences
X
ielg
au - Agh jelh
These are sums of microeconomic input coefficients measured as deviations

from their (weighted) average; see (2.18). They are multiplied by the corre¬
sponding fj/xj, and the products are weighted by the same w’s as those of
(2.18). Hence (2.20) is the covariance of discrepancies between macro-
economic and sums of corresponding microeconomic input coefficients on
the one hand and final demand as a fraction of total output (fj/xj) on the
other hand. [Note that we should really speak about “negative” covariances,
because we have A^ — Za^ in (2.20), not Za^-A^.]
One conclusion from (2.19) is that the first-order bias vanishes when the
percentage of total output absorbed by final demand is the same for all
industries within each industry set. This follows directly from the fact that,
under the condition stated, fj/xj can be brought before the summation sign
in (2.20). Note that the conclusion holds for whatever values of the input
coefficients an.
Another conclusion is that the first-order bias is zero (for whatever final
demand values) if Zaij, ie Ig, is the same for all jelh, for each (g, h) combi¬
nation. This follows from (2.18), because Agh — Za^ is then zero for all relevant
values of the subscripts. The condition implies that aggregation should take
place such that there is a homogeneous industry input structure in the follow¬
ing sense: All industries which are aggregated to the set Ih require the same
amount of Agh dollars supplied by the set Ig for each dollar of output, for
each of the M2 combinations of g and h. In matrix form this condition
amounts to
(2.21) EAW'E = EA
which is, of course, equivalent to Pt = 0; see (2.16). We shall call (2.21) the
weak condition of industry input homogeneity, the adjective “weak” will be
justified immediately.
Note that (2.21) does not ensure that the matrix P of the aggregation bias
Pf vanishes. For P = 0 we need
(2.22) AW'E = A
see (2.13). This is the strong condition of industry input homogeneity, which
amounts to the following: All industries which are aggregated to the set Ih
require the same amount of
I aUwj
jeih
dollars of input supplied by industry i for each dollar of output, for each
i=l, m and each h — 1, M. Evidently, the strong condition deals with
inputs supplied by individual industries, whereas the weak condition confines
itself to inputs supplied by industry sets.
9.3. The Aggregation Bias of Primary Demand Predictions
The extension of the aggregation analysis to the demand for primary inputs
is largely analogous, but there are some subtle differences. We imagine that
the in' primary inputs are aggregated to M' primary input sets P,,..., Pw; the
arrangement is supposed to be such that Pt consists of the first primary in¬
puts, which are followed by those of P2, and so on. We define the M' x m'
summation matrix E* of “primary aggregation,” the form of which is similar
to that of E as defined in (2.1): Each row corresponds to a primary input set,
each column to an individual primary input, and the element is 1 if the input
of the column falls under the set of the row and 0 otherwise. Consider then
the primary input coefficient which corresponds to the first primary set and
the second industry set:
I I 2 ynj
he Pi jel
(3.1) = X X bhJWj
X
This is the (1, 2)nd element of E*BW'. [Compare (2.3).] We conclude from
(1.10) that the macroeconomic variant of this equation pretends that E*y is
equal to
E*BW (I - EAW')_1Ef
whereas actually
E*y = E*B(I — A)-1f
By subtraction we find the aggregation bias P*f of primary demand, where P*

can be written as follows:
(3.2) P* = E*B [W' (I - EAW')~ *E — (I — A)“]]

= E*B [W'E (I — AW'E)-1 — (I — A)-1]
= E*B[W' (E(I - A) W'}~ *E - (I - A)-1]
The second line is obtained from the first by means of (2.12). The third line is
9.3 AGGREGATION BIAS OF PRIMARY DEMAND PREDICTIONS 329
obtained from the first by replacing the MxM unit matrix by EW', see (2.6).
By postmultiplying the expression in square brackets on the last line by
(I-A)W' we obtain W'-W' = 0 and hence:
(3.3) P*(I - A)W = 0
which is analogous to (2.15).

An obvious question is: When the strong homogeneity condition (2.22) on
the industry input structure is satisfied, is the aggregation bias matrix P* then
necessarily zero? Answer: No. Consider the second line of (3.2), which is
simplified to
(3.4) p* = E*B(W'E — I) (I — A)-1
when (2.22) is true. When the additional condition
(3.5) E*BW'E = E*B
is also satisfied, we do indeed have P* = 0. On comparing (3.5) with (2.21) we

conclude that (3.5) amounts to a weak homogeneity condition with respect
to the primary input structure. This conclusion will be verified in the next
paragraph.
The results obtained can be summarized as follows. (1) A sufficient
condition for P = 0 (zero aggregation bias of intermediate demand for what¬
ever final demand values) is the strong industry input homogeneity condition
(2.22). It is a strong condition in the sense that it deals with the inputs bought
from every single industry. (2) A sufficient condition for the zero value of the
first-order matrix P! is the corresponding weak condition (2.21); it is weak
because it confines itself to inputs bought from industry sets. (3) A sufficient
condition for P* = 0 (zero aggregation bias of primary demand for whatever
final demand values) is the strong condition (2.22) on industry inputs com¬
bined with the weak condition (3.5) on primary inputs. We shall now show
that (4) a sufficient condition for P* = 0 [dealing with the first-order bias of
primary demand, to be defined in (3.6) below] is the weak “primary”
condition (3.5). For this purpose we expand the inverses on the second line
of (3.2):
P* = E*B[W'E + W'E(AW'E) + ... - (I + A + ...)]
and define
(3.6) Pf = E*B(W'E — I)
with P*f as the first-order aggregation bias of primary demand. [Note that
we should use only one term in each expansion, because we already have one
input coefficient matrix (B) before the brackets.] We observe that P* satisfies
(3.7) PfW' = 0
in view of (W'E — I)W' = W' (EW') — W' = 0. This result is analogous to (2.17).
We also observe that (3.5) is a sufficient condition for P* = 0 as asserted
above. This condition implies that ZbtJ (sum over iePk), which is the (k,j)lh
element of the M'xm matrix E*B, should be equal to the corresponding
element of E*BW'E, which is
(3.8) £ YJhirwr=Bkh say (jelh)

iePkrelh
and which is, in turn, the primary input coefficient corresponding to the sets
Pk and 4; see (3.1). Evidently, condition (3.5) implies that the industries
which are aggregated to some set Ih require the same amount of Bkh dollars of
input supplied by the primary set Pk for each dollar of output, for each
h — 1, ..., M and each k— 1, ..., M'. This homogeneity condition is weak,
because it deals with primary input sets rather than individual primary inputs.
Note that the corresponding strong condition
BW E = B
is not required, neither for P* = 0 nor for P* = 0.

We can obtain an expression similar to (2.19) for the elements of the first-
order bias P*f of primary demand. The kth element of the TIT'-element vector
E*BW'Ef is M
E
h=l
bu e
jelh
fj
and the kth element of E*Bf is
Z Z
iePkJ=l
bufj= Z Z (Z
h=l jeIh\iePk
hij)fj
/
Hence, using (3.6), we find for the klh element of P*f:
(3.9) (Ptf). = E E (a. - E

h = 1 j e Ih \ iePk
t,)fj
/
M /
fi
=Z Z
h=l jelh
wj\Bkh- Z
i6 Pk
bu
JWj
fj
Z (z n)J Z
h=l\ielh jelh
wj(Bkh
\
- Z
iePk
bij
y
9.4 THE INFORMATION CONTENT OF AN INPUT-OUTPUT TABLE 331
Here again, as in the case of (2.19), we can conclude that each of the M'
elements of the first-order bias is equal to the weighted sum of M covariances
dealing with input coefficients (here: primary input coefficients) and final
demand as a fraction of total output. The interpretation is entirely analo¬
gous to that of (2.19), so that we can refer to the discussion below that
equation.
9.4. The Information Content of an Input-Output Table
We now turn to the following problem: If it is decided to aggregate the

m + m' industries and primary inputs to M+M' industry and primary sets,
this can be done in a very large number of ways; but is it possible to formulate
a sensible criterion according to which one particular aggregation method is
better than all others? We shall try to answer this question by means of in¬
formational concepts.1 Let us start by writing the m equations (1.1) in the
following form:
*11 + *12 + ••• + *1 m = *1 ~ fl
(4 *21 + *22 + ••• + * 2m = *2 ~ f2
*inl + *m2 T T Vmm *m fm
and the m' equations (1.9) as
yn + y i2 + + y lm =1T
(4.2)
y 21 + L22 + •• + y 2m =y 2
y m' 1 + ym'2 + ••• + ym'm — yn
The left-hand side of these m + m' equations consists of m(m + m') non¬
negative terms which add up to Ixt (sum over i from 1 through m), which is
the aggregate output (or input) of all m industries. If we divide both sides of
each of the m + m' equations by Ixh the m(m + m') terms on the left add up
to 1 and can therefore be formally regarded as probabilities of a bivariate
distribution, so that they can be handled by means of informational concepts.
1 The exposition is largely based on H. Theil and P. Uribe, “The Information Approach
to the Aggregation of Input-Output Tables,” Report 6503 of the Center for Mathematical
Studies in Business and Economics, The University of Chicago (1965). Reference should
also be made to an earlier and similar approach by J. Skolka, The Aggregation Problem
in Input-Output Analysis (Czech with English summary), Czechoslovakian Academy of
Sciences, Prague, 1964.
In the case of (4.1) this leads to
(4.3) Pu= *‘J i’J — !’•••’m

Z
k= 1
Xk
i.e., to interindustry flows measured as fractions of the aggregate output

(or input) of all industries. In the case of (4.2):
y*j h = 1,..., m';j = 1

(4.4) m
Z1
r=
xk
which are the primary inputs absorbed by the various industries, also
measured as fractions of aggregate input. Finally, the terms on the right in
(4.1) and (4.2) become
Xj ~fi= Zi
Pi. m m
i = 1,...,m
k=1
Z xk Z
k= 1
(4.5)
yh
Pm + h,. m h = 1,...,m
Z
fc = 1
xk
which are, respectively, the total intermediate demand for the products of the
;th industry and the total flow of the hlh primary input, both divided by Ixk.
We have thus succeeded in expressing all elementary flows in terms of an
(,m + m’) x m array (Pij), the elements of which behave as probabilities in the
sense that PijÔ, ZIpij= 1. The algebra which follows will be simplified
considerably when we interpret the array as being square of order m + m',
which is performed easily by adding an appropriate number of zeros:
(4.6) Pij = ® i — 1, ...,m + m’; j — m + \, ...,m + m'
Also, it will be convenient to write
(4.7) n — m + m'
for the total number of industries and primary inputs, and to use the word
“sector” whenever we mean either an industry or a primary input. The reason
is that the analysis which follows is symmetric in industries and primary in¬
puts. An example of this notation and terminology may be instructive. Take
the case of three industries and one primary input, hence /? = 4 sectors; then
9.4 THE INFORMATION CONTENT OF AN INPUT-OUTPUT TABLE 333
the array (p^) is of order 4x4 and its fourth column consists of zeros, for
example:
Receiving sector j
Supplying sector i 1 2 3 4 Sum pi
Pij
1 .05 .05 0 0 .1
2 .05 .1 .05 0 .2
3 .1 .05 .1 0 .25
4 (primary) .2 .1 .15 0 .45
Sum p j .4 .3 .3 0 1
The last row contains the column sumspj, defined as follows:
(4-8) P.j = -ppL-
Z xk
k= 1
0 j = m + 1, ...,n
Note that we have
(4.9) Pt.^P.i i — 1,m
because p i-pi, is equal tof/Ixk for z = 1, m, which is nonnegative.

Consider now the expected mutual information of the bivariate array .):
(4.10) I=Hpu 1°« —

i=t j=l Pi.P.j
This will be called the information content of the input-output table,1 which
can be justified as follows. Suppose that the array (p;j) is characterized by
independence: Pij=Pi.Pj for all pairs (z, j). Then we would not need the
input-output table at all, because the separate flows from one sector (primary
or secondary) to another can be derived directly from the corresponding
marginal totals: total input by industries, total intermediate demand by
industries, and the total flow of each primary input. The information content /
of the table is zero in that case. We do need the table when there is no
independence, so that the separate flows cannot be computed from the
marginal totals. The corresponding / will then take a positive value, which
becomes larger and larger when the flows deviate more and more from the
independence pattern.
1 This measure was proposed by J. Skolka, loc. cit.

It is of some interest to consider the independence pattern from the stand¬

point of input-output prediction of intermediate and primary demand. In
the present notation the (i,j)th industry input coefficient is written as
VU XjjlZxk PjJ
(4.11) £J< j i,j = 1
Xj/ZXk P-j
which is simplified to aiJ=pii when there is indeed independence. Hence the

industry input coefficients are then independent of the index of the receiving
industry, which amounts to strong homogeneity of the industry input
structure throughout the economy. In that case the total intermediate demand
for the products of industry i is
mm m
(4-12) zt = Z *ij = Z atjXj = Pu Z XJ

j= i j= i j=i
i.e., intermediate demand 2,- is proportional to aggregate industry output.

After summation over i we find
or equivalently
m
m z/i
(4.13) Z *1 =
Z
k = m+ 1
Pk.
On combining (4.12) with (4.13) we find
(4-14) z, = A V fj i = 1,.... m
I
k = m+ 1
P,k
which implies that the total intermediate demand for the products of each
separate industry is predicted to be proportional to the aggregate final
demand for the products of all industries. This is a variant of the final-
demand-blowup extrapolation method. It is actually a very simple variant,
because the method usually consists of predicting zt to be proportional to the
corresponding/;,1 which - contrary to (4.14) - requires that the individual
final demand values/l5 ...,/m of the prediction year be known.
1 See, for example, Applied Economic Forecasting, Section 6.5.

9.5 THE INFORMATION DECOMPOSITION EQUATION 335
For primary demand we proceed in the same way. The (h,j)ih primary input
coefficient is
yhj/ZXk Pm + h, j h = l,..., n — m(= m!)
(4.15) bh; =
xtlExk P.J j = 1
which is simplified to bhj=pm+lu in the independence case. This amounts to

strong homogeneity of the primary input structure throughout the economy.
We have
yn= Z bhjXj = X xj h = 1, ...,n - m(= m')
j=i J= i
and hence in view of (4.13):
Pm + h,.
(4.16) /y h = 1,..., n — m ( = m ')
Z T/c.;=i
k = m+ 1
the interpretation of which is analogous to that of (4.14).
9.5. The Information Decomposition Equation of Input-Output Aggregation
We now return to the subject of aggregation and suppose, as we did in

Sections 9.2 and 9.3, that the m industries are combined to M industry sets
and the m —n — m primary inputs to M' primary input sets. Since the in¬
formation approach handles industries and primary inputs symmetrically
and simultaneously, we can simplify the notation by writing
(5.1) G — M + M'
for the total number of sets and by indicating the sets themselves by Sg,
g = 1, ..., G. They will be numbered such that the industry sets are first and
the primary sets last:
— h $m+i = Pi
— hi Sr. = PW
The notation of the macroeconomic flows is then straightforward:

M
M
Pij g,h = l,. ,.,G

i E Sg j E Sh
(5.2) I P.H-= z P, <7 = 1,- ,.,G

h =1 ieSg
I p.„-= Z P.J h = 1,. • G

0=1 jeSh
These are, respectively, the flow from Sg to Sh, the total flow originating
in Sg, and the total flow going to Sh, all measured as fractions of the aggre¬
gate output of all industries.
The information content of the aggregated table is
G G
PajL
(5.3) /o = I I P9h log
g=lh=l PgP.H
and the obvious question is: What is the relationship between this 70 and the
information content I of the original table as defined in (4.10)? We shall
illustrate this problem with the numerical example that was given below
equation (4.7), assuming that the first two industries are aggregated. Hence
5'i=(l, 2), 52 = (3), S3 = (4) with the following array:
Receiving set Sh
Supplying set Sg Si S' 2 S3 Sum
P9h
s, .25 .05 0 .3
s2 .15 .1 0 .25
S3 (primary) .3 .15 0 .45
Sum Ph .7 .3 0 1
The corresponding /0 value is
.25
(5.4) Io = -25 log + .05 log
(■3) (.7) cm .Wi
which should be compared with the I of the table before aggregation:
.05 .05
(5.5) I = .05 log + .05 log + ... = .102 bit
U)04)
These outcomes illustrate the inequality
(5.6)
In words: The information content of an input-output table cannot increase

by aggregation (and it will usually decrease). Our next task is to prove this
inequality.
For this purpose we need some symbols indicating halfway aggregation.

When summing the p’s over a row index in a set Sg we write
(5.7) rgj= Yj Pu g = l,...,G;j = l,...,n

ieSg
This is the flow from set Sg to the individual sector j, measured as a fraction
of aggregate industry output. When summing the p’s over a column index in
a set Sh we write
(5.8) cih= Y Pu i = l,...,n; h = 1,...,G
which is the flow from sector i to set Sh, measured in the same way. The
following summation rules will be obvious:
G
Z r9j = P.j Y rgj Pgh

9=1 jeSh
(5.9) G
Y cih = Pg» Y cih = Pi.

ieSg h= 1
We now consider the difference I—I0, which is asserted to be nonnegative

in (5.6). This difference can be written as follows:
rgh
1 - Jo = I I Z Z Pu log —- - log
g= 1 h= 1 leSg jeSh Pi.P.j P,P.K
The logarithmic difference in brackets can be written as
iog + log AA + log AT*

P.jlPj PiJPg. Vi c»
Pgh Pgh
so that we conclude
f gjlPgh
I-*o= Y Y Y rgj lo§
g= 1 h= 1 jeSh P.jlP.h
Cihl Pgh
+ Y Y Y cth log
g= 1 h = 1 ieSg PiJPg.
G G
+ 111 lPu 1°8

g=lh = lieSgjeSh rgj cih
Pgh Pgh
This result can be written in the following alternative form:
(5.10) I = Io+Y PJ.H + Z Pgh. + Z Z Pghgh

h= 1 9= 1 9= 1h= 1
where is the column effect of Sh:

G
1 gjlPgh
(5.11)
P.j/P.h
9=1 jeSh
and Ig is the row effect of Sg:

G
Cihl Pgh
(5.12) 9-
log
PiJPg.
and Igh a cell effect dealing with the flows from the sectors of Sg to those of Sh:
PijIPgh
(5.13) log
rgj cih
j e Sg je Sh
xPgh xPgh
Equation (5.10) is the information decomposition equation of input-output

aggregation. Our task is now to interpret its various terms. It will turn out
that Ih, Ig and Igh can all be regarded as information expectations of indirect
messages. Since they are all multiplied by nonnegative coefficients in (5.10),
we conclude that / cannot be smaller than I0, which proves (5.6).
Let us start with the column effect I h as given in (5.11). We conclude that
this effect is a weighted average of
IgjlPgh
(5.14) -- log
gh P.j/P.h
j eSh
which is the expected information of the indirect message that transforms the
prior probabilities p j/Ph,jeSh, to the posterior probabilities rgj/PghJeSh. It
is positive unless
(5.15) !jl = Ed or Vgj =Pgh for each jeSh

Pgh P.h P.j P.h
in which case the value of (5.14) is zero. The ratio rgj/pj is the fraction of
sector/s total input which it buys from Sg. The ratio Pgh/Ph is the fraction of
set Sfs total input bought from Sg, sector j being one of the sectors of the
buying set Sh. The two ratios are equal for each jeSh when all sectors of Sh
have the same input coefficient for the supplies of Sg. As soon as this is not
the case, the value of (5.14) will be positive. Furthermore, Ih as defined in
(5.11) is a weighted average with nonnegative weights Pgh/P.h of expressions
of the form (5.14). The conclusion is that Ih = 0 if and only if Sh is charac-
terized by weak industry input homogeneity and also by weak primary input
homogeneity, and I h>0 otherwise. The verification of this conclusion is as
follows:
(5.16) if g< M
P.j
i G Sg
Y, bjj if g — M = k > 0
iePk
Now if rgj/p j is the same for all jeSh for each g = 1, ..., G, we must have
(5.17) Yau = Agh jeSh; g =

ielg
which amounts to weak industry input homogeneity of Sh [see (2.18)], and

also
(5.18) I bu = Bkh jeSh; k = 1,

iePk
which is weak primary input homogeneity of Sh [see (3.8)]. If (5.17) and (5.18)
apply to all purchasing sets Su ..., SM, we have
(5.19) EA = EAW'E E*B = E*BW'E
which amounts to weak input homogeneity, both primary and secondary, of

all M purchasing industry sets Al5 ..., SM;1 see (2.21) and (3.5).
This “double” weak input homogeneity condition ensures a zero value of
the second right-hand term of the decomposition (5.10), TPh7h. When the
condition is violated by any Sh, i.e., when there is a set Sh consisting of
sectors whose input structures differ (primary or secondary), Ih and hence
IPhI.h take positive values. We conclude that IPhIh is an informational
measure of the input heterogeneity of the aggregation procedure, and that
Phlh is the part of the total input heterogeneity which is to be attributed
to Sh.
In the same way we can regard the third term of (5.10), ZPg Ig , as a measure
of the output heterogeneity. We see from (5.12) that Ig is a weighted average of
CilJPgh
(5.20) log
PiJPg.
1 Note that the industry sets Si,..., Sm ( = /i, ..., hi) are the only purchasing sets, because
there are no flows going to the primary sets. The input heterogeneity (to be defined in the
next paragraph) is confined to the industry sets.
which is the expected information of the indirect message which transforms

the prior probabilities pJPgieSg, to the posterior probabilities cih/Pgh, ieSg.
It is positive unless
(5.21) —= — or ih _ gh por each ieSg

P9h Pa. Pi. P,
in which case the value of (5.20) is zero. The ratio cJpL is the fraction of
sector Fs total output which it sells to Sh. The ratio Pgh/Pg. is the fraction of
the total output of Fs set Sg which this set sells to Sh. Condition (5.21) implies
that all sectors of Sg ship the same percentage of their total output to Sh;
going back to (5.12), we conclude that Ig = 0 if and only if this is true for all
buying sets Sh. When this is the case, we may indeed say that the supplying
set Sg is characterized by (weak) output homogeneity. The weighted sum
SPg lg in (5.10) measures the output heterogeneity of the aggregation pro¬
cedure as a whole.
We now turn to the third term of the decomposition (5.10), which is a
weighted average of Igh as defined in (5.13). This is the expected information
of the indirect message whose prior and posterior probabilities are
1 gj Cih j Pij • c • c
—
p p
- and p
- teSa,jeSh
gi j n
rgh rgh rgh
respectively. Again, it is nonnegative and zero if and only if there is a pairwise

equality of prior and posterior probabilities. The special zero case implies
that the flow from i to j, measured as a fraction of the total flow from Fs
set Sg to j's set Sh, can be computed according to the rules of stochastic
independence. This follows from
see (5.7) and (5.8). The total cell effect IIPghlgh measures the extent to which
the separate flows within each macroeconomic flow deviate from the in¬
dependence pattern.
The information loss /-/0 of our numerical example is .072 bit [see (5.4)
and (5.5)]. It is a matter of straightforward computation to find that the total
input heterogeneity ZPhIh = .034 bit, the total output heterogeneity IPg Ig =
.033 bit, and the total cell effect SIPghIgh = .005 bit. Hence in this case the cell
effect is of minor importance, but the input and output heterogeneity are
sizable and even larger than the information content /o = .030 bit after aggre-
9.6 AN APPLICATION TO DUTCH INPUT-OUTPUT TABLES 341
gation. Note, however, that the three components of the information loss do
not have the same meaning with respect to the prediction of primary and
intermediate demand. The information approach treats rows and columns
symmetrically. It leads to a row or output heterogeneity effect, to a column
or input heterogeneity effect, and to a cell effect. As we know from Sections
9.2 and 9.3, it is only the input effect that counts when we predict primary
and intermediate demand given final demand. This is no defect of the in¬
formation approach; it enables us to evaluate the merits of the aggregation
procedure on the basis of the input heterogeneity alone. This will be illustrated
in the remaining sections of this chapter.
,
9.6. An Application to the Dutch Input-Output Tables 1949-1960
We shall now apply the ideas set forth in the two previous sections to the
annual Dutch input-output tables in the years 1949-1960. These tables are
based on a 35-industry classification. Following C. B. Tilanus we shall
aggregate these to 11 industry sets.1 Four sets consist of only one industry,
seven sets contain two industries or more. The following are the 11 sets:
(1) Food, beverages and tobacco consisting of 3 industries: Food manu¬
facturing industries (animal products), Food manufacturing industries (other
products), and Beverage industries and tobacco manufactures
(2) Agriculture, forestry and fishing (a one-industry set)
(3) Textiles, footwear and other wearing apparel consisting of 2 industries:
Textiles and Footwear and other wearing apparel
(4) Chemicals and petroleum refineries (a one-industry set)
(5) Metal industry consisting of 5 industries: Basic metal industries, Metal
products and machinery, Electrical machinery and apparatus, Transport
equipment, and Other metal products and diamond industry
(6) Construction (a one-industry set)
(7) Electricity, gas and water (a one-industry set)
(8) Other manufacturing industries consisting of 7 industries: Coal mining,
Crude petroleum and salt mining, peat cutting, etc., Wood and furniture,
Paper and paper products, Printing, publishing and allied industries, Leather
and rubber products (excl. footwear), and Earthenware, glass, lime and
stoneware
(9) Trade (margins), consisting of 2 industries: Wholesale trade and
Retail trade
1 C. B. Tilanus, “Thirteen Aggregated Input-Output Tables, The Netherlands 1948-1960,”

Report 6423 of the Econometric Institute of the Netherlands School of Economics (1964).
(10) Transport, storage and communication consisting of 3 industries: Ocean

and air transport, Other transport, and Communication
(11) Other service industries consisting of 9 industries: Banks and other
financial institutions, Insurance, Ownership of residential housing, Medical
and related services, Business services. Recreation services, Hotels, cafes and
restaurants, Other personal services, and Goods and services not elsewhere
classified.
We shall apply no aggregation to primary inputs. There are three such
inputs, and hence also three primary input sets:1
(12) Imports
(13) Wages
(14) Gross profits.
As a whole, therefore, there are 14 sets Sx, ..., iS14 consisting of 11 industry
sets Su ..., and 3 primary sets S12, S13, S14.. They are obtained from
38 individual sectors, 35 of which are industries and 3 are primary inputs.
The information content / of each of the twelve 38-sector tables is pre¬
sented in Table 9.1, together with its components according to the infor¬
mation decomposition equation (5.10). It turns out that / is fairly constant
around a level of about .9 bit. The information content I0 after the aggre¬
gation to 14 sector sets is on the average slightly larger than .5 bit; the
aggregation reduces the information content of the table by a little over
40 per cent. The input heterogeneity is primarily responsible for this re¬
duction; it amounts to about .2 bit on the average and accounts for almost
60 per cent of the information reduction /— 70. The output heterogeneity is
less than .1 bit on the average and accounts for only 25 per cent of the re¬
duction. The cell effect accounts for less than 20 per cent.
An obvious question is which industry sets are mainly responsible for the
reduction of the information content. This question is answered in Table 9.2,
which contains the contributions PhIh of the various sets to the total
1 Actually, there is aggregation with respect to primary inputs, because the Netherlands
Central Bureau of Statistics distinguishes between the following seven: (a) Imports, (b)
Wages and salaries, (c) Employers’ social insurance fees, (d) Depreciation, (e) Indirect
taxes, (f) Subsidies (treated as a negative input), and (g) Other income such as profits and
interest. Primary input (12) consists of (a), (13) of (b) and (c), and (14) of (d) through (g).
Since Subsidies are negative inputs and Other income is also negative for some industries
in some years, and the information approach can handle only nonnegative numbers, it is
impossible to proceed on the basis of the seven individual primary inputs. They are
therefore combined to the three listed in the text prior to the aggregation procedure
that will be analyzed. [In one year (1948) even the value of Gross profits is negative for
one industry; that year is deleted from the analysis.]
9.6 AN APPLICATION TO DUTCH INPUT-OUTPUT TABLES 343
TABLE 9.1
THE INFORMATION DECOMPOSITION OF TWELVE INPUT-OUTPUT TABLES
Year of the Information

Information Input Output Cell
input- content after
content heterogeneity heterogeneity effect
output aggre¬
/ ZP.hl.h SPg.Ig. ' H Pghlgh
table gation Io
In IO'4 bit
1949 8546 4788 2154 922 683
1950 9046 5228 2275 869 674
1951 9205 5414 2243 903 645
1952 9090 5336 2151 919 685
1953 8899 5201 2069 953 677
1954 8998 5202 2184 932 680
1955 8825 5262 1980 938 644
1956 8754 5170 1985 949 650
1957 8859 5177 2023 1005 654
1958 8705 5059 1978 990 678
1959 8715 5103 1958 980 673
1960 8588 5036 1926 980 645
Average 8853 5165 2077 945 666
As a percentage of I
1949 100 56 25 11 8
1950 100 58 25 10 7
1951 100 59 24 10 7
1952 100 59 24 10 8
1953 100 58 23 11 8
1954 100 58 24 10 8
1955 100 60 22 11 7
1956 100 59 23 11 7
1957 100 58 23 11 7
1958 100 58 23 11 8
1959 100 59 22 11 8
1960 100 59 22 11 8
Average 100 58 23 11 8
As a percentage of 1 —In
1949 57 25 18
1950 60 23 18
1951 59 24 17
1952 57 24 18
1953 56 26 18
1954 58 25 18
1955 56 26 18
1956 55 26 18
1957 55 27 18
1958 54 27 19
1959 54 27 19
1960 54 28 18
Average 56 26 18
TABLE 9.2
CONTRIBUTIONS OF INDUSTRY SETS TO INPUT AND OUTPUT HETEROGENEITY
Textiles, footwear
wearing apparel
Transportation,
communication
Year of
manufacturing
and tobacco
the input-
storage and
Total
and other
beverages
industries
industries
(margins)
output
industry
service
table
Food,
Trade
Metal
Other
Other
5
o
Input heterogeneity (P.nl.h in
1
1949 581 80 61 200 145 208 878 2154
1950 768 62 66 216 120 216 827 2275
1951 809 65 100 245 95 196 732 2243
1952 761 52 91 217 101 207 722 2151
1953 716 50 96 204 108 190 705 2069
1954 747 57 98 204 123 197 757 2184
1955 678 44 99 214 110 193 643 1980
1956 669 38 110 203 111 182 672 1985
1957 672 38 135 204 104 186 685 2023
1958 615 27 123 187 109 174 742 1978
1959 642 30 132 186 122 165 681 1958
1960 585 28 156 185 130 155 686 1926
Average 687 48 106 205 115 189 728 2077
Output heterogeneity (Pg.Ig. in 10-4 bit)

1949 98 15 233 463 0 47 66 922
1950 105 13 204 441 0 48 57 869
1951 97 11 235 463 0 36 61 903
1952 104 13 231 467 0 42 62 919
1953 99 14 242 477 0 53 67 953
1954 98 14 217 476 0 61 67 932
1955 93 14 245 471 0 50 65 938
1956 90 13 271 468 0 44 62 949
1957 96 11 277 507 0 43 71 1005
1958 103 12 267 490 0 43 75 990
1959 109 12 269 468 0 46 77 980
1960 109 6 277 453 0 45 90 980
Average 100 12 247 470 0 47 68 945
input heterogeneity as well as the contributions Pg Iq to the total output

heterogeneity. The table is confined to the seven industry sets which contain
more than one industry; the contributions of the one-sector sets are obviously
zero. It turns out that the largest values are those of the input heterogeneity
of Other service industries, which consist of 9 separate industries; the average
9.7 PARTIAL DISAGGREGATION OF THE FOOD INDUSTRY 345
of the corresponding PhIh is slightly larger than .07 bit. The input hetero¬
geneity due to Food, beverages and tobacco is just below .07 bit on the
average; the output heterogeneity Pg Ig due to Other manufacturing industries
is third.1
What are the conclusions to be drawn from this kind of numerical data
with respect to appropriate input-output aggregation? We know, of course,
that as far as the aggregation bias is concerned, only the input heterogeneity
is a matter of concern. Let us then imagine that minimization of the total
input heterogeneity is our criterion. [We shall consider the relationship
between this criterion and the aggregation bias in more detail in Section 9.8.]
In principle, this is straightforward. Given the n sectors from which we start,
we can aggregate them to a fixed number G of sector sets by taking all
possible combinations and choosing the aggregation method which is char¬
acterized by the smallest value of the total input heterogeneity IPhIh.
It is also possible to take certain restrictions into account which indicate
that one does not want to combine particular sectors. Mathematically, the
problem is of the combinatorial type, and the necessary computations may
be sizable. We shall confine ourselves to a more modest goal in the next
section.
9.7. Partial Disaggregation of the Food Industry Set
Consider Su the Food industry set, which is the second largest contributor
to the total input heterogeneity. We write the information content I0 as the
sum of four terms, the fourth of which does not involve :
(7.1)
We shall disaggregate into two subsets:

(1) Food manufacturing of animal products (a one-industry subset)
(2) Other foods, beverages and tobacco consisting of 2 industries: Food
manufacturing of non-animal products and Beverage industries and tobacco
manufactures.
1 The zero values of the output heterogeneity of the Trade industry set are simply due to
the fact that only Wholesale trade produces for intermediate demand. There are no flows
from Retail trade to any industries.
The flows between these subsets, measured as fractions of the aggregate

output of all industries, will be indicated by qrs, where r or s=l refers to
Food manufacturing of animal products, and r or 5=2 to the other subset.
The flows from these subsets to S,„ h = 2, ..., G, will be indicated by Qrh
(measured in the same way), and the flows from Sg to these subsets by Q'gs>
g — 2, ..., G. We obviously have:
2 2
Z Z
r= 1 s= 1
QrS = Pi 1
(7-2) Z Qrh = Plh h = 2, ...,G

r— 1
E Q’„ = p„ g = 2,..., G
S= 1
The information content of the table after this partial disaggregation is then
Qrh
(7.3) T*
Jo Z Z 4rs log
r = 1 s= 1
+ Z Z
r =1h = 2
Qrh log
Qr.Q, Qr.P.h
e: P9h_
+ Z Z
g=2s=l
Qgs log + I Z p9h log
Pg.Q.s g=2h = 2 Pg-P-h
where
2 G
Qr. ~ Z
s= 1
Qrs+ Z
/i = 2
Qrh r=l,2
(7.4)
2 G
(?.s “ Z
r= 1
CGs + Y. Q'g
9=2
0* ^ — 1?2
represent the total flows originating in the rth subset and going to the 5th
subset, respectively.
On comparing (7.3) with (7.1) we conclude that the difference I* — 70 can
be written as the sum of the three terms:
(7.5) I* -/0 = Pndn+ YPlh^lh+

h=2
PglAgl Z
9=2
where
2 2
9rs , tfrslP 11
(7.6) *11 = — log-
Pit Qr. (L
r= 1 s= 1
Pl.P 1
9.7 PARTIAL DISAGGREGATION OF THE FOOD INDUSTRY 347
s= 1
The first right-hand term, of (7.5), PnAlu is the information loss which is
incurred when the four subset flows within Food, beverages and tobacco are
aggregated to one set flow. Equivalently, it is the gain in information content
which is obtained when we disaggregate the set flow to four subset flows. We
conclude that when this partial disaggregation is applied (13 sets plus 2 sub¬
sets instead of 14 sets), the information content of the table becomes /*
(instead of I0) and the total cell effect becomes EIPghIgh — P11All, which is
less than or equal to the original value IIPghIgh (sum over g and h from 1
through G).
Given the remarks made at the end of the previous section, the most
interesting thing is of course the reduction of the input heterogeneity. For
this purpose we consider the last right-hand term of (7.5) as well as (7.8). They
deal with the expected information of the indirect message whose prior
probabilities are the shares QJPA of the two subsets in the total input of
their set Sx and whose posterior probabilities are the corresponding shares
Q'gs/Pgi °f the input bought from a particular set Sg. We have Agl =0 if and
only if
Qgs = 2^ or Q'gs = Pg 1
5 = 1,2
Pgl P. 1 Q.s P. 1
which means that, as soon as Si is characterized by some (weak) input

heterogeneity, Agi will be positive. It will be clear that IPglAgl is the re¬
duction of the informational value ZPhIh of the total input heterogeneity
which is obtained by the partial disaggregation. In the same way we have to
subtract ZPlhAlh from IPg Ig to find the new value of the total output
heterogeneity. This is immediately seen when it is realized that Alh = 0 if and
only if
which means that the two subsets supply the same percentage of their total
output to Sh.
TABLE 9.3
THE GAIN IN INFORMATION CONTENT OBTAINED BY PARTIAL DISAGGREGATION
OF THE FOOD INDUSTRY SET
Year of the Reduction of Information

input-output input output cell content of the
table heterogeneity heterogeneity effect 15-sector table
E Pg\A(jl E PihAih PnAn n
1949 343 16 132 5279
1950 544 20 147 5939
1951 552 22 145 6133
1952 579 22 162 6099
1953 539 23 145 5908
1954 583 25 138 5948
1955 503 21 123 5909
1956 496 22 123 5811
1957 493 25 119 5814
1958 456 29 129 5673
1959 475 32 130 5740
1960 425 33 116 5610
Average 499 24 134 5822
Note. All figures are in 10-4 bit.
Table 9.3 contains the three components of the gain which is obtained by
the disaggregation, together with the information content /* after the dis¬
aggregation. It turns out that the input heterogeneity due to the Food
industry set is reduced by almost 75 per cent (see the first column of Table 9.2).
The total input heterogeneity is reduced by almost 25 per cent. We note in
passing that there is a minor reduction of the output heterogeneity; that there
is (on the average) a 20 per cent reduction of the total cell effect; and that the
average I* exceeds the average I0 by a little less than 15 per cent, so that the
information loss due to the aggregation of the 35 industries (I-I0 and I-I*)
is reduced from .37 to .30 bit on the average.
9.8. Input-Output Prediction Before and After Disaggregation; A Minimax

Interpretation of the Input Heterogeneity Criterion
We shall now subject the disaggregation of the previous section to a test by

predicting total intermediate demand for the products of each industry set as
well as the total flow of each primary input set, using equation (2.9) and the
9.8 A MINIMAX INTERPRETATION 349
corresponding “primary” prediction equation. We confine ourselves to the

aggregation bias proper by taking the input coefficient matrices and the final
demand vector of the same year. Hence we do not really predict and our
“forecasts” would be trivially perfect if there were no aggregation. To increase
the comparability of the numerical results in successive years, we shall
express the discrepancies in logarithmic form, not in current guilders. Thus,
let the observed value of the total intermediate demand for the products of Sg
be Zgt in year t, and the value computed from (2.9) be Zgt, then the loga¬
rithmic error is
(8-!) egt = log
where log stands for natural logarithm. The procedure for primary demand is
completely analogous. The individual errors are given in the Appendix of this
chapter. Here we confine ourselves to the mean square errors:
where t takes the 12 values 1949-1960. These mean squares are presented in
Table 9.4 for all industry sets and for the three primary inputs, both for the
(7=14 sets of Section 9.6 and after the partial disaggregation of the Food
industry set. The main conclusions are the following two. (1) The dis¬
crepancies are all very small; if we take square roots, we obtain root-mean-
square values which vary between \\ and 6 of one-hundredth of 1 per cent
in the case of the intermediate demand for the products of the 14 industry
sets, and still less in the case of primary demand. [The latter result is not
really unexpected, because the first-order aggregation bias of primary
demand vanishes when there is no aggregation of primary inputs.] (2) For
most of the sets the disaggregation has virtually no effect on the mean square
error, and in the only case in which the effect is fairly substantial - Agri¬
culture, forestry and fishing - the disaggregation raises rather than lowers
the mean square error.
How should we explain these results, particularly the latter, given that the
disaggregation was rather successful with respect to the input heterogeneity?
To answer this question we go back to equation (2.19), which describes the
elements of the first-order aggregation bias of intermediate demand as a
weighted sum of certain covariances, and to the corresponding equation (3.9)
for primary demand. These covariances deal with two sets of “variables.”
One is final demand measured as a fraction of total output; this cannot be
TABLE 9.4
MEAN SQUARE ERRORS OF INTERMEDIATE AND PRIMARY DEMAND PREDICTIONS BEFORE AND
AFTER PARTIAL DISAGGREGATION OF THE FOOD INDUSTRY SET
Mean square error

Group or subgroup before after
disaggregation disaggregation
1. Food, beverages and tobacco 471 472

la. Food manufacturing (animal products) 6265
lb. Other food, beverages and tobacco 398
2. Agriculture, forestry and fishing 263 332
3. Textile, footwear and other wearing apparel 1186 1186
4. Chemicals and petroleum refineries 462 460
5. Metal industry 1873 1909
6. Construction 2233 2204
7. Electricity, gas and water 3700 3669
8. Other manufacturing industries 1361 1363
9. Trade (margins) 928 926
10. Transport, storage and communication 2218 2226
11. Other service industries 5394 5400
12. Imports 18 18
13. Wages 41 40
14. Gross profits 25 25
Note. All figures are to be multiplied by 1CL10.
controlled by the aggregation procedure, because final demand is exogenous

with respect to the input-output model. The other concerns discrepancies
between micro- and macroeconomic input coefficients, and these dis¬
crepancies can to some extent be controlled by the aggregation procedure.
We shall now show that the criterion of a minimum total input heterogeneity,
given the number of industry and primary sets which one decides to choose,
is approximately equivalent to the minimization of a weighted sum of
variances of precisely the same input coefficient discrepancies. After this has
been shown we shall be able to formulate a more direct link between our
desire for a nearly-zero aggregation bias and the input heterogeneity criterion.
Our starting point is the familiar approximation
iy (h-
Ij
cij log
bj
1
2
(bj - aj)2
2L bj
where the a's and b's are positive numbers satisfying Zctj = Zbj = 1 and log
stands for natural logarithm. We interpret the a's as rgj/Pgh and the b's as
Pj/P.i,, which gives
(p, _ !jj\2
V In. rA ~ 1 V PA-
L r„ pJPj. ~ 2 L pJT.»
j'eS/, j eSh
= 1(EA2 v
2VV LWi pj
jeSh
the unit of measurement here being the nit rather than the bit. [The quadratic
approximation requires, of course, that Sh does not have too much input
heterogeneity with respect to the supplies by Sg.] On combining this result
with (5.11) we find
(8.3)
9=1 jzSh
which means that the input heterogeneity Ih of Sh is (approximately) equal

to a weighted sum of the following G expressions:
V Ed(Ik-r0i)2 g = 1,..., G
Li P.h \P.h P.j)
jeSh
Now p.j/P.h is nothing else than Wj as defined in (2.4). Furthermore, as long

as g takes one of the values 1, ..., M (the case of industry inputs) PgJP.h is
simply Agh, see (2.18), and rgJ/p j is equal to Xau, ieSg = Ig. This proves the
first line of
(8.4) if g ^M
jeSh jelh
W; if g -M =k>0
l
jelh iePk
The second line deals with the case in which the supplying set Sg is a pri¬
mary input set (g — M=k>0). Then Pgh/P.h = Bkh and rgj/p J = Xbij, iePk, see
(3.8).
On combining (8.3) and (8.4) we obtain the following expression for the
total input heterogeneity:
M 1 MM ( Agh — Z aij)
(8.5) X PJ„ »- X
h—1 g= 1 h = 1
X A, X
jelh 1gh
- -
1 M' M (bu-
\
X O
iePk /
+ ^fc=l/i=l
?Z jelh B kh
where use is made of the fact that the ratio P.lJPyh’ which occurs between the
two summation signs of (8.3), is equal to the reciprocal of Agh if g^M, and
to the reciprocal of Bkh if g-M=k>0. Let us now compare (8.5) with the
first-order aggregation bias (Ptf)0 of (2.19), which we divide by the total
output of all industries in order to obtain expressions which are comparable
with the informational measures:
M /
(8.6)
(pA
m X
h= 1
X "y U*
P„ jelh \
I atJ)J fj
lelg
Z1 Xi
i=
Consider the first-order bias (P*f)fc of primary demand, (3.9), in the same
way:
(8-7)
(P*f)*
-m
•y
-= v
x
Z h=l
p
P.h Z
V
jelh
(n
wj[ Bkh
\
- V
Z iePk
l Vj
bu v
/ Xj
i= 1
It is immediately clear that (8.6) and (8.7) have one thing in common with
(8.5), viz., that their covariances involve precisely the same sums of micro-
economic input coefficients (Lo.y, iel6 and Zb^, iePk) as those which occur
in (8.5); and the way in which they occur in the latter equation is as a weighted
sum of variances around their averages (the macroeconomic input coefficients
Agh and Bkh). Minimizing the total input heterogeneity ZPhlh amounts
- approximately - to minimizing this weighted sum of variances. Of course,
this criterion does not guarantee that the bias (8.6) and (8.7) will be closer to
zero than under any other criterion; Table 9.4 provides an instructive
counter-example. Using statistical terms, we can say that each separate
covariance of such a bias is equal to the product of the two standard devi¬
ations and the correlation coefficient. As stated earlier, the aggregation
procedure cannot control the standard deviation offj/xj because the/’s are
exogenous with respect to the input-output model; for the same reason, it
cannot control the correlation coefficient either; but it can control the
standard deviation of Ea^-, ie Ig, and of Eb^, iePk, and by minimizing this, one
forces the covariance to lie in a smaller interval around zero. That is, if p is
the correlation coefficient and ay and ay the two standard deviations, ay
being controlled, then the range of variation of the covariance pay o’2 is a
multiple ay of ( — ay, ay); and the width of this interval (for any given a2) is
minimized when ay is minimized. We may therefore say that the criterion of a
minimum input heterogeneity has a minimax character: It minimizes the
maximum absolute value which the aggregation bias can take.
Some qualifications are in order. We had to approximate the aggregation
bias by the first-order bias before we were able to write it in the convenient
covariance form. We had to approximate the input heterogeneity by its
leading quadratic terms before we were able to write it in the convenient
variance form. Also, the argument presented at the end of the previous para¬
graph was in terms of one single covariance, whereas the total number of
components of the bias of intermediate and primary demand is M+M', each
of them consisting of M covariances. The criterion of a minimum EPhIh is
obviously scalar; it amounts - approximately - to the minimization of a well-
defined weighted sum of the variances of the input coefficient discrepancies,
see (8.5). The question arises whether the weights are sensible from an
economic point of view. A definite answer to this question is of course im¬
possible when nothing is specified about the (exogenous) final demand values;
also, one may want to make the answer dependent on the specific use which
is made of the input-output predictions. But given our ignorance in these
respects, we may decide to minimize the variances of the input coefficient dis¬
crepancies “uniformly.” Specifically, consider the industry input coefficient
discrepancies
(8.8) Agh- Y, Qij jelh

ielg
When Agh is very small, the corresponding Eai} will also be small and so will,
generally, the standard deviation of their discrepancies be. When Agb becomes
larger, the standard deviation will increase, but on the average presumably
less than proportionally. This presumption is related to the sampling theory
of the binomial distribution: If p is the probability, n the size of the random
sample, and p the estimator of p, then the standard error of p is sjp{ 1 —p)/n,
which increases less than proportionally with/?; whenp is small, the standard
error is approximately proportional to yjp and hence the variance to p. If we
apply this idea to the discrepancies (8.8), their variance becomes proportional
to Agh, and a uniform minimization of these variances requires that we divide
the corresponding squares by Agh. This is indeed what the minimization of
ZP.i,I.h as given in (8.5) amounts to, both with respect to industry input co¬
efficients (Agh) and with respect to primary input coefficients (Bkh).
Table 9.5 contains the logarithmic enors (8.1) for all years 1949-1960. It
deals with all thirteen secondary and three primary sector sets as well as with
the two subsets of Food, beverages and tobacco. There are two lines for each
sector set. The upper line is concerned with the case before partial dis¬
aggregation of the Food industry set (14 sector sets), the lower line with the
case after partial disaggregation (13 sets plus 2 subsets). The figure upper left
is 183, which should be multiplied by 10-6; this is the 1949 value of the
logarithmic error for (Food, beverages and tobacco) before partial dis¬
aggregation. The figure below this one, 159 x 1CT6, is the logarithmic error
for Sl after partial disaggregation. The corresponding prediction is obtained
by adding the predictions for the two subsets of S1.
355
o — 80 — 80 it ^ fN 00 80 — Tf r- fN it it f 00 CO
196(
00 oo in it 00 08 CO CO O O co co in n in in
co <o 08 fN <N CO oo co CO
1 y 1 1 1 1
~~
i i 1 7 1 1 i i
Tf O OO CO — 00 CO CO it oo r- — P- CO 00 CO it —• O V8
08 fN Cl 80 in CO it fN CO co fN fN fN 1
in co — <N (N fN (N fN 08 08 55 | |
BEFORE AND AFTER PARTIAL DISAGGREGATION OF THE FOOD INDUSTRY SET
08 — T—
| | | | | |
' 1 1 1
T'- in 08 O 08 08 p- 80 co fN r" r- 80 P- OO 08 08 08 «n it
00 — fN fN fN it -t 08 08 CO CO | |
fN fN fN <N co ro <N fN it 3" fN fN CO CO 1 1
i i 1 1 1 i i [ i i i
in co P~ 08 — fN 80 80 fN it — fN r- fN 00 08 O', 80
p- o o p- <N — fN o o O O o o 1 o o
fN Cl fN (N it Nf CO co co co "t Tt | | | |
1 1 i 1 1 1 1
O 00 00 OS rf fN «n 80 co it CO it it O 08 — o m oo 00 It in 80 fN
SO p- •n >n o o 80 80 o o CO CN oo oo o o
'n 08 in in fN fN 80 80 it it fN fN
Os
80 80 O O P- 80 08 08 80 80 m in 80 80 O 08 00 00 in «n o o
in p- 08 08 o o fN fN 08 08 <n in o o it co •n m
co co 08 co co CO co fN <N fN fN fN fN fN fN "t It
Os 1 1 i i
1 1 1 7 77 i i I 1 1 1
o co p- o it 08 CO Tf OO 08 in co •n — — fN p^ co P^ CO
it it in co 80 80 1C. rj- O 08 80 «n O 08 08 08 it -t fN fN — fN p- r-
<N fN 08 —' fN ~ — fN — r- p* p- 80 t- -f
| | i i 1 1 i i
1 1 1 1 1 1 1 1 1 1 1 1
p- CO — co O —1 00 it fN 08 08 fN — co rj- co — — 80 OO 80 00 —
08 in © p- OO <N <N 80 "n it it 08 08 O O fN fN o o
<N CO CO •n n fN fN It it CO CO CO CO
08 1 i 1 1
"" 77 1 1 1 1
fN o p-~ 80 — — p- in ->t CO CO — 00 80 — oo r- — fN 08 p» »n
IN fN 80 00 80 00 — 08 it co >n *n 80 in in in in in IN | co co CO fN
m r~" r- CO (N CN fN CO CO 1 1
08 —< —
1 l i i i i i i i i 1 1 i i
7
co O (N O CO co CO P~ co n it 80 co it 08 O o o fN co CO It •n — m 80 CO —
— <N it 80 it 80 80 OO 00 p» p- CO CO r- r- CO 00 08 08 08 08 00 oo
fN fN p~ it fN fN 80 80 fN fN p- p- 00 00 r- t— CO CO 08 08
08 i i 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 1
p- co p- o fN fN P~ 00 oo r-~ 00 fN co — 80 CO — o — 08 oo m — in P-~ CO 08 80 08 so
•n m 80 p~ it (N >n it 00 00 it <n — fN in n r- 80 It co 00 oo fN co
co co o fN fN fN fN fN fN co co CO CO co co
08 1 ] l 1 1 1 i l
i i 7 '7 1 1 1 1 1 1
1
08 co 08 fN co oo 08 O »n 08 fN «n O 00 80 08 t}- in r-~ co CO 08 •n o O O Tt — — 08
oc in «n o 08 08 08 O <N co 80 *n o o It CO m in U- C" —O 80 P- fN fN
in in CO CO fN fN fN fN fN fN
106,
fN co
1 1 i 1 1
i i 1 1
MULTIPLIED BY
<D V) fN (N (N fN fN fN fN fN fN CN fN fN fN
-2 " + + + + + + + + + + + + + + + +
it co CO co it co it CO it CO it CO id- co It CO it co It CO it co it CO Tt CO it co It CO
1°
O G
(8.1),
o
a C 1A cd
JO . C O £ Ofl
cd
"C
C/I c t> > u cd u C VI
o 3 G 00
(A CA *2 X) u cd a cd 3
cd
X)
0) 3 O 7 cd ’>
CO <0
>
XJ 6
O
U
<u
JO
3
(A i— cd
o
XJ
C a
3 o
cd
6
0,
Lt
o © VI
O
a
‘5 <D <u l~ a
*d
o
Uh 6 O
Lh
+3 cd £
<u
75
c o
Ih
<D
<U
XI G
L
ID O
a
<D
CD
VI
VI
O
x a JO JO
cd JO ^ a Ih
PH < H cd o s u E 5 H H 5
0
1 0
fN cd It in 80 p^ 00 08 o fN cd it
1 1
CHAPTER 10
INFORMATION MEASURES IN THE ANALYSIS
OF INTERNATIONAL TRADE
10.1. Predicting Trade Flows from Total Exports and Total Imports
The objective of this chapter is the application of information concepts to
the analysis of international trade flows, both for prediction and for the
measurement of concentration. Prediction will be our first concern.1 We
imagine that the world is divided into n regions and write ytj for the flow of
goods and services from the zth region to the jth, measured as a fraction of
world trade. Hence the double sum (i,j= 1, ..., n) of the yi} is equal to 1.
Furthermore, we write
n n
(i.i) u. = Z yu y.j = E u,
for the total exports of the z'th region and the total imports of the yth region,
respectively, both expressed as a fraction of world trade.
Suppose that we know total exports and total imports of each region, and
hence also yL and y p i,j— 1, The question is: Can we predict the flow
from the zth region to the jth, for all pairs (i,j), from these marginal totals?
One very simple answer is
(1.2) hi = yt.y.j
which amounts to the assumption of import-export independence. It means
that the exports from i to j are supposed to be large when i exports much and
j imports much, and that the flow from i to j becomes smaller when either i
exports little or j imports little or both. Although there is undoubtedly some
truth in this assumption, it is clearly of a very approximate nature. Some
region i may export little to some other region j in spite of large values of
1 The analysis described in Sections 10.1 add 10.2 is based on P. Uribe, C. G. de Leeuw
and H. Theil, “The Information Approach to the Prediction of Interregional Trade
Flows,” Report 6507 of the Econometric Institute of the Netherlands School of Economics
(1965).
357
358 INTERNATIONAL TRADE
and y j, simply because their distance is large or because of political troubles.

However, many of these causes have a more or less permanent character,
such as distance and sometimes political troubles. It is therefore conceivable
that one can improve on the prediction method by taking account of the
distribution of flows in some earlier year.
Let us assume, therefore, that we have at our disposal numerical data on
the flow from i to j in some earlier year, for all pairs We write xtj for
this flow when measured as a fraction of that year’s value of world trade, and
(1-3) Xi.= Y,xij x.j = I xtJ

j =1 i=l
for the fractions measuring that year’s total exports of i and total imports
of j, respectively. We shall then usually find that there is no import-export
independence in this year. For some pairs (/, j) we have xij>xix j, which
means that the exports from i to j are above the independence level, for other
pairs Xij<xLxj indicating that trade is below the independence level. Equiva¬
lently, consider the logarithmic ratio
log
xl.x.J
which is the mutual information between the exporting region i and the
importing region j (see Section 2.4). This mutual information is positive,
zero, or negative when the exports from i to j are above, at, or below the
independence level.
When the forces determining the deviations from the independence pattern
may be assumed to be approximately constant over time, a rather obvious
procedure is to predict on the assumption that the mutual information values
do not change from the year of the x’s to that of the y’s. This leads to the
following forecast of yiJ, based on the individual flows of the earlier year and
the marginal totals yL and yy
(1.4)
yi.y.j
Tu - XU
xix.j
These forecasts should add up to 1 when summed over i and j. Actually,

however, they do not satisfy this constraint.1 But this defect is remedied
1 Another constraint that may be violated is Take, for example, *«/==.3,

Xi. =x.j = .4,yi. = y.j = .8. Theny'tj — 1.2. Such violations will be rare, however, except
when the changes from the year of the x’s to that of the /s are very large. It is assumed
here that they do not occur.
10.1 PREDICTION OF TRADE FLOWS 359
easily when we divide all forecasts by their sum:
(1.5) n n
Z Z Vhk
Even after this adjustment the forecasts are not really satisfactory, because
the sum over j of y'-j- is in general not equal to the given value yu, nor is the
sum over i equal to y j. An additional adjustment is therefore necessary.
The formulation of this second adjustment should be made dependent on
the criterion of the quality of the forecasts. A rather obvious criterion is the
information inaccuracy:
(i.6) i=H yu log ~

i=ij=i yu
since both the predictions ytJ and the quantities predicted ytj are nonnegative
and add up to 1. Our task is to adjust the forecasts y-'- such that they satisfy
the marginal constraints; clearly, we want to change the y-j as little as possi¬
ble, because we prefer to retain the underlying idea of constant mutual in¬
formation values. If our measure for the difference between two sets of
probabilities is the information inaccuracy as in (1.6), the obvious adjustment
procedure is that of minimizing
t!
lu
(1.7) I Z yh los
i=l7 = 1 yu
subject to
n
Z 9ij = yt. i = 1,..., n

7=1
(1-8) n
Z yu = y.j
i=1
j = i,-,n
The ytj thus derived may be called two-stage information forecasts. The first
stage is based on the assumption of constant mutual information values.
This leads to the yk of (1.4), followed by the yk of (1.5) after the proportional
correction. The second stage consists of minimizing, subject to the marginal
constraints (1.8), the information inaccuracy (1.7) in which y". plays the role
of the “observed” value and y;; that of the prediction.
To solve the minimization problem we construct the Lagrangian expression
n n y'' n / n \ l \
Z 7Z= 1 y"u lo§ fij

1=1 ir - I M I h- - yu) - Z /m Z 9u ~ y.j)
i=1 \7 = 1 J 7 = 1 \i=l /
We then differentiate with respect to yu (under the assumption that log
stands for natural logarithm) and equate the result to zero:
The solution of this system of equations (with the unknown yu in the de¬
nominator) is rather awkward. The procedure is simplified when we replace
the information inaccuracy (1.7) by its leading quadratic term:
n n
(1.9)
The approximation error is small when the y'(j violate the marginal con¬
straints to a limited extent only (which is usually the case). If we then form a
similar Lagrangian expression, differentiate with respect to ytj, and equate
the derivative to zero, we find
(1.10)
which is solved more easily.1
10.2. An Application to the Trade of Eight Regions

In the previous section we developed two prediction methods. One is the
import-export independence prediction (1.2), which is quite naive and whose
performance should therefore be regarded as a lower limit to that of more
sophisticated methods. The second is the two-stage information forecast,
amended according to the quadratic approximation (1.9). We shall now apply
both procedures to import-export data published by J. Waelbroeck for the
years 1938, 1948, 1951-52 and 1959-60.2 These data are based on a division
of the world into eight regions; they are presented (in the form of the
fractions ytj) in Table 10.1. There are two zeros on the diagonal; they refer to
the intra-regional trade of regions which consist of only one country.
1 A straightforward procedure is the Stephan method. See, for example, W. E. Deming,

Statistical Adjustment of Data (New York: John Wiley and Sons, Inc., 1944), pp. 121-124.
2 See J. Waelbroeck, “Une nouvelle methode d’analyse des matrices d’echanges inter-
nationaux,” Cahiers economiques de Bruxelles, No. 21 (1964), pp. 93-114. Note that in the
analysis described here Japan has been allocated to the Rest of the world, whereas
Waelbroeck takes it as a separate region. This is motivated by the zero flow from Other
E.E.C. countries to Japan in 1938, followed by positive flows in later years, which would
have led to infinite inaccuracy values. Note further that Germany should be interpreted
as Western Germany after the War, and that Yugoslavia has been allocated to the Rest
of the world, not to the Communist countries.
10.2 AN APPLICATION TO THE TRADE OF EIGHT REGIONS 361
cd «- £ Tt-ONM. D-nnioaN
r^CNNOONCOTj-’—rt
no on oo’ © © on o
m — — -h in o
— MNO'tci'o — m tJ-
in to o no nj-o <n
NO (N CO NO ' «(N ro Nt -H oi Is- r-
intô — oooroô rn
IMPORT AND EXPORT SHARES IN PERCENTAGE FORM: 1938, 1948, 1951-52, 1959-60
o\i^(N»o — ONor- t- ©'<tooinr--r'-a\«—

t^-oncN'o'ONq rs © O © (N © co on on —_ -H rj- ON CN CO no o ON
.<N — NO Tt —< .(N —■
lO(NXlONONOO\CO ON
(N't(Nin-CONDO <N n no o coo'll-
- T}- t"- rf C
■> NocNr- 00 t"« (N CN C
- O — ~ NO <N « * ’ O«
o
NO
OsI
lo
Os
TfiOO'tOsP'OO'O Tf
tXJIOlOOOhCNj-O
— (N —< — — <N CN CO —« *— — — <N — CN CN i—i eo co
ON CD
' — o —
-ooOP'Co^-ioM in
© lO NO NO CO CO —,NO CO
CO.NO
ONNOOOIOOM- CO
O Nt n ON ON T) — ON <N
NO Nf.CO K to CO -
a>
4>
SG .2co
<D +-»
3 -r G <d *C C <D *
O -3 c o ’C G O£
S ° S- c2 O3 GG °. GP 2~ c2
3 u
r • 3 l» 3 u
oo . -a0 < oo o OO O£ °2< O °

• 2, • °
o S . T3 oo o£
<D (D S gH „ - ij a>H t« <D U. £PH
C . w(/, aj c 4>
U g>H 1/5 <D
c* (D
1g»W2*i ’£ •£ . ft, ■= -C
: >n W'GI
5 £ >>•-. J s SW*M C *5
< c u pj G<~ << 3 : £w W
S c S S 2
£o
£ to ■Sc e s.«i s SStsO i e (D . u S ° •C £ o
r c_ aS Ut, z.U i>N SS
fa 3 !- x: •= xi O D
5s«£5’__ os »£ c£ O <D OS ut ct! o 1u) —
ZbOOPOUb ZbOODO UC* Z jOODOUbC ZjOODO UCZ
- cj CO rf loi NO CO ^(NmTj-in^r'W —■ (N CO Tf IT) NO h- CO - n co rf <n no t" co
The naive prediction method (1.2) requires only data of the prediction
year. On comparing (1.2) and (1.6) we see that the information inaccuracy of
this method is simply the expected mutual information of the import-export
pattern. The values in the four years are as follows:
1938 .2061 bit

1948 .2540 bit
1951-52 .3479 bit
1959-60 .3532 bit
The figures increase monotonically, which implies a trend away from inde¬
pendence.
As stated above, one should expect smaller inaccuracy values when a two-
stage information forecast is used. This is pursued in Table 10.2, which
contains these values for each year with each of the earlier years as a base.
The figures are all less than the independence values shown above; in the case
of 1959-60 with 1951-52 as base the reduction is even larger than 90 per cent.
TABLE 10.2
INFORMATION INACCURACY VALUES (IN BITS) OF TWO-STAGE INFORMATION FORECASTS
Year to be predicted
Base year
1948 1951-52 1959-60
1938 .0995 .1553 .1701

1948 .0906 .1479
1951-52 .0338
The inaccuracy values which we discussed until now refer to the prediction
of all individual shares yu of world trade. We may also be interested in the
destination distribution of the exports from a given region i,yiJ/yi_,j= 1,n,
or in the origin distribution of the imports by a given region j,yijly,J. i= 1,..., n.
The corresponding inaccuracy values are
n
i=l
We have yijlyi=yj and yijjy]=yi_ in the case of the independence pre-

10.2 AN APPLICATION TO THE TRADE OF EIGHT REGIONS 363
diction (1.2), so that the logarithms in the two formulas of (2.1) are both the
logarithm of y^/y^yj. Hence J, is then the average of the mutual infor¬
mation values involving region i as exporter, and Ij is the average of those
values which involve the/'1 region as importer. In the case of the two-stage
information forecast the two logarithms in (2.1) can also be simplified; we
can write them as log (y^/ytj) because yt,=yL, y.j—y.j-
The information inaccuracies IL for the export shares and I j for the import
shares are given in Table 10.3. In general terms they confirm the overall
picture of Table 10.2. For example, the two-stage information forecasts are
better than the corresponding independence forecasts with only 6 exceptions
out of 96. These exceptions all refer to the case in which the early postwar
year 1948 is either predicted or used as a base, and most of them deal with
Germany between its defeat and its economic recovery. The destination and
origin forecasts for 1951-52 are generally better when 1948 is used as a base
than when 1938 is used instead. Disregarding Germany, we find that this
applies to 11 cases out of 14. Similarly, the forecasts for 1959-60 are better
when the 1948 base is replaced by 1951-52; there are only two exceptions to
this rule.
We conclude this section with three remarks:
(1) The two-stage information prediction procedure is rather close to the
so-called RAS method, which was developed by Stone and Brown for the
adjustment of input coefficient matrices in input-output analysis. This re¬
lationship is considered in more detail in the Appendix of this chapter
(Section 10.A).
(2) The development over time of the individual mutual information values
log
yu
yi.y.j
is in many cases also instructive. Take, for example, the Communist countries
(region 7), which have substantial J7 and 17 values according to Table 10.3.
If we substitute i=j=7 in the mutual information formula, we obtain the
following figures:
1938 .619 bit
1948 2.663 bits
1951-52 2.973 bits
1959-60 2.408 bits
These values are all positive, which indicates the obvious fact that the intra-
Communist trade is above the independence level. Moreover, the postwar
figures are substantially larger than that of 1938, thus indicating that there
' ^ "d r- oo
m r-
(S(N
m
«n
nt ON E- fN
ON On on On
vo vo
00o
Tf 04 T}* T—1 00 VO
coOroh
<D
o
£
On VO
t-h O
rn NO
r-H O
VO
O oooo
ON Tf ON O O m
t-h o
co ro 7
OOO
vo ro n n
O O T-H o
d O VO on on on vo vo on t-h -h E'¬ 00 r- on ON -H (N VO
d O r-- r- r- on HOOCOm EN m Is* vo m -h r- \t
E
O on
oo
on on (N
■rf On t-h
on on h n
Hh-HO
-h (N
OO Tfr
CO
I"""
04 Ol
T-H 04
On t-h on t-h
t-H 00 O t-h
E
INFORMATION INACCURACY VALUES (IN BITS) FOR EXPORT SHARE AND IMPORT SHARE PREDICTIONS
• 00
0) r*
43 E* ■£U. r- o
on VO
O
on
vo
On
on
vo
oo -H CO (N
Tf O vo to
^ o
on 00
on OO On
M OO
04 00 on On
On VO On t}-
E" On fN on on Monoort on tj- 'tori 00 on rf 04
^ O <N O O <N r-H O O 04 o m t-h o m h (O (O
ofc. §
W o
o
§
T3 g
oo
^h 7 00 *o
I T
’"h c£ 04 01
VO *0 —1
On O on On <N tJ- MmMVO N ^tOoo On (N on
£-§ on on
on on
X}
On
on rt
on r- rj-
•O VO 00 t On
On
^ *0 t-h r- on
On
on vo 00 m
C “ O "-i Tt o o ’-HTt rf i"- <n
on O O O
t. t-h on
> ^ O
On 04 on on
cn O O *hh MhhH
mOOO
D •-
<3 9
-s:
05 9 9 C
cn 9
u.n- « "ST
CD w s- O m 05 T-H O OO Co ^h r- r- 00 > OO ON <N Tfr VO 05 on ON VO ON
•3
w e
on
on
vo
<N
r- Tt m
on (N CN
on on On m
MN CO h
v on th
g O o-
On O vo
Nt d n-
O 00
CO T-H OO TH
tj*
Ow § £ — O o ÔO O ^ o o o T-H T—> O HH T- O O
o h5 &
I
TABLE 10.3
3
m Tj- VO t-h tJ" O on On
00 M vo vo t On00 on (N —1 vo
c m
on
on
r-
Q Tfr hH
o <n r-
vo m
co m on
rtf-
n
OO 04
Tt —
on r-m hOhND
Nf- cc n
on r- O
s O vo on <N on m (N on O o fNj t-H NO m — vo O
»H
<u
o
cj
c.a CO T|- 00 04 ON fO O on On rt hh m t-H <N r- t-h
h 7 ^
r-
m On on Tfr —h t* O On ^ t— VO o 00 ON o
s « on <N T- 4 M H-'MHt vO vo 00 vo r-- 00t}- rj-
04 O O O O O O Tt O
on
O T-H m O cn O
ni
43 O O' T-H O- m (N 7- on 00 h- m O 't M vo
'C J- vo m in n- l- T+ _ VO <N
on on
T—< <N — ON ON ON O vo m m (N
O a on (N on OOnOOro § <N Tf <^n
T-H O C4 t— O NhhO (N O mOO
Nt00 h H
m O O O
2!
”0
o
X
£ <N (N
d on on
1
o 00 OO OO OO OO 00 00 00
1
m m rj- OOOOh
m on m m tcj- m rt on
ON ON ON ON ON ON ON ON ON
T-H T-H T-H
ON (On ON
<D <D <d a> (U <D CD <L>
C/) C/D C/5 CA CS> C/5
0) u <D <D (U
C/5 C/5 C/5 C/5 C/5 C/5
c3 aj d d d d d d d d cj d
X) JD X) X) X X X X X XXX
/—s
^S' -s O ^S'
(Nj r-H (N H <N T-H (N H CN T-H (N T—H
10.3 concentration: origin, destination, composition 365
was an additional concentration of trade within the group. These features are
not unknown, of course, but it is interesting to see how they can be measured
quantitatively. We shall make use of such information values on a large scale
in the last section of this chapter, where the development of the European
Economic Community is considered.
(3) The two-stage information prediction procedure has a much wider
range of application than foreign trade alone. Take an arbitrary array of non¬
negative fractions, xiJk, the triple sum of which is 1. Suppose that the corre¬
sponding values yijk in some later period are to be predicted, given the
marginal values y{ , yj, y k. The first stage leads to the preliminary forecast
Vi ik Î ik
after which this is adjusted to y"jk such that the triple sum is 1. The second
stage amounts to the minimization of a chi-square, which leads to equations
of the type (1.10) with an additional Lagrangian multiplier vk.
10.3. Concentration with Respect to Origin, Destination and Composition for

Separate Countries and Commodity Groups
In this and the remaining sections of this chapter we shall be concerned

with a three-dimensional problem in international trade. Specifically, the
subject is the development in the years 1961-1963 of the export-import flows
of 183 commodities among eight countries. The countries are listed in
Table 10.4. The X’s and F’s are their symbols; the X’s will be used when the
TABLE 10.4
EIGHT COUNTRIES AND THE PERCENTAGE OF TOTAL EXPORTS COVERED
Percentage of exports covered

Country Symbol
1961 1962 1963
Belgium* Xi, Yi 69.1 72.4 75.9

Netherlands x2, y2 58.1 60.7 67.6
Germany Xs, Ys 43.7 45.9 49.0
France x4, y4 45.2 48.1 49.0
Italy Xs, Y5 48.3 51.3 51.3
United Kingdom Xe, Y6 30.4 32.6 32.9
Canada x7, y7 78.8 81.0 77.8
U.S.A. X8, YS 39.7 39.2 39.6
* Belgium-Luxembourg Economic Union (B.L.E.U.).

TABLE 10.5
DESCRIPTION OF TEN COMMODITY SETS
Number Percentage
Symbol Description of share
commodities in 1961
Si Food and live animals 33 11.4

S2 Beverages and tobacco 4 2.4
s3 Crude materials (inedible) except fuels 29 12.9
S4 Mineral fuels, lubricants and related materials 5 5.4
S5 Animal and vegetable oils and fats 4 .5
S6 Chemicals 16 7.1
S7 Manufactured goods classified chiefly by material 50 25.6
S8 Machinery and transport equipment 18 26.3
So Miscellaneous manufactured articles 18 7.0
S10 Commodities and transactions not classified 6 1.3
according to kind
country is the exporter, the F’s when the country is the importer. Since we
consider only the trade among these countries, the data to be used cover only
part of their total exports and total imports. The last three columns of the
table contain the percentage of total exports covered. These percentages range
from about 30 to 80; they are on the low side for the United Kingdom and
the United States, which is not surprising in view of their important economic
ties with countries outside the group of eight considered here.
The 183 commodities are three-digit code groups used by the O.E.C.D.
(Organisation for Economic Co-operation and Development). We shall
indicate them by Zk, k — 1, ..., 183. In some of our explorations these three-
digit groups will be aggregated to 10 one-digit groups, to be indicated by
Sg, g = 1, ..., 10. These groups are listed in Table 10.5, together with the
number of three-digit groups in each one-digit group as well as the percentage
share of the one-digit groups in the total trade among the eight countries in
1961. To simplify the terminology we shall speak about commodities and
commodity sets when discussing the Zk and the Sg, respectively. For further
details we refer to the Appendix to this chapter (Section 10.B).
The following notation will be used. We write pijk for the exports of
commodity Zk from country Xt to country Yj, measured as a fraction of the
total trade among all eight countries in all 183 commodities. Hence the pijk
are all nonnegative and their triple sum (i, j= 1, ..., 8; k= 1, ..., 183) is
equal to 1. Summations are indicated by dots, so that
183 8 183 8 8
Pij. = Z Pijk P.j. = Z Z Pijk P.. =

P..k Z Z Pijk
i= 1 j= 1
are, respectively, the total exports from Xt to Yj, the total imports of Yj, and
the total trade in Zk, all measured as fractions of the eight countries’ total
mutual trade. Aggregation (partial summation) is indicated by a superscript.
For example, we shall write
8 8
Pijg Z Pijk P -g Z P- k Z Z Z Pijk
for the exports of commodity set Sg from Xt to Yj and for the total trade
in Sg, respectively (again, measured as fractions of the value of aggregate
trade). The superscript 3 serves to indicate that the third subscript (g) has an
aggregative character. Thus, in this notation the last column of Table 10.5
specifies the 1961 values of 100p3g, g = 1, ..., 10.
Our first objective is to measure the concentration of international trade
for each individual country. This problem is similar to that of the industrial
concentration which we considered in Sections 8.1 through 8.3; in fact, we
shall use the same entropy measure. But there is a difference to the extent
that the present concentration problem has a two-dimensional nature for
each country. Consider any exporting country Xt; then there is concentration
with respect to customers (importing countries Yx, ..., Y8) and also with
respect to commodities. For example, it may be that a country’s exports are
heavily concentrated on one commodity but quite diversified with respect to
its customers because this commodity is sold everywhere. Similarly, there is
concentration with respect to suppliers and commodities in the case of any
country’s imports, and with respect to suppliers and customers for any
particular commodity.
Specifically, consider any exporting country XL and its concentration with
respect to customers. Its total exports are measured by pt , and the pro¬
portions going to the various customers arePnJpi...,Pis./Pi..- The entropy
of this distribution is
8
(3.1) Hx,(Y)= y
Pi.. Pij.
which is our inverse measure of concentration. For the concentration with

respect to commodities we consider pLl/Pi._, ■■■,Pi.mlPi.., which are the
commodity shares of Xt’s export basket. The inverse concentration measure

is the entropy of this distribution:
183
(3.2) H*(Z)= )
L-i Pi.. Pi.k
fc = 1
For importing countries Yj we can proceed in a completely analogous

manner. We have 8
(3.3)
i= 1
for the entropy of the distribution over suppliers, and
183
(3.4) Hr,(Z) = V
Li P-j.
log
Pj.
P.jk
k= 1
for the entropy of the distribution over commodities.

The entropies (3.1) through (3.4) are shown in Table 10.6. With the
exception of Canada all export entropies by destination vary between 2.3 and
2.7 bits; since the maximum value is log 8 = 3 bits, this range corresponds to a
lower limit of almost 80 per cent of the maximum and an upper limit of
90 per cent. The Canadian values are much lower, which indicates the relative
insignificance of the Continental European market for that country’s exports.
The Canadian import entropy by origin is even smaller, and it is also de¬
clining. For the United Kingdom and the United States, too, the import
entropies (by origin) are below the corresponding export entropies (by
destination). For these three countries, therefore, the imports exceed the
exports with respect to supplier-customer concentration. We have the oppo¬
site picture in the case of Belgium, the Netherlands and Germany, while
France and Italy do not show a consistent pattern in all three years.
The export and import entropies by commodities show a more uniform
picture. The latter exceed the former in most cases. This reflects the tendency
of individual countries to specialize in their production for exports but to
diversify their demand for imports, which leads to a positive excess of the
export concentration over the import concentration.1 Nevertheless, there are
two exceptions: France and the United States. Their export concentration
over commodities is less than the corresponding import concentration.
1 See, for example, M. Michaely, “Concentration of Exports and Imports: An Inter¬

national Comparison,” The Economic Journal, Vol. 68 (1958), pp. 722-736.
ENTROPY VALUES FOR INDIVIDUAL COUNTRIES
We obtain additional insight by applying the aggregation of commodities

to commodity sets. Consider the entropy defined in equation (3.2):
10
^ pt y Puc *
log Pi-9 + log Pi;
J
■ 1
Pi.. L Pi.g
keSg
Pi.k Pi.gi
10 10
= yPLa VPu ■k 1
3~
Pig , V
log-+ ) - log-3"
1 Pi-
L Pi.. L pi
g= 1 keSg
\g Pi.k L Pi..
g= 1
Pi g
This can be written as
(3-5) HXi(Z) = HXt(S) + HXiS(Z)
where HX.(S) is X;’s export entropy by commodity sets:

10
Pi.g . Pi..
(3.6) HXi(S) — log T
Pi: Pi.g
9=1
while HX.S(Z), to be written as
10
(3-7) HX.S(Z) — V
U Pi..
—'9Hx.Sg(Z)
g= i
is the total entropy of X/s exports over commodities within sets. It is a
weighted average of individual within-set entropies:
Pi.k , Pi.g
(3.8) ^WZ)= ) — log""*
Pi.g Pi.k
keS „
For imports we have an analogous decomposition:
(3.9) Hy.(Z) = Hrj(S) + ff,jS(Z)
where 10
H,(S) = ^log^
P.i P.j,
9= 1
10
(3.10) *wz)=y
Lj
^^s,(z) P.j.
9= 1
HfjS>(z)= y
Li P-jg P.jk
keS„
The right-hand terms of the decompositions (3.5) and (3.9) are shown in
the last twelve lines of Table 10.6. It turns out that the export entropies over
commodity sets vary between slightly more than 2.2 bits and about 2.8 bits.
The maximum value is log 10 = 3.32 bits; hence the lower limit is just below
70 per cent and the upper limit about 85 per cent of the maximum. This
percentage range is about the same as that of the export entropies over the
individual commodities, for which the maximum is log 183 = 7.52 bits. For
the import entropies the percentage ranges are from about 80 to almost 90,
which is higher. When considering commodity sets rather than individual
commodities, we find that the export concentration exceeds the import
concentration in a majority of the cases, but there are several exceptions: the
Netherlands, France, United States, and Italy in 1962 and 1963. When
considering commodities within sets (see the last six lines of the table), we
find that the rule holds except for Belgium, France, and the United States in
1962. We recall in this connection that France and the United States are the
only exceptions to this general rule for the total concentration by commodities
[the left-hand sides of (3.5) and (3.9)].
There is also the concentration problem of individual commodities with
respect to suppliers and customers. The corresponding entropies are
8
) —log —
L p..k Pi.k
i= 1
8
yîog^
Hz,(Y) =
u P..k P.jk
j= i
Since there are as many as 183 commodities, this would lead to a very large
number of entropy values. We shall therefore confine ourselves to a more
limited goal by concentrating on commodity sets rather than individual
commodities. The entropies are then
ws,(M =
(3.12)
Hs.m =
which are inverse concentration measures with respect to origin and desti-
nation, respectively, of commodity set Sg. These measures are presented in

Table 10.7. It turns out that there is more concentration with respect to
origin than with respect to destination, which could be expected. The
maximum entropy value is log 8 = 3 bits and the largest observed values in the
case of destination exceed 2.99 bits (Chemicals, S6); 11 values (out of 30)
exceed 2.9 bits, 20 are 2.8 bits or higher. The origin entropies, on the other
hand, are less than 2.8 bits in all cases except five. There is one commodity
set (Beverages and tobacco, S2) for which the origin entropy exceeds the
corresponding destination entropy consistently; for one other set (Manu¬
factured goods classified chiefly by material, S7) the pattern is different in
different years. As a whole, however, the picture is very regular.
10.4. Aggregative Measures of Concentration
In the preceding section we measured concentration for a particular

exporting country, for a particular importing country, and for a particular
commodity set. There exists a natural way to combine such measures.
Consider, for example, the composition entropy of X/s exports as defined in
(3.2). If we take the weighted sum of these entropies with the corresponding
export shares as weights, we obtain
8 8 183
p,hXi(z)=y P,, y tog

Li L pl. Pi.k
=H (z)
i= 1 i= 1 k= 1
8 183 8
Pi.k log — - ) Pi., log 1 = H(X,Z)-H(X)

Pi.k Li Pi..
i=1k= 1 i=l
In other words, this weighted sum of the entropies over composition of

individual exports is nothing else than the average conditional entropy of Z
given X, that is, HX(Z), which is equal to the excess of the two-dimensional
entropy H(X,Z) over the one-dimensional entropy H(X). This is entirely in
accordance with the conditional entropy definition given in Section 3.1. We
can regard HX(Z) as an inverse concentration measure for the exports of all
eight countries with respect to their composition. In the same way we can
weight the entropy values (3.4):
8
I P.j.HYj(Z) = HY(Z) = H(Y,Z)-H(Y)

j= i
This is an inverse concentration measure of all imports with respect to
10.4 AGGREGATIVE MEASURES OF CONCENTRATION 373
W
-J
m
<
H
composition. Consider also:

8 8 183
HXy(Z) = H(X, Y,Z) - H(X, Y) = ^ ^ ^ pijk log ^

i=l j = 1 k = 1
8 8 183 8 8
PiJ.
V PiJk ,
)
Pi J-
n- lQg - = Pij.HXiYj(Z)
Z_J Pij. Pijk
i = 1J=1 fc = 1 i= 1 j= 1
This is the weighted average (with weights j^y.) of the entropies over compo¬
sition of the individual flows from a particular country to a particular
country. We know from Section 3.3 [see equation (3.2) of that section] that
Hxy{Z) is always smaller than HX(Z) and Hy(Z), at least not larger. This is
easy to understand in the present context. In the case of Hxy(Z) we take both
the exporter and the importer as given; it stands to reason that there is then,
on the average, less uncertainty (more concentration) with respect to the
composition of the basket.
These derivations show that there are several composition entropies for
the trade among the eight countries as a whole, starting with the un¬
conditional entropy H(Z), followed by the single-conditioned entropies
HX(Z) and Hy(Z), and concluded by the double-conditioned entropy Hxy(Z).
It is instructive to consider this numerically for one year, 1961 say, for which
purpose we present the figures in the following array:
log 183 H(Z) HX(Z) Hy{Z) Hxy(Z)

7.516 6.673 5.966 6.292 5.559
100 88.8 79.4 83.7 74.0
100 89.4 94.3 83.3
The first column contains log 183 = 7.516 bits, which is the maximum
entropy value in view of the number of commodities. The observed un¬
conditional value H(Z) is 11.2 per cent lower (see the second line, where the
maximum is put equal to 100). This entropy measures the dispersion of
aggregate trade with respect to composition irrespective of origin and desti¬
nation. Next is HX(Z), which measures the dispersion with respect to
composition, given the origin but irrespective of the destination. Its value is
10.6 per cent below the unconditional H(Z) \ see the third line. This difference
is a measure for the average gain in knowledge as to the composition when
we know the origin of the export baskets. We then have Hy(Z), which per¬
forms the same service except that origin and destination are to be inter-
10.4 AGGREGATIVE MEASURES OF CONCENTRATION 375
changed. The figures show that knowledge of the destination is on the average
less useful than knowledge of the origin, which is not really surprising in
view of the results obtained in the previous section. Finally, we have HXY(Z)
measuring the composition dispersion given both origin and destination.
The uncertainty reduction from the level of H(Z) is then 16.7 per cent.
The various unconditional and average conditional entropies with respect
to composition are given in the upper half of Table 10.8, both for the
composition in terms of the 183 individual commodities and for that in
terms of the ten commodity sets. The commodity set entropies are obtained
in a completely analogous manner. For example:
8 10 8
1
Hx(S) = H(X,S)-H{X) Pi.q 3 Pi., log
Pig Pi..
i=19= 1 i- 1
8 10 8
Pi..HXi(S)
i= 1 9=1 £= 1
see (3.6). The results show that, both with respect to commodities and with
respect to commodity sets, knowledge of origin is more informative than
knowledge of destination. The unconditional entropy over sets is relatively
smaller than that over individual commodities. On the other hand, the
additional uncertainty reduction obtained by knowledge of origin or desti¬
nation or both is comparatively small for sets, so that the double-conditioned
entropies Hxy(Z) and HXY(S) are both about 75 per cent of the corre¬
sponding maximum. These “relative” statements are all based on the last two
columns of each array, which contain the entropies in percentage form
averaged over the three years.
The lower half of Table 10.8 deals with origin and destination entropies in
an analogous manner. The maximum is always log 8 = 3 bits. This is followed
by the unconditional entropies, H{X) and H(Y), and by three single-
conditioned figures. There are three such figures, not two, because the
commodity specification may be either in terms of individual commodities or
in terms of commodity sets. In the former case we have
183 8
HZ(X) =
I
k= 1
P..k
i= 1
Puk
P..k
log
P^k
Pi.k
376
CONDITIONAL AND UNCONDITIONAL ENTROPIES

INTERNATIONAL TRADE
10.5 THE COMMON MARKET VERSUS THE REST 377
and in the latter:

10 8
9=1 1=1
It stands to reason that HZ(X)^HS(X), because the Z specification is more

detailed than the S specification. In the same way we have HYZ(X)^HYS(X),
which deal with double-conditioned entropies that are defined in a straight¬
forward manner. The figures indicate that the origin concentration exceeds
the destination concentration systematically. In particular, the double-
conditioned origin entropy HYZ(X) is very low; it is almost 50 per cent less
than the maximum value.
10.5. Two Country Sets: The Common Market and the Rest
An obvious question is to what extent the import-export data of the eight

countries enable us to describe the development of the Common Market in
informational terms. It should be stressed at the outset that the analysis
which follows gives only a partial answer, since a more complete analysis
should include the development of imports and exports during years prior to
1961. The method is therefore more important than the results, although
these results are not without interest either.
We shall need some additional symbols indicating aggregation over
countries within country sets. There are two such sets: the European Economic
Community (E.E.C.) and the Rest. The former will be indicated by Qx or Ri,
depending on its role as exporter or importer, and it consists of the first five
countries: Xx, ..., X5 or 7l5 ..., Y5. The Rest consists of the last three
countries (X6, X7, X8 or Y6, Y7, Y8) and is indicated by Q2 or R2. We follow
the superscript notation for aggregation over countries also. For example,
183
plfg = Z I
ieQrkeSg
Pijk Pis2= X X X
isQrj^Rsk— 1
Pijk
are, respectively, the exports of commodity set Sg from the countries of the
country set Qr to country j and the total exports from the countries of
country set Qr to those of country set Rs, both measured as fractions of the
aggregate trade among all eight countries. The flows p]2 are shown in
Table 10.9. They indicate that the Common Market share of exports in¬
creased from 50.3 to 53.8 per cent, that its import share increased from 54.7
to 58.4 per cent, and that the share of the mutual trade of the E.E.C.
countries increased from 36.8 to 41.2 per cent.
TABLE 10.9
SHARES OF TOTAL TRADE OF E.E.C. AND THE REST
Country set
E.E.C. Rest Sum E.E.C. Rest Sum E.E.C. Rest Sum
of origin
1961 1962 1963

E.E.C. .368 .134 .503 .388 .131 .519 .412 .126 .538
Rest .179 .319 .497 .174 .307 .481 .172 .290 .462
Sum .547 .453 1 .563 .437 1 .584 .416 1
We shall now evaluate these developments in informational terms.

Consider the following mutual information values:
Pis2
(5.1) log^-^ r,s = 1,2
Pr..P.s.
These values, which are completely analogous to the measures used for the
intra-Communist trade at the end of Section 10.2 under (2), are shown on the
first line of Table 10.10 for each of the three years. They appear to be positive
whenever r = s, negative whenever r^s, which means that the trade within
each country set is above the independence level, whereas the trade between
the two sets is below that level. For r = s = 2 the successive figures increase
monotonically (indicated by + upper right). Hence the trade among the
three countries of the Rest increases continuously relative to the independence
level. For r^s the figures decrease algebraically in the three successive years
(indicated by - behind the time series). This means that the trade between
the country sets decreases relative to the independence value. It is somewhat
surprising to find that the figures for r = s = 1 also decrease, which implies
that the trade among the Common Market countries is declining in the
direction of the independence level. One should realize, however, that such a
development necessarily takes place when the import and export shares of the
E.E.C. increase more and more. In the limit we havep\\ =p\ =p\ =1, for
which the value of (5.1) for r = 5=l is zero.1
1 The E.E.C. share of the total imports of the E.E.C. increases (see Table 10.9). It is
(.368)/(.547) or 67 per cent in 1961, 69 per cent in 1962, and 71 per cent in 1963. Similarly,
the E.E.C. share of the total E.E.C. exports increases: 73, 75, 77 per cent in the three
successive years. But when we divide the intra-E.E.C. trade by the product of E.E.C.
imports and exports (all divided by the aggregate value of the trade among the eight
countries), the resulting ratio declines over time. In this connecuon we refer to footnote 1
on page 358, which shows that an increase of the marginal probabilities may lead to a
probability of the bivariate distribution which exceeds 1 when the mutual information
value is not changed.
10.5
TABLE 10.10
MUTUAL INFORMATION VALUES (5.1) AND (5.2): E.E.C. VERSUS THE REST
THE COMMON MARKET VERSUS THE REST
379
We can also derive similar mutual information values for the separate
commodity sets:
yrsg 'Ip! 0 r,s = 1,2
(5.2) log
„13 „23 0 1,...,10
Pr.g P.sg =
3 „3
P..g P..g
These, too, are given in Table 10.10. We find that they all have the same sign
pattern as that of all commodities combined; also, that they increase over
time for r = s = 2 with only one exception and that they decrease algebraically
for in a majority of the cases. [The absence of + or — indicates that the
development is not monotonic.] For r=.s = l, dealing with the trade among
the Common Market countries, the picture is more diffuse. It is interesting to
observe that S4 (Mineral fuels, lubricants and related materials) is an ex¬
ceptional commodity set to the extent that its E.E.C.-Rest trade pattern
tends to move in the direction of independence.
Another extension of the mutual information formula (5.1) is in the
direction of the exports from and the imports by individual countries. When
considering total exports from countries to country sets we have the following
mutual information values:
Pis.
(5.3) log i = 1, ...,8; s = 1,2
Pi..P2s.
Similarly, for total imports by countries from country sets:
i
(5-4) log -P- r — 1,2; j = 1, ...,8
Pr..P.j.
These values are presented in Table 10.11. It turns out that the E.E.C.
countries have a perfectly consistent sign pattern: The information values
(5.3) and (5.4) are all positive when the trade is with the E.E.C. group, thus
indicating that this trade (both exports and imports) is above the independ¬
ence level, and the signs are all negative for trade with the Rest group. The
three countries of the latter group are divided. For Canada and the United
States we have a consistent opposite sign pattern; in particular, the Ca-
nadian-E.E.C. mutual information values take substantial negative values
(varying from -2.36 to -3.07 bits). For the United Kingdom, on the other
hand, the picture is more diffuse. Its exports to E.E.C. and to the Rest were
close to the independence level in 1961, but in the later years there was a
movement toward Continental Europe and away from the English-speaking
TABLE 10.11
MUTUAL INFORMATION VALUES (5.3) AND (5.4): E.E.C. VERSUS THE REST
FOR INDIVIDUAL COUNTRIES
Country 1961 1962 1963 1961 1962 1963
, _ P2<2.
.o8 P‘\ !°g „
■p
*».. p .1. P.i.. P .2.
Belgium .495 .481 .455 -.983 -1.025 -1.061

Netherlands .485 .440 .432 — -.950 -.884 -.973
Germany .405 .393 .375 — -.716 -.742 -.778 —
France .442 .446 .417 -.817 -.906 -.919 —
Italy .246 .272 .246 -.367 -.448 -.437
United
Kingdom .005 .053 .076 + -.006 -.071 -.114 —
Canada -2.359 -2.646 -2.703 — .980 1.057 1.131 +
U.S.A. -.352 -.397 -.447 — .335 .388 .459 +
, plj.
log log ■
Pi.P.i. P P
Belgium .579 .534 .495 -.996 — .954 -.932 +

Netherlands .446 .387 .372 — -.658 -.582 -.604
Germany .394 .413 .393 -.551 -.638 -.653 —
France .478 .472 .428 — -.729 -.781 -.741
Italy .188 .229 .268 + -.218 -.296 -.390 —
United
Kingdom -.104 -.104 -.143 — .098 .105 .150
Canada -2.894 -2.947 -3.069 — .906 .956 1.018 +
U.S.A. -.491 -.552 -.587 — .369 .426 .474 _J_
countries on the American continent. The imports, on the other hand, drifted
toward the latter countries.
Finally, we shall consider a combination of the extensions (5.2) and
(5.3)-(5.4) by taking both separate commodity sets and individual countries
(either as exporters or as importers). The formulas are:
, Pfsllp3.g i = l,-,8
(5.5) log
Pi P% s = 1,2
P^.g P^.g 0 = 1,..., 10
13 ,3
r— 1,2
(5.6) logÛ j = 1.8
Pr.g P.jg
3 3
O
CS5
11
P..9 P-9
Table 10.12 contains a summary; the information values (5.5) are given in
the first set of three columns for s= 1 (exports to E.E.C.) and in the second
set for 5= 2 (exports to the Rest group), the values (5.6) in the third set for
r= 1 and in the fourth for r — 2. Our comments are as follows:
(1) The sign pattern is stable in the sense that different signs in different
years of the same mutual information value (5.5) or (5.6) are relatively rare.
For example, we have a change in sign of (5.5) for S4 in the case of France
and also in the case of Italy; but in 153 cases out of 2 x 8 x 10= 160 there is
indeed stability as to the sign. [Note that the sign of (5.5) for 5= 1 is neces¬
sarily opposite to that of s — 2, and similarly for (5.6) with s replaced by r.]
(2) The positive values of (5.5) for the exports of the E.E.C. countries
(/=1, ..., 5) to the E.E.C. (5= 1) apply to all commodity sets in all years with
only 7 exceptions out of 150. The major exception is Italy’s S5. That country’s
exports of Animal and vegetable oils and fats to the E.E.C. is consistently
below the independence level, and hence the exports to the Rest above that
level.
(3) There are more exceptions to the positive sign of (5.6) for the imports
by the E.E.C. countries (j= 1, ..., 5) from the E.E.C. (r = l). The number of
negative signs is 15 out of 150. The Netherlands is a major exception with
respect to St and S2 : Its imports from the E.E.C. of Food and live animals
and of Animal and vegetable oils and fats are almost 1 bit below the inde¬
pendence level.
(4) The values of (5.5) for the exports of the United States to the E.E.C.
are negative except for 9 positive signs out of 30. Beverages and tobacco and
Crude materials (inedible, fuel excepted) are the main exceptions; for these
the exports to the E.E.C. are above the independence level. The values of (5.6)
for the U.S. imports from the E.E.C. are also mostly negative. The number of
exceptions is 7 and the major exception is Animal and vegetable oils and fats.
(5) All Canadian values of (5.5) for its exports to the E.E.C. are negative,
and many of them are substantial in absolute value. Those of (5.6) for imports
from the E.E.C. are negative too except for Beverages and tobacco.
(6) As could be expected, the picture of the United Kingdom is more
diffuse than that of any other country: 17 positive and 13 negative values
for (5.5) in the case of exports to the E.E.C., 15 positive and 15 negative
values for (5.6) in that of imports from the E.E.C.
+1 1 ++ + 1 1 +
f" co -rt <n <n © SO SO m so Tt co — © in ©

-Ou'iN’t OS SO Os cot}mOM co 00
MVI-HON't MOST} I-" Tt <N © - 3 —• © (N
E.E.C. VERSUS THE REST FOR INDIVIDUAL COUNTRIES BY COMMODITY SETS I7II III 1 1
*0
.080
© Tt Tt OO co in ©
.275
— so in U' Tt
.131
os m Tt so »n OS 00 so SO so so M co
(N<OOsM 3 M OD t} SO Tt CN © —
1 171 III 1 1
Tf IT) Tj- O <N e- r- o oomt'-coo O M Os

Os — co 00 © 00 so n NN-NO MOO©
CN Tt Os © © — OO T} CO t} M © M — 0 CN
f • 17 f r r r ’ r * f
1 l + i 1 1 +
3-
119
650
Tt (N SO in — OS OS n so —1 so Tt
.251
vs os - 0 n «n i"- co Ttcoh-O-
(NOs^CMon co 00 r- u- m co 0 in
1 171 1 1 1
(N —1 © co —i os — n ■3- in «n os © Tf SO CO
sot}- — — — © r- so MMMt}- »n © t"
(N os so r^- co Tf f" f" u- in co © m - M SO
0
1 1 7 1 U 1 1 1
0
<3
00 © so Os
.160
-.295
Tt — CO (N co in so 0 -h
-.581
SO so — so © Tt © Tt O t} u u E- in
CN so so so © m in u Os so co — co
*r f 7r
TJ
c
<3 1
a 1 1 ++ S’ + 1
*vso
(N OS (N © CT <N SO —1 Tt — —< © OS co CO (N
Tt — TfSO — 00 00 r— r» Tt 1 u-~r'-~ •n © CN
*01 OS CO so Tt os co 0) so 0 n - r} co E- M
•n ~ 1 1 I
Co 7 1 1T 1 1 1 1 1 1 1 1
TABLE 10.12
Co
.442
.496
r>inos(Noo in co <n CO Tj- t^- so SO
.711
m cN — co so Os n 00 in 0 r} m co
00 © (N so M Os M 0 (N — — so
7 177 1 1
•n—
II
| . |
1
CO © CO © Tt •n so Tt OS CN CO Tt CO os os 00
© OS OO SO Tt © OS SO SO <N © Tt CN Tt (N in
C' so CO Os T) — r- cn tJ- (N — — SO co SO <N
7 1 11 1 1 1
in-
1
.
1
|
1
|
1 1
1 1 + 1 1 + 1
(5.6):
— OS <N © <0 00 (N in — os 00 *n co t"" E- O

© — OO I-" 00 co Os os os — — os COMM
in co co Tt <N M COM M M CO M t} M *n M
71 17
AND
.000
.159
(N © co- Os Os OS T} OS CO Os M
.491
Tt 04 co fN SO M M co co 00 *n 00
(5.5)
in co rf Tf co — Tt (N <N 00 — — *n
—1 — CO
1 1 | |
MUTUAL INFORMATION VALUES
so Tt >n C" © © Tt — — — <N <N (N CO — Tt

h- Tt © <N Os r- so »n t} © N M Os Tt OS SO
incos}t}d © © <N Tt © — CN SO Os m co
71 'T
c
O3
O e a
TJobo T3obo
T3cr> 3 T3Cfl 3
ai 33 '
g s- aj u
3
^« ■
3 >>
S «- 3 o
C J2 3 2
_3 .
3 r- o
*53)5 u 5- “■S< 3 a> C o
b0~ S £ — •a
75CQ- « £ 3 a
PUS
13 0) <U 2 3
coZOfc£ POP
384

INTERNATIONAL TRADE
+ ++ +++ + i i + i + + +
vi u so n co nn so VlOVlOSVl © E'' <n

co (N co oo
so st os Os
st © CN SO st (NO — -tOSMst St — E- OS © CO CO st CN © ©
— 1 ~ 1 1 1
1 1 1 1 1 f 77 1 7 1 1
III
C/)
CO l E'' — OS 00 so so 00 (N st os O O VCO
(N SO co co o so so CO Os — © (N i/O <N OS CO CO
vt O Ol U U (NOSO © SO © CN IVS st — VO OS OS (N CO co CO os ©
6 ~ — — |
o I 1 1 1 1 1 1 1 1 1 1 1
II 1 1
£
wsMxjcon oo o r- •/o © — © Tt © IV) © © oo so co st
(NOsOSSfOO st OS (N OS (N — E" © e- e- © so st r- oo vo
so Tt O «0 <N (N OO O - VS OS M t st © so os 00 — co (N •V) OS —
II 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 + 1 II 1 + 1 1
cn so oo m co VS Is- CO st uo © — so
CO os o so os — 'OON CO OS © SO st e- © (N cn —
SO rl O — m rn d) st st st St (N St V (N M CO — IV) O
Os
ts co |
7 l I 7 I
1 5 1 1
d s*
SO rf IO M SO (N OO •VS © SO Os OS d> — «V) OS CO St
w
<N rmosvs- —< 00 OS M t M 1— SO so oo e- S co — <n so r-
w n O -'t ri (N OS O st vs st st (N st CO so Ci, V) V M M M (N (N ©
.S’ a
e '7 f i 7 i 7 r
O d>
u- <N CN SO —< O vs Tt 00 CO M M vs n St — CO — © r- cn co IV) © CO

— OS (N 00 *0 — © CO VO VO so r- co iv) o cn
VO no-stn CN 00 © t V t V M St <N so IV) IV) — co CN CO co —
Os *5 &
.o ’<N f |'<N I* ’ CO
&5 a
5 l 1 C3 1 1
-c “Q
U i +1 1 1 C
++ ■8 +
a 1 1 1 ++
o
o
Co Os O VO rf i/s in co —* OS co M CO t t CO & co r- — — co
co St *V) CN SO St co — <N t V 00 co co — — e-
Os f' sf n n so n st d>
E- St OS — CO © (N IV) oo r- in — oo — — r-
os
S-
m i i i i’7 i*7 r o 7 l l 7 l
i i 1 1
SJ |
to — f'nos so — oo >2, © 00 »0 E- 00 Os — CO IV) (N st SO iV) CO © CO
CN ~ - h- st Tt i/'s so O S’ M V t t CO VMOS co iv) r- © oo
& so oc st r j st st (N CO C r- st os — <N © — »V) © co st — co
co
1 1 1 1 1 i| "1 |i 7
1 |i | 1 1 | 1
H
st — Ov E- Os st O SO © (N © SO — IV) — CN SO — Ov (N IV)
U t OS O OtN — SO st © st (N so CO 00 VO CO SO IV)
SO vi u st cin I/O (N (N CO OS - M © © IV) SO VO st 00 IV) © os IV)
1 1 II 1 1- 1 “Mil 1
17 17 i
+ 1 + + 1 i i i i 1 1 1 + 1
r- t- so — Tt st IO st © OO (Nr- — CO IV) 00 — st
so — o co r- SO CO CO h- OS — IV) — vo iv) iv) iv) r-
SO nnM-- cn cn co CO V) st st — © 00 VO IV) CO CN st CO © IV) oo
Os
1 7 1
7 f 7 1
d M n -h u- Os (Nr- I"" co — SO CO st (N (N © n os so r- © so co
Os vs st rn M CO — st 00 00 so — SO st SO CO st — so r- — CO M M
w (Nnn-M (N COM CO «n st IT) — © so t" SO CN CN St st © st e-
w Os CO | co 1 (N |
o 1 1 1 1 1 1
ici-sonh' 00 OS 1— oo st vo co — © CO St IV) SO IV) os ©
OO st in (N OS VS OS — OS 00 C- CN t (N IV) iv) •V) 00 OO IV) St © so r-
SO NMN-" M co M CO V st V) <— VO CO CN st CO © St vo
Os
11 7 1 7 1
>>
c
3
o
U ao 6
o
s
o
© © ©
00 00 to
© G © G © G
G K. G ^
cl & 2 fi* G 5 C J2 C u
• 5 u. S « —1 • 3 v- ctf o
1 53 55 8 3 0) c d) ,2«co
‘5b-5 £ Sit* M’S u M3-5 uS —
B§“! 13 o tu 2 p a g^ 13 <j> 2 c g^
»zona DUD nzois DUD wZOfc- DUD
+ + +++ + 11 +
'ÔMTI'OO
<30 *n IS) 1
IN ■
IN VO t-«
© CO oo
© Tf vo o
ro — — Tf VO
is, IN — ©
©—©
-(Nnm- © Tf © © r- (■" (N —* — 00 Tf
IN
77 1 1 1 1 *- 1 7 | '7 1
in in Tf vo oo
(N co vo r" —
oo in in
IN O' I"- l/~> m
r- Tf 00
IS)
© ON
is) 1
—©
r- vo
in ~ in m — CO CO — Tf © CO — IN Tf GO IN
IN
77 1 1 1 1 " 1 | | 17 ■' |
loncooon On co CN ON © <n in oo r- vO —
© OO On IN VO
Tf IN U-) IT) —1 ro 't OO
Tf co IN
— © On ©
OO On ^ ON ©
is i •n Tf
© — 00 00
IN
77 i ' 1 1 - 1 7 1 1 1
sc
i i i 1 + 1 i ++ + 1 1
o
f" On — ro Tf — Tf ON — CO CO Tf © © in
ro Tf Tf on i---
ro co — —• o
© co co
© CO ©
bes: — r» 1 CO
r-~ m <n e- —1 VO — ro
on — r-
<N i3 ^ |
1 o
o
1
CO VO VO © VO © IN © CO ON VO vo Tf 00 Tf ON
vow, — (NW
CO CO <N IN ©
Tf —1
— NT ©
OO r- is,
—* r- r-
r-~ vo ro VO IN
co Tf <n
On On co
<N <4>
'7 r
S-3
«n VO — CT\ ©
~ ON Tf —• 00
© in —
On © co
Tf <N IS)
CO On
l- in ON Tf O
ro f- ro
— VO Tf
Tf ro IN IN © —< wo — 00 VO — vo © ON t— IN
IN o
| sc
so
1 1
l ++ + l + '■S 1 1 ++
<3
On — so in on I"* —« E" cc © E" © vo vo co
E- O ro IN O © <N ON N W N CO N CO Tf ©
NS — IN Tf — © CO VO VO X. © IT) ON I 00 © On 00
■5 — —
1 1 1 1 1
77 l | i “r*
u II
c <3
uo VO © OO © IN
« WOCOVO
— © co — o
I"- — (N
on r~* in
On vo 00 vo ro
oo On Tf ON IN
r- r- r- co in
OO ON ro
© on r-
77 u i
O i i i 17
5
vo «n vo Tf oo © © IN OO m
e^ •- 5 ro Tf — vo in
W
— Nf CJ — — -7 00 VO
d
IS) IS)
r-. no ON © r- 00
-I 77 r r r 1 1 1 1 1
M o
PQ
co
< 11+11 1 ++ ++ 1 1
H
Tf on t"~ t"- r-
OO wooo
— <N -H
CO Tf ©
OO vo OO VO
ON — Tf ON IN
00 ro CO h-
in on is,r- is,
Tf in © ro 00
i i i 1CO ~
1 1
~ CN m © in
00 VO co — 1/-)
On VO ON
© — OO
IN IS)
VO ro
r- ro ©
10
IN <n ON Tf Tf Tf IWN f" © 00r-
1 <N~
'7 1 1 1
(N W OO N - m co ON CO CO On IN VO
ON OO © VO IN IN © <n
—©— Tf Tf m Tf co © *— >n
1<N ~
177 ‘ 1 1
o
U £
C/)
T3o60
X3 _c
C
c«c E-Si &
2
Isas B^< .3 S S 8 -o -g J
m-S s s CO’S £ S '3 S &0
■3 « « 2 a « « «> 2 «j S <3 •
mZOu*; DuD cqÔPh^ DOS
(7) As in Table 10.11 each of the three-element time series is followed

by + or by — except when the development is not algebraically monotonic.
The percentage of nonmonotonic cases (no + or —) is almost 50. In Table
10.11 it is only 20. This difference reflects the increase in erratic behavior
when the object of analysis is more microeconomic in nature. In this case:
when we shift from total trade between countries and country sets to the
trade in particular commodity sets. A still more detailed analysis (com¬
modities rather than commodity sets, or country-to-country trade rather
than country-to-country set trade) would undoubtedly be interesting from
the standpoint of the detail involved, but the decrease in regularity would
almost surely continue.
10.A. The RAS Method for Adjusting Bivariate Frequency Tables
The RAS method,1 when described in the terminology of Section 10.1,

amounts to the following. One multiplies each xu by a number rt which is
specific for the exporting region and by a number Sj which is specific for
the importing region. These numbers are to be chosen such that the marginal
constraints are satisfied. Hence:
(A.l) y*
flj = WjSj i,j = 1, • • n
n
(A.2) £ yi = xusj = yi. i = 1, • ..,n
j= 1 j= i
n
(A.3) tyfj = Sj X riXij = y.j J = 1, • ..,n
i= 1 i= 1
where y* is the RAS forecast of yu. [The word RAS is due to the original use
of the method in input-output analysis; the input coefficients, usually de¬
noted by dij, are the object of adjustment and the adjusted value is of the
form rfiijSj.]
The economic interpretation of the procedure is as follows. When the
share ytJ of the exports from i to j exceeds the earlier value xtJ, this may be
due either to the fact that the exports of i show an overall increase, or to the
fact that the imports of j show a general increase, or to the fact that there is a
special increase of the exports from i to j which is neither related to i’s total
exports nor to y’s total imports. It is assumed that there is no such special
effect, in which case it is reasonable to say that r; measures the effect of the
change in the share of i in total exports (r;> 1 if the change is positive,
r,-< 1 if it is negative), and that Sj measures in an analogous way the effect of
the change iny’s share of total imports.
If we substitute in (A.l) the ratios yJxL and y.j/xj for rt and Sj, re¬
spectively, we find that y* is equal to y'u as defined in (1.4). They are not
1 See R. Stone and A. Brown, A Computable Model of Economic Growth. Volume 1 of

the series “A Programme for Growth” (London: Chapman and Hall, Ltd., 1964).
388
10.B DETAILS ON IMPORT-EXPORT DATA 389
really equal, of course, because y*j does and yb does not satisfy the marginal
constraints. But if we adjust the yb such that these constraints are satisfied, it
stands to reason that the result (yy) will not differ very much from y*-.
Hence the RAS forecast and the two-stage information forecast will be
approximately equivalent as long as the yb of (1.4) do not violate the marginal
constraints too seriously. In fact, there are no appreciable differences between
the results of the two methods when applied to the data of Section 10.2;
reference is made to the article by Uribe, De Leeuw and Theil quoted at the
beginning of this chapter.
10.B. Some Details on the Import-Export Data of Eight Countries
The data analyzed in Sections 10.3 through 10.5 are taken from the
O.E.C.D. Statistical Bulletins on Foreign Trade, Series C: Trade by Commod¬
ities. All flows are measured in units of $1000. The 183 commodities are
those three-digit groups for which the transactions are sufficiently large; all
groups for which transactions exceed $ 10,000 are specified separately, and
the remainder is summarized under one of the three-digit groups of the last
one-digit group (S10).
For 1963 there are no complete data on Dutch exports with respect to
composition. For the purpose of the computations described in the text these
remainder items have been allocated proportionally over all commodities.
The size of these items, expressed as a percentage of Dutch exports to the
seven other countries in 1961, is as follows: Belgium 7.4; Germany 4.8;
France 7.2; Italy 12.4; United Kingdom 9.4; Canada 21.3; U.S.A. 16.3.
CHAPTER 11
CONTINUOUS INFORMATION THEORY
AND A MULTIPLICATIVE DECOMPOSITION OF
PREDICTION ERROR VARIANCES
11.1. The Entropy of a Continuous Distribution; The Case of the Normal

Distribution
When a distribution is discrete with probabilities pu p2, p„, its entropy

is defined as
n 1 n
I P,-log - = - l>logPi
i= 1 Pi i= 1
Suppose now that the distribution is continuous with density function/( ).

Then the role of the probabilities is taken over by the density function
evaluated at all possible values of the argument, and summation is replaced
by integration. The entropy of a continuous distribution is therefore defined as
(1.1) H = - J f (x) log f (x)dx
As an example we take the normal distribution with mean p and variance a2:
1 (x - p)2
(1.2) / (*) = exp
<7 J2li
The logarithm of this density function can be written as the sum of two parts:
... 1 (x — p)2
log/ (x) = - loga^/2n - -— 2
1 G
where log stands for natural logarithm. [We return to natural logs and nits in
this chapter.] Hence the entropy is the sum of
OO
(log (Jyj2n) J f(x)dx \ogOyj2n — 2 log 2tc<j2

— 00
390
11.1 THE ENTROPY OF A CONTINUOUS DISTRIBUTION 391
and
OO
— 00
We thus obtain for the entropy (in nits) of a normal distribution with mean /.i
and variance a2:
(1.3) H = \ \og2neo2
Hence the entropy is one-half of the logarithm of a multiple 2ne of the

variance a2. Equivalently, the entropy is a linear function of the logarithm
of the variance, \ log 2ne being the constant term of this function and \ log a2
the part which is proportional to the logarithm of the variance. It will be
noted that we can equally well say that H is a linear function of the loga¬
rithm of the standard deviation a, because \ log a2 and log o are equal.
We conclude from (1.3) that the entropy is independent of the mean /t. In
fact, this is natural. The entropy is a measure of uncertainty. When we know
that some unknown quantity is a random drawing from a normal distribution
with given mean /< and variance a2, the only source of uncertainty is the mere
fact that there is variance. The larger the variance, the greater the uncertainty;
hence we should indeed expect the entropy to vary positively with o2 and to
be independent of /r. It can be shown that out of all distributions with a given
variance, the normal distribution is the one with the largest entropy. This
statement is subject to the proviso that the range of variation is ( —oo, oo).
When the range has a finite lower and upper limit, the distribution with
maximum entropy is the rectangular distribution over this range. When the
distribution’s range is (0, oo) and when the mean /./ is fixed, the maximum
entropy case is that of the negative exponential distribution, the density
function of which is ixe~xx where a is the reciprocal of ji.1
Although the continuous extension of the discrete entropy measure is
straightforward, there is nevertheless an important difference which ought to
be stressed. A discrete distribution may refer to qualitative characteristics
such as colors, ideas, and authors of books, but it may also deal with
quantitative variables, e.g., the number of accidents on a certain highway per
week. Continuous distributions are evidently more closely related to discrete
distributions of the latter type than to those of the former. Suppose then that
1 For a proof of these statements see, for example, S. Goldman, Information Theory
(New York: Prentice-FIall, Inc., 1953), pp. 127-134.
392 CONTINUOUS INFORMATION THEORY
we deal with such a quantitative distribution and write xux2, ■■■, x„ for the
values taken by the random variable with probabilitiespy,p2, • • ■,pn■ Suppose
further that we change the unit of measurement such that xt becomes kxh
i = l, ..., n, k>0. Clearly, this has no effect on the entropy — Ipt logpt. But
when the distribution is continuous, such a change in the unit of measurement
does have its effect on the entropy (1.1). It is easily seen that in the normal
case the change leads to a new value of H which exceeds the old value by
log k [see (1.3)]. We concluded in the previous paragraph that the dependence
of the entropy on a2 is as we should have expected, but the dependence on
the unit of measurement does represent a difference compared with the dis¬
crete case. On the other hand, when we consider two random variables,
X and Y, both of which are continuously distributed and whose values are of
the same dimension, then the difference of their entropies does not change
when their (common) unit of measurement changes.1 This will be particularly
important for our purposes in this chapter, where we shall be mainly inter¬
ested in entropy differences.
The objective of this chapter is the application of the normal entropy
measure (1.3) to the evaluation of a number of forecasts made by the Nether¬
lands Central Planning Bureau and of certain estimates of recent changes
made by the Netherlands Central Bureau of Statistics. The basic idea is quite
simple. Suppose that the distribution of the prediction error (or estimation
error) is normal. The implication is that the conditional distribution of the
realization, given the forecast (or estimate), is also normal. The entropy of
this conditional distribution is then a natural measure for the uncertainty of
the realization, given the forecast or estimate, or equivalently for the fore¬
cast’s or estimate’s inaccuracy with respect to the realization. The series of
successive forecasts and estimates is described in Section 11.2. A simple
distribution law for their errors is formulated in Section 11.3; its parameters
are estimated and its validity is tested in Section 11.4. The later sections are
devoted to an evaluation of the results in informational terms.
11.2. Predicting the Future and Estimating the Past
Consider any fixed year t. In the year before, t-1, the Central Planning
Bureau computes forecasts of the changes in t of certain macroeconomic
1 See Goldman, loc. cit., pp. 139-140, where the following statement is proved. Let
xi,.... xn be a set of random variables whose joint distribution is continuous; transform
them to another set y\,..., yn\ then the new entropy is equal to the old entropy minus the
expectation of the logarithm of the Jacobian of the transformation.
11.2 PREDICTING THE FUTURE AND ESTIMATING THE PAST 393
variables. During year t these forecasts are revised in the light of new
evidence. When year t has passed the Central Bureau of Statistics takes over
the task of the Planning Bureau; in September of t +1 it releases preliminary
estimates of the changes that took place in t. These estimates are revised in
t + 2, and revised again in / + 3. As a whole, there are seven systematic
attempts to predict or estimate the changes that will take place, or are taking
place, or took place in year t: “will take place” as long as / is future, “are
taking place” when t is the current year, and “took place” when t has passed.
Each of these seven is an attempt to reduce the uncertainty with respect to the
changes in t. Their chronology is shown on the time axis of Figure 11.1. The
/-I t t* 1 t+ 2 t +3
S D S D S S S
12 3 4 5 6 7
Fig. 11.1. Seven successive forecasts and estimates
shaded area above t indicates that this is the year whose changes are predicted
or estimated. The individual stages in the process of uncertainty reduction
can be briefly described as follows:1
(1) In September of t — 1 the Queen opens the new session of Parliament,

the Minister of Finance presents his budget proposals for year t, and the
Central Planning Bureau releases the preliminary version of the Central
Economic Plan of year t. The figures of that document form stage 1 of the
process of uncertainty reduction (indicated by 1 below the S of September in
Figure 11.1). These figures are based on an econometric model which de¬
scribes a number of endogenous variables in terms of current exogenous
variables and lagged (endogenous and exogenous) variables. The exogenous
changes in t are to be predicted extraneously; the changes in t— 1 are partly
known, partly not known yet, in which case they have to be estimated; given
these predictions and estimates, the corresponding endogenous predictions
for t can be computed. Two amendments must be made, however. First, the
results are not accepted mechanically. If it turns out that a figure is implausible
in the light of other evidence, it is usually revised. Second, the model used is
1 For details see Applied Economic Forecasting, Chapters 3 through 5, particularly

Section 5.1.
frequently revised too. There is no single model that was in continuous use
during the period whose data will be considered here, viz., the ten-year period
1953-1962.
(2) In December of t— 1 the Central Planning Bureau completes the

official version of the Central Economic Plan. [It was actually a few months
later in some of the years of the first half of our period, but December of
t— 1 is presently the typical month.] The figures of this official version form
stage 2 in the process of uncertainty reduction, indicated by 2 below the D of
December in Figure 11.1. Their derivation is identical to that of stage 1, but
the underlying data are of course more complete.
(3) In September of t the Central Planning Bureau releases the pre¬

liminary version of the Central Economic Plan of year t+1. This document
also contains figures of the current year t, and these form stage 3. Two-thirds
of the year has passed, but the actual uncertainty with respect to this year is
larger than this fraction suggests, because statistical data lag behind the
events which they describe. Therefore, model forecasts still play a certain role
with respect to this stage.
(4) In December of t the official version of the Central Economic Plan of

year t+1 is completed. Its figures on the changes in t form stage 4 with
respect to this year. Model forecasts play hardly any role in this stage.
(5) The task of the Central Planning Bureau with respect to year t has now
come to an end. If the object of uncertainty were sunshine or rainfall on a
particular place and a particular day, the problem would cease to exist at the
end of that day, because sunshine and rainfall can be measured accurately
without delay. This is, in general, not true for macroeconomic variables.
When the year has passed there is still uncertainty as to the changes that took
place. There is no longer any variability of the “true values’’ of these changes,
as there was when year t was still (partly or wholly) future; if the variable is
well-defined, its value in year t and its change from the level of t— 1 are now
fixed. But there are the unavoidable delays in the compilation of the com¬
ponents of macroeconomic time series, which cause initial estimates of past
changes to be far from perfect. Past and future are admittedly different to the
extent that it is possible, to some degree, to influence the future, whereas it is
impossible to change the past; but past and future are frequently similar to
the extent that our knowledge of what happened in the past and our knowledge
11.2 PREDICTING THE FUTURE AND ESTIMATING THE PAST 395
of what will happen in the future are both imperfect. One may use different
terms for this imperfection (prediction errors for imperfect forecasts and
estimation errors or measurement errors for imperfect estimates of past
values). In both cases, however, we have to do with the same fundamental
phenomenon of uncertainty. One should of course expect that there will be
less uncertainty when year t has passed than when it is still future; but there
is no a priori reason to assume that therewill be an abrupt decline of uncertain¬
ty as soon as we are in January of year t+1.
The Central Bureau of Statistics continues the task of the Central Planning
Bureau with respect to year t. Its first (preliminary) estimates are released in
September of / +1; they form stage 5 in the process of uncertainty reduction
with respect to year t.
(6) One year later, September of / + 2, the Central Bureau of Statistics

publishes revised estimates. This is stage 6.
(7) Finally, in September of t + 3 the definitive values are published. These

figures, which are based on national accounts data which by then have
become complete, will be regarded as the “true values” even though it does
occur from time to time that additional revisions are published later on
(e.g., when new census data become available).
Five examples are presented in Table 11.1, all dealing with the government
sector or the international sector. The measure of change is the log-change,
i.e., the change in the natural logarithm from the level of t — 1 to the level of t.
For example, the figure upper left indicates that in September 1952 the log-
change in the government wage bill during 1953 was predicted to be .0705.
This corresponds approximately to a 7 per cent increase. No revision was
made in December, but in September 1953 it was raised to .0760. There was a
larger increase (to .0935) in December 1953. Then 1953 has passed and the
figures from stage 5 onward are provided by the Central Bureau of Statistics.
Its September 1954 estimate is .0941, which is only slightly larger than the last
Central Planning Bureau value. There was no further revision in September
1955, but the definitive figure (that of September 1956) is lower than its four
most recent predecessors.
The other variables of the table are characterized by a number of dots
which indicate missing observations. Actually, the forecasts and estimates
were not always presented in the rigorous systematic fashion described above,
and this description is therefore to some extent an idealized picture. As a
TABLE 11.1
SEVEN SUCCESSIVE FORECASTS AND ESTIMATES OF THE LOG-CHANGES IN FIVE VARIABLES,
1953-1962
Stage 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962
Government wage bill

1 705 1561 469 402 989 797 667 825 478 770
2 705 908 848 680 1310 697 671 1033 451 831
3 760 1124 1089 1196 1062 315 770 1089 816 1371
4 935 1336 934 1233 1026 321 214 1117 933 1264
5 941 1443 1054 1238 1358 282 183 1029 751 1178
6 941 1350 1054 939 1300 381 143 973 705 1119
7 749 1542 1235 1071 1197 485 306 988 961 1223
Other government consumption

1 2994 -377 649 286 178 -90 -294 392 1106 816
2 1914 -171 257 592 -587 -598 -208 497 939 1035
3 1354 -60 853 315 227 -640 0 488 770 1098
4 849 646 1033 715 592 -453 -202 753 755 1100
5 . 0 -770 -649 1141 525 1392
6 61 -1244 -556 1202 631 1410
7 995 940 963 1478 75 -1171 -617 1427 845 1018
Government investment
1 2086 488 -1165 431 315 -263 1972 1080 1124 971
2 2086 237 -758 1596 -780 -141 1398 1067 1328 971
3 2986 218 564 996 705 -182 1484 1519 1089 639
4 4298 -454 1244 1804 360 -1133 1574 1016 1133 1105
5 134 -773 1636 959 1299 1519
6 . . 1152 -1027 1309 863 1082 1267
7 4483 -854 948 1385 1837 -699 1289 978 1067 1057
World trade
1 296 0 677 583 583 296 373 583 770 639
2 198 0 392 723 488 296 344 583 770 378
3 770 677 . -202 583 770 583 677
4 908 . . 1398 392 583
5 . • • . 535
6 . 554 594
7 276 843 1098 935 611 40 843 1328 488 751
World price level

1 0 100 198 100 0 198 100 100
2 -613 -459 0 100 296 -101 0 202 99 0
3 -576 -408 100 198 392 -202 0 100 -253 100
4 -324 100 392 296 -101 -202 100 -253 100
5 • . . -202 . -50
. 6 -325 49
7 -502 -284 149 363 237 -315 -233 149 -284 -20

11.3 THE DISTRIBUTION OF ERRORS 397
whole, we shall consider 19 variables; for details we refer to the Appendix of

this chapter.
11.3. The Distribution of the Prediction and Estimation Errors; A

Multiplicative Decomposition of Error Variances
We shall write xith for the log-change in the ith variable in the /,h year
according to stage /?. Hence the stage 7 value is written as xitl; since this is
regarded as the true value, the prediction or estimation error is of the form
xith — Xiti- The objective of this section is to formulate a simple distribution
law for such errors. Suppose, for example, that this law is as follows: The
error xith—xit7 is a random drawing from a normal population with mean
/j.ith and variance ofth. Then the conditional distribution of the true value ;qt7,
given the forecast or estimate xith, is also normal and its mean is xith — [xith and
its variance of(/l. Hence the entropy of this conditional distribution, which will
be used as a measure for the quality of the forecast, is \ log 2neofth. We
mentioned in Section 11.1 that the mean of this distribution does not and
should not play a role when measuring uncertainty. The bias of the forecast
or estimate (measured by Huh) is therefore disregarded completely. It is easily
seen that this is reasonable from the standpoint of forecast (and estimate)
evaluation, at least when the bias is known, because one can correct for this
bias so that the variance is indeed the only source of uncertainty with respect
to the true value, given the forecast formulated.
We shall make the simplifying assumption that all error distributions are
normal with zero mean. Hence pith = 0 and ofth is the mean square error. It will
turn out in the next section that, as far as the evidence goes, the normality
assumption is acceptable, whereas the zero-mean assumption is slightly in
error. The mean square error is therefore systematically larger than the
variance, so that we tend to underestimate the quality of the forecasts and
estimates. However, it will appear that this effect is rather small. The reason
for our over-simplified model is the overwhelmingly large number of potential
parameters. When we have a mean nith and a variance ofth for each error
xith — xit , we are faced with the problem of adjusting 2660 parameters, be¬
7
cause there are 19x 10x7 = 1330 combinations of the indices i, t and h. We
may delete all combinations with h = l (in view of pit7 = oft7=0); still,
2280 parameters remain. Their adjustment is out of the question for the
simple reason that there are only 1140 observed values of the errors xith-xit7,
one for each (/, t, h) combination. [The actual number is even less because of
the missing observations; see the dots of Table 11.1.] If we impose the means
to be zero, so that only variances remain, the number of parameters to be
adjusted is reduced by 50 per cent. Still, the number which remains (1140) is
by far too large compared with the number of observations. Additional
simplifications are therefore necessary.
Consider then the variance o2th of the error xith — xitl for a particular
(zth) variable, a particular (/th) year, and a particular stage h. It stands to
reason that this variance is a decreasing function of h, given i and t. This is
the “stage effect” on the variance, which indicates a reduction of the un¬
certainty with respect to the change in the z‘h variable in year t caused by the
additional data which become available when we move from one stage to the
next. There may also be a “year effect” on the variance, which reflects the
possibility that one year is predicted and estimated, on the average, more
successfully than another. And there is almost certainly an effect of the
variable estimated and predicted, some being “good” and some “bad” in the
sense that the errors are usually small for some variables but usually large
for others. We shall formalize these intuitive ideas by postulating that the
error variances can be decomposed in the following multiplicative manner:1
i = l,...,19
(3.1) £(xith - xitv)2 = AfBfCi t = 1,..., 10
h = 1, ...,6
where At (the positive square root of A2) measures the inaccuracy corre¬
sponding to the zth variable, Bt the inaccuracy corresponding to the tth year,
and Ch the inaccuracy of stage h. [Note that we can apply (3.1) to stage 7 also
by defining C7 = 0.]
It is appropriate to give a more detailed justification of the variance de¬
composition (3.1). It is multiplicative, not additive. The additive decomposi¬
tion Di + Et + Fh is much more familiar, since it is basic to classical variance
analysis. It is doubtful, however, whether additivity is appropriate in the
present context. To show that this is a plausible statement, let us assume that
the variances do not change from one year to the next, so that Bt of the
multiplicative model and Et of the additive model are constant (independent
of t). For example, take B, = 1 and Et = 0, in which case the two competing
specifications are simplified to A2C2h and Dt + Fh, respectively. Let us also
normalize such that Cl = 1 and Ft= 0, so that A(2 = Z>,. stands for the error
1 The same decomposition (applied to the same data) was used by H. Theil and M.
Scholes in “Forecast Evaluation Based on a Multiplicative Decomposition of Mean
Square Errors,” Report 6511 of the Center for Mathematical Studies in Business and
Economics, The University of Chicago (1965). Several numerical results are taken from
this paper, which does not, however, contain any applications of the entropy concept.
11.4 THE ADJUSTMENT OF PARAMETERS 399
variance of the stage 1 forecasts of the z'th variable. Consider then two
variables for which these variances are . 1 and .01, respectively. If F2 = — .003,
the variance of the stage 2 forecast of the first variable is .097, which is a very
small reduction compared with the stage 1 value, whereas that of the second
is .007, which amounts to a relatively substantial uncertainty reduction. This
is evidently unrealistic. It is far better to assume that the percentage effect is
the same for all variables, and this is precisely what the multiplicative model
(3.1) performs. Note also that the standard deviation of the errors under
condition (3.1) has the same simple multiplicative form, AjBtCh, whereas it
takes the more cumbersome form of the square root of Di + Et + Fh in the
additive case. Note finally that the multiplicative form is particularly con¬
venient with respect to the logarithmic entropy expression (1.3). This aspect
will be pursued later in this chapter, especially in Sections 11.5 and 11.7.
The right-hand side of (3.1) involves 19 coefficients At for the 19 variables,
10 “year coefficients” Bt, and 6 “stage coefficients” Ch, hence 35 coefficients
as a whole. There are two multiplicative degrees of freedom, since one may
multiply each At by a>0 and each Bt by /?>0 and each Ch by 1/a/? without
affecting the right-hand side of (3.1), for whatever (/, t, h) combination.
Hence the number of coefficients to be adjusted is 33, which is a sizable re¬
duction compared with the total number of 1140 variances in the left-hand
side of (3.1). The price to be paid is that (3.1) is a restrictive assumption. In
particular, it implies that there is no interaction among the three determining
factors. Suppose, for example, that in the course of time the stage 1 variances
have decreased, whereas those of the other stages remained constant. This is
excluded by (3.1). It can be handled if we multiply the right-hand side of the
equation by a double-subscript term (subscripts t and h = 1 in the example) in
the same way as conventional variance analysis handles interaction by
adding a double-subscript term to Dt + Et + Fh. However, this would lead to a
sizable number of additional coefficients to be estimated, and it will therefore
not be pursued here.
11.4. The Adjustment of the Parameters of the Error Variances; Tests for
Normality
We proceed to the numerical specification of the parameters Ah Bt, Ch of
the variance (3.1). Our starting point is the ratio
(xith
(4.1)
A ?BtzC h2
which in view of (3.1) has unit expectation for each (/, t, h) combination. Let
us multiply such a ratio by Af. We then obtain

ith '•itl )2
/ = !,...,19
B<C2h
the expectation of which is Af, for any (t, hi) combination. For any given i the
average of these ratios over all (/, h) combinations has the same expectation
Af, and it stands to reason that the sampling variation of this average around
Af is less than that of the individual ratios. An obvious adjustment equation
is therefore
(Xith Xitl )
(4.2) = Af i = 1.....19
n. BfCl
t h
where stands for the total number of {t, h) combinations of the zth vari¬
able. [This number is 10 x 6 = 60 when there are no missing observations; the
actual numbers are given in the last column of Table 11.2 below.] Note that
the A’s, B’s and C’s of (4.2) are adjusted values (point estimates), not the
“true” parameter values of the model (3.1).
We can apply a similar procedure to Bt and Ch. When multiplying the
ratio (4.1) by Bf we obtain another ratio whose expectation is Bf for any
(z, h) combination. We equate the average of all these ratios to the (adjusted
value of) Bf:
1 ' (xith ~ xUt)7
(4.3) t = 1,..., 10
n.t. AfCl
where n u is the number of (i, h) combinations of year t. In precisely the same

way we have
(x ith litl)
(4.4) = Ch h = 1,...,6
AfBf
the divisor before the summation signs being the number of (z, t) combinations
of stage h.
The 35 adjustment equations (4.2), (4.3), (4.4) are nonlinear in the co¬
efficients, so that an iterative procedure has to be applied. The procedure
followed consists of a number of rounds, each consisting of three steps. The
steps of the first round are as follows:
Step 1. Take Bf = 1, t = l, ..., 10, and Ch2 = i(7-/z), h = 1, ..., 6. [This
amounts to assuming, in the first step at least, that the variances do not
change in the successive years and that they decrease linearly in the successive
stages.] Then compute the left-hand side of (4.2), which gives the first-round
value of Af, z'= 1, ..., 19.
Step 2. Take Cl = i(7—h), h = 1, 6, and the Af found in Step 1.

Compute the left-hand side of (4.3), which leads to numerical values of B2.
Divide each of these by their geometric average; the resulting ratios are the
first-round values of B2, normalized such that their geometric average is l.1
Step 3. Take the A’s and B’s found in Step 1 and Step 2, respectively.
Compute the left-hand side of (4.4), which leads to numerical values of Ch2.
Divide these by the value of C2, and the resulting ratios are the first-round
values of Ch2 normalized such that Cx — 1.
The second round consists of another set of three steps. In the first we take
the B’s and C’s obtained in the previous round, substitute these into (4.2)
and thus obtain the second-round values of the A’s. These new values,
together with the C’s just mentioned, are then used to obtain - with the aid
of (4.3) and after appropriate normalization - the second-round B values,
and so on. The results are shown in the first column of Table 11.2. We
conclude that the zfs differ widely, that the B’s indicate rather substantial
differences between the various years, and that the successive C’s show their
expected decline. However, we postpone a more detailed discussion of the
numerical results until the next section. Here we consider the extent to which
the assumptions of normality and zero mean are justified. For this purpose
we define the following standardized error:
xith ~ x itl
(4.5)
V
& AiBtCh
Under the assumptions made this standardized error is a normal variate with
zero mean and unit variance for each (i, t, h) combination. To test the zero-
mean assumption we compute the average standardized error for each
variable, each year and each stage:
a,= i = 1.19
t h
(4.6) £ = ^^4* t = i,..., 10

i h
h = 1,..., 6
n..hLL
1 Normalization rules are needed because of the two multiplicative degrees of freedom.
The rule followed here implies that both the B s and their squares have unit geometric
mean. The advantage of this particular rule will be clear when we consider the average
prediction performance for each variable over the ten years; see footnote 1 on page 420.
U-iO c V)
_o
O
X) >
6 o
z X)o
3 v>
ADJUSTED PARAMETER VALUES OF ERROR VARIANCES AND SOME MOMENT MEASURES OF STANDARDIZED ERRORS
/
O
'v>
mriOMnn-N 'tOOMTj-nCOrtO
TfOf'lf'lCOw <N 00 O
Tf SC SO GO <N CO <N SO CO — Tf
OOÔVOTf — CN<OTj- ouNDOcor'ici OS CO —
CT\ OS
— <N — Tj-Mcosot^coio
I— OS OS'O'O — COINOSf^
COI^-Tt — fNTt
CO U- CO CO — IN
OS OS O — <N
3 fNINCNINTf—HTfcO(N(NCN(N|/")INCNC'IC'|co(N cococofNTj-co(NfNco<N (N (N (N CO co co
<D c/d
__ D1/3
TD sOTtU'COCNiOSOcoi^-r^coxJ-COSOOSOSOfNIN
(NONCO'ONTjOOO- — 00 O ~ ro
" a co-OSTfrfior'Tj-'OOF-'Or^Tj-ir, OOY,^ co OO O so Os
* coOÔ'OO'O" O^co — co co — — coOco — (NSOO'cicoOcoOCN TfOOO'tO
* ■ \ f \ \
a3 TtOS'OCO'tf'tÔCOOs-I'OcOOO’t'OOOYiOS soIO CO — so (N Os OS co Tf O0 O l/~l SO IN r-'

co r- os cn —■ oo
o — corôs — oooscor^-fN — coosin — sorîo
vs Ttco tmn co o co co — co — cn cn ~ soo - co
_ r^r"toscs(N'c--0(N
(Ninvocors —— coo — ^
w
n co n in m —
5
Ttoo©—'SO©f^©cO(N so — r- co in
(v, (NSOO’CIOSCOSO©’— CO U- CO N CO lO CO so (Mo COV-)OOSOfN©<NOOSOOO C3« I so co CN ©
^ Os © <N IO TJ- (N lO so so so SO Tt —< © CO © 00 fN — CQ (SjOCOOCOMSOIOIOTj- (j so IN — — ©
~r © in <n —' os K so’ Tf — in" © co © co iri CN co ~ —h ' —I
©
_ »-H 00 ©
—©i— —< <N CN —
—< (N
IN ~ CN CO
GO
a
o > £
CTj ««i
D «.y 3
55'^
i~ u tt
o o-n
G X>
a a o
£W>
— (N CO ^ in so
<u00 00
<o a) 4) o a)
lOYllOYiYllCUO'OSOSO 3 ctf 00 300 aj00 00
cO'tVSSOIÔOOSO — (N
-C'|COTj-iUSONOOO\©«NcOTt'inSOUOOOs
ctj
Os<0\OsOsOsOSOsOSOsOs
s s s s <s ts
These averages are given in the second column of Table 11.2. It appears that
they are negative in a substantial majority of all cases, both for variables and
for years and for stages. This is due to a combination of two factors: (1) most
of the true (stage 7) changes are positive, and (2) most of the forecasts and
estimates are correct as far as the sign is concerned, but underestimation of
the size of the change occurs more frequently than overestimation. This
leads to a majority of negative errors and of negative averages of standardized
errors. The zero-mean assumption is therefore not really acceptable.
The normality assumption may nevertheless be tenable. To test it we
consider skewness coefficients, defined as the third moment around the mean
divided by the third power of the standard deviation. On comparing (4.5)
and (4.2) we conclude that the A’s, B’s and C’s are adjusted such that the ,
standardized errors of the ;'th variable have a unit second moment around
zero. Hence the standard deviation of these errors is equal to the square root
of 1 — af, where a; is their mean as defined in (4.6). Therefore, the skewness
coefficients of the standardized errors for each of the 19 variables are of the
form
j life* - «;)3
(4.7) Skewness - *" ~-
(1 - a;)
In the same way, we can test for deviations from mesokurtosis by computing
kurtosis coefficients:
1
ni
ZZ (Zi,h-«d4
t h
(4.8) Kurtosis =---w~2
(1 - a fy
These measures are presented numerically in the third and fourth columns of
Table 11.2, both for the 19 variables and also for the 10 years and 6 stages
[for which the appropriate extensions of the definitions (4.7) and (4.8) are
straightforward]. It appears that the skewness is positive for 10 variables,
negative for 9; positive for two stages, negative for three, and zero (in three
decimal places) in one case; positive for five years, negative for five other
years. Hence there is no evidence of a consistent skewness pattern for the
data as a whole. The kurtosis coefficient is less than 3 (the mesokurtic value
of the normal distribution) for 14 variables out of 19; the median is 2.76.
For the separate years and stages the kurtosis values are more evenly
distributed around 3. We conclude that on the whole there is not much
evidence against the hypothesis that the errors are normally distributed
around means which are mostly negative. This idea is corroborated by

Figure 11.2, which contains the frequency distribution of all 1063 standardized
errors. The smooth curve is that of the normal density function with
(4.9) ££ith = - } &gth = 1 implying var = 1 - (- = li¬
lt is easily seen that little is changed when we modify the original assump¬
tions on the error distributions such that they follow the specification (4.9).
Using (4.5) we find that <§£ith = — \ implies
(Xith ~ Xitl) =— iAiB,Ch

or equivalently
(4.10) £ (xin\xith) = xith + \AiBtCh
This means that the bias towards change underestimation is removed when
we add to the predicted (or estimated) log-change one-quarter of its root-
Fig. 11.2. Frequency distribution of 1063 standardized prediction and estimation errors
mean-square error.1 Note that we take xith as fixed in (4.10), and that we
consider the random variation of the true log-change xitl, given xith. Under
this interpretation we have for (3.1):
where the left-hand expression stands for the second moment of xitl around
xith, given xith. When we subtract from this conditional second moment the
square of the conditional first moment around xith, we obtain the conditional
variance of xit7:
We conclude that by using the mean square AfS^Cl we overestimate this

variance only slightly. Moreover, this overestimation disappears when we
consider differences between entropies, for the entropy of xitl given xith is
(4.12) H{xitl\xith) = \ \og2ne(\^ A2i BfCl)

= i log ft + i log 27zeA
Hence each entropy is overestimated by a constant amount of
|ilogfs-| = -0323 nit
when we disregard the factor ff in (4.11). Needless to say, this statement

presupposes that S£ith = — \ holds for each and every (/', t, h) combination.
It is plausible that there are actually differences between different variables
(and years and stages), but it is almost equally plausible that the relative
reduction of the conditional variance of xitl below AfBfC,l is usually small,
so that such differences are of minor importance.
The last remarks indicate that the adjustment procedure described in this
section has an approximate character. It is impossible to avoid this. The total
number of observations is 1063 and therefore substantial compared with the
usual numbers in econometric analyses, but its importance is easily exagger¬
ated. We are interested in the distribution of the true value xitl, given
successive conditions xitl, ..., xit6. Since the realization is the same xitl for
each of these six conditional distributions, it stands to reason that our data
1 When the predicted or estimated log-change is negative, this addition leads to a revised
value which is algebraically larger than the original value, so that the correction is in that
case certainly not designed to remove the bias towards change underestimation. However,
this bias is not particularly clear for negative changes, and the number of changes of this
sign is rather small; see Applied Economic Forecasting, Section 3.3.
are characterized by a sizable degree of interdependence. Moreover, one

should expect an interdependence over the 19 variables for each stage and
year, particularly for the first few stages which are based on econometric
model techniques. We shall therefore not attempt to apply significance tests
to the skewness and kurtosis values of Table 11.2, nor to compute standard
errors for the adjusted values of the A’s, B’s and C’s.
11.5. The Information Gain of the Next Stage
In this section we consider a fixed variable in a fixed year and ask the
following question: What is the reduction of uncertainty with respect to the
log-change of this variable when we move to the next stage? In other words,
until a certain moment we have the data of stage h, and then suddenly the
data of stage h + l are released. What is the gain obtained from these new
data? The answer in informational terms is simple. The entropy of xit7,
given xith, is \ log 2neafth, where <jfth is the variance of the error xith — xitl. The
new value given xit h+1 is \ log 2neaft<h+1, and hence the decrease in entropy is
(5.1) H(xitl\xith) - H(xitl|x,t>fc+1)

= \ log2neofth - ^ log2neo^h+1
= i log C1 2h - i log Cl+l = log Ch-

Ch+i
This follows directly from (4.12); it holds both when we apply the correction
factor {-f and when we disregard that factor, and the result log (CJCh+1)
applies to any variable and any year because it is independent of i and t. We
shall call the natural logarithm of Ch/Ch+1 the information gain of stage
h + 1.1 These gains are given in the first three columns of Table 11.3, both for
the ten-year period as a whole and also for the two halves of this period.
[The adjustment of the ,4’s, B’s and C’s for these five-year periods is com¬
pletely analogous to that of the ten-year period.] The last three columns of the
table are based on the idea that one should take account of the different
lengths of the time intervals between successive stages. Take e.g. stage 2,
1 Note the similarity with the information gain of weather forecasts defined in Section 1.4.
Here we take the logarithmic difference of two successive standard deviations of the
realization, given the forecast (auu and cru,h+i), or equivalently the logarithmic difference
of two successive C’s. There we took the difference of the information contents of two
successive messages on the realization (successive in the sense: before and after the fore¬
cast), which amounts to the logarithmic difference of the reciprocals of two successive
probabilities.
11.5 THE INFORMATION GAIN OF THE NEXT STAGE 407
TABLE 11.3
INFORMATION GAINS OF SUCCESSIVE STAGES
Information gain Information gain per month

Stage 1953 1953 1958 1953 1953 1958
h+1 -1962 -1957 -1962 -1962 -1957 -1962
2 .196 .257 .153 .065 .086 .051

3 .476 .396 .560 .053 .044 .062
4 .321 .330 .336 .107 .110 .112
5 .056 .130 .019 .006 .014 .002
6 .145 .072 .212 .012 .006 .018
which is released in December of t— 1, three months after stage 1; compare

this with stage 3 which is released nine months after its predecessor (Sep¬
tember of t). It stands to reason that more data become available during a
nine-month interval than during a three-month interval, so that the in¬
formation gain of stage 3 gives an overly optimistic picture of the quality of
that stage. A natural correction procedure is that of computing information
gains per month, which amounts to a division of the information gain
log (CJCh+1) by the number of months between the stages h and h+1. The
results are shown in the last three columns of Table 11.3; the figures are
obtained by dividing the corresponding value in the first three columns by 3,
9, 3, 9 and 12 in the five successive lines. These divisors correspond to the
September-December release dates described in Section 11.2.
The outcomes show that stage 4 (December of the year to be predicted)
has the largest information gain per month, viz., about one-tenth of a nit.
Stage 2 is next and stage 3 is third (except for the second five-year period in
which their places are interchanged). The uncertainty reductions of the last
two stages, which represent preliminary estimates of changes in the recent
past, are of the order of 1 per cent of a nit per month and therefore very small
compared with those of the earlier stages.
How should we explain these differences? One tentative explanation is the
following. The best two stages (no. 2 and 4) are both “definitive” in the sense
that they are based on the official Central Economic Plans. All other stages
are “preliminary,” i.e., either preliminary versions of the Central Economic
Plans or preliminary estimates by the Central Bureau of Statistics. It is
certainly conceivable that preliminary versions receive less care than definitive
versions do, which would explain the difference in uncertainty reduction per
month between stages 2 and 4 on the one hand and all other stages on the
other hand. It remains true, however, that the results of the last two stages
prior to the definitive stage 7 are surprisingly modest. This feature can be ex¬
plained, to some extent and under appropriate conditions, if we assume that
the stage 7 data are also subject to error. Let us write xitQO for the “really true”
value of the log-change in the /th variable in year t. The observed errors can
then be written as the difference between the corresponding “true” error and
the stage 7 error:
Xith Xit7 i.Xith Xitoo) (Xitl Xitcc)
On squaring both sides and taking the expectation we find
(5.2) S(xith - xit7f = S(xith - xit00)2 + S(xit7 - xi(00)2

- 2S[(xith - xito0)(xit7 - xUaJ]
Suppose now that the stage 7 errors are normally distributed with zero mean
and that they are uncorrelated with the true errors of the earlier stages. Then
the cross-moment on the second line of (5.2) vanishes. Suppose also that the
true errors satisfy
i = 1,..., 19
(5.3) S(xith - xilo0)2 = A2B2C2 t = 1,..., 10
h = l,-.,7
where At and B, are equal to the corresponding coefficients of (3.1). Then,

using (5.2) with the third right-hand term put equal to zero, we conclude
from (3.1) and (5.3):
A]B2Cl = A2B2C2 + A2B2C2
and hence
(5.4) Cl^Cl-C2 h = 1,..., 6
This result implies that the “true quality” of stage h as measured (inversely)
by C2 is always better than the quality measured by Ch2, and that the
difference is a constant (C2) in the sense that it is independent of h. Hence the
effect is larger for the later stages with their smaller Ch values, which means
that their information gains are increased by this correction.
To apply the correction we have to specify a numerical value for C2,
which is of course unknown. It is reasonable to assume C\^C2 and hence,
in view of (5.4), C2^\C\, because otherwise stage 7 would be worse than
stage 6. Hence:
(5.5) C2 = OCl where 0^0^:^

11.6 INDIVIDUAL VARIABLES AND YEARS 409
TABLE 11.4
INFORMATION GAINS PER MONTH OF SUCCESSIVE STAGES UNDER ALTERNATIVE CONDITIONS
OF IMPERFECT STAGE 7 ESTIMATES
Stage h +1 <9 = 0 6 = .1 9 = .2 6=3 0 = .5
II
2 .065 .066 .067 .068 .068 .069
3 .053 .054 .055 .057 .058 .060
4 .107 .113 .119 .126 .134 .143
5 .006 .007 .007 .008 .009 .010
6 .012 .013 .015 .016 .019 .021
7 00 .092 .058 .035 .017 0
Table 11.4 contains the information gains per month (corresponding to the
ten-year period) of the stages 2 through 7 based on the true errors under
alternative assumptions on 0. They are of the form log (CJCh+1) divided by
the relevant number of months (3, 9, 3, 9, 12, 12 for the successive values of
h +1). Note that stage 7 too has its information gain, which is infinity when
0 = 0 (corresponding to perfect stage 7 estimates) and declines towards zero
when 0 approaches \ (in which case stage 7 is no longer an improvement over
stage 6). It is seen that the effect of stage 7 errors on the information gain of
stage 2 is very small. It is much larger for the gain of stage 6, which increases
by almost 80 per cent when 0 increases from 0 to i.e., when we imagine
that stage 7 is equally good (or bad) as stage 6 rather than perfect. Never¬
theless, even under this extreme condition the information gains of stages 5
and 6 are modest compared with those of the earlier stages. Our results are
not too much different from those obtained when we assume stage 7 to be
perfect. This is not too surprising, because the stage 7 error variance is subject
to the limit From now on we shall assume again that stage 7 is
error-free.
11.6. A Mean-square-error Evaluation of Individual Variables and Years
In the previous section we confined ourselves to the C’s of the error

variance specification. We now turn to the ff’s and B’s and consider the
following two questions. (1) Table 11.2 shows that Bf2 is subject to a sizable
downward trend, from about 2 or 3 to less than What is the cause of this
trend? (2) How do the forecasts and estimates compare with naive extrap¬
olations? The second question is particularly relevant for stage 1, because
our evaluation of the successive stages in the previous section is based on a
comparison with the preceding stage, and stage 1 has no predecessor. An

obvious procedure is to define a naive extrapolation method as the predecessor
of stage 1.
Turning to the first question, one possibility is that the declining trend
of B2 is due to an increasing competence of the forecasters; but it is equally
well possible that the trend was in favor of changes which are easier to
predict. Our line of attack will be in the direction of the latter effect, because
it can be measured more easily, at least under appropriate conditions. When
a declining trend remains after account is taken of the second effect, one has
to ascribe this residual trend to other causes, and an increasing competence is
then a very real possibility. The problem is, therefore, to find out which
changes are relatively easy and which are relatively difficult to predict. It is
rather simple to find examples of “difficult” changes, e.g., those which were
induced by the Korean War. [This war was actually prior to our ten-year
period.] Such cases are very specific, however; what we need for our present
purposes is a general statement, preferably in numerical terms, indicating the
degree to which a change is difficult to predict and to estimate. We shall adopt
the viewpoint that, on the average, large changes are predicted and estimated
with larger error than small changes.1 We should be careful about the meaning
of “large,” however. Our basic data are comparable in the sense that they
all have the same dimension (each xith is a log-change per year), but the
annual changes in a variable like Private employment are of course normally
closer to zero than those of Private investment, say. In addition, there are
years of substantial changes and years of small changes. When formulated in
this way, the size of the change in any particular variable in any particular
year has a two-dimensional nature: this size may be large because the
variable is normally subject to substantial changes and/or because the year is
one of exceptional development. It will prove useful to formalize this idea by
assuming that the (true) log-change xitl in the zth variable in year t is the
outcome of a stochastic process which satisfies
(6.1) Sxl, = t;2b;2 i = 1,..., 19; t = 1,..., 10
where Aj measures the degree of change of the z'th variable and B[ the degree
of change of the tth year.
1 There are situations in which this is almost trivially true. Suppose that a change is large
because it differs substantially from the expected value, and that the prediction is equal
to that expectation. Then it is necessarily true that the change and the prediction error
are both large.
11.6 INDIVIDUAL VARIABLES AND YEARS 411
The specification (6.1) is particularly convenient in relation to the mean-

square-error model (3.1). They can be combined as follows:
(6.2)
Hence the ratio of the mean square prediction or estimation error to the
mean square successive difference of the variable can again be separated in
terms of the three indices. [Remember that xitl is a change or successive
difference in terms of natural logarithms.] In particular, consider the stage 1
forecasts under the normalization Cj = 1:
(6.3)
When a zero change is predicted, the error is — xitl; hence the left-hand
denominator can also be regarded as the mean square error of no-change
extrapolation. We conclude that in (6.3) we compare the mean square stage 1
error with that of a particular type of naive extrapolation. This is precisely
the subject of the second question that was asked at the beginning of this
section. Moreover, the right-hand side of (6.3) contains a term in t, which
deals with the variation over time both of the mean square forecast error (Bt)
and of the mean square successive difference (B't). If it is true that the de¬
clining trend of Bt (see Table 11.2) is due to the fact that successive later
years are easier to predict, and if it is also true that “easiness” can be mea¬
sured by the coefficient B't of the mean square successive difference specifi¬
cation (6.1), then we should expect BJB' to be constant. At least, this ratio
should not show an upward or downward trend over time. Hence an analysis
of (6.3) should also be instrumental in answering the first question raised at
the beginning of this section.
It is worthwhile to consider the special case in which all years are predicted
and estimated equally well and in which, moreover, all years are characterized
by the same degree of change: Bt = B’t — 1, t = 1,..., 10. The left-hand ratio of
(6.3) is then equal to the square of Since there is no heterogeneity over
time in this case, one may estimate AJA'i by whose square is defined as
(6.4)
This Ui is nothing else than the inequality coefficient of the stage 1 forecasts
for the /th variable,1 which thus turns out to correspond to a special form of
the more general specification (6.3). The latter specification is indeed more
general, because it takes account of different performance levels (Bt) and
different degrees of change (B't) in different years. Conversely, one may
generalize the inequality coefficient in accordance with (6.3) by dividing the
error xin—xitl by Bt and the change xitl by B't; this amounts to adjusting
errors and changes such that their second moments are constant over time.
We shall therefore call the ratio AJA'i the weighted'inequality coefficient of the
stage 1 forecasts of the z'th variable.
The numerical specification of A\ and B[ is analogous to that of Ah Bt, Ch
described in Section 11.4, but it is simpler due to the fact that there are no
C’s. Each round consists of two steps, based on
(6.5) 1 i = 1,---, 19
10 Lb?
x
(6.6) 1Y ^-*'2
19 La'2 ‘
t = 1,..., 10
i
In the first step we take B't2 = 1, t= 1, . .., 10, and compute the left-hand side
of (6.5), which gives the first-round values of the A'2. These are substituted
into (6.6) in the second step, which leads to B'2 values; again, normalization is
such that their geometric mean is 1. The second round proceeds from these
first-round values, and so on. The results for the B’s are shown in Table 11.5.
The first column contains B2, taken from Table 11.2 and reproduced here for
the sake of convenience. The second column contains B'2 and the fourth
BjB't. It is seen that the latter ratio shows some fluctuations over time, but no
upward or downward trend. We conclude that the downward trend of the
mean square forecast errors is reduced to stationarity when these mean
squares are expressed as fractions of the corresponding mean-square no¬
change extrapolation errors. Hence there is no evidence of an increasing
competence of the forecasters when this competence is defined in relation to
the level of difficulty as measured by the mean square successive difference.
[The other columns of the table will be discussed in the next section.]
The results for At are shown in Table 11.6. The first column contains lOOd,-
(the square root of 104A2, taken from Table 11.2). This is the root-mean-
square error of the stage 1 forecasts of the /th variable for the ten-year period
1 Inequality coefficient “new style,” see Applied Economic Forecasting, Section 2.4.
11.7 AN INFORMATIONAL APPRAISAL OF STAGE 1 413
TABLE 11.5
THE COEFFICIENTS Bt, B't, Bt OF TEN YEARS FOR THE MEAN-SQUARE SPECIFICATIONS
(3.1), (6.1) AND (7.1)
Year B't B? B? Bt! B't Bt/B"t
1953 2.284 2.411 3.381 .97 .82

1954 3.058 1.501 1.193 1.43 1.60
1955 1.380 1.143 .638 1.10 1.47
1956 1.061 1.205 .844 .94 1.12
1957 .826 1.023 2.085 .90 .63
1958 1.200 .720 4.138 1.29 .54
1959 .627 .743 1.190 .92 .73
1960 .580 1.303 .781 .67 .86
1961 .563 .495 .362 1.07 1.25
1962 .482 .568 .158 .92 1.74
as a whole, expressed - approximately - in percentage changes (because

the xith are log-changes and hence approximately relative changes). The
expression “for the ten-year period as a whole” is to be interpreted as: after
correction by means of Bt for the different performance levels in the separate
years of that period. The second column of the table contains 10CL4-, which is
the corresponding root-mean-square no-change extrapolation error in per¬
centage form. The fourth column contains the weighted inequality co¬
efficients AJA'j. It is seen that most of these coefficients are less than 1, which
indicates that at the level of stage 1 the prediction competence is above no¬
change extrapolation. [At the level of stage 2 we have to multiply AJA'i by
C2 = .821, which implies that the weighted inequality coefficients of that stage
are all less than 1 except for Residential Construction.] The median ratio
AJA; is .67; the lower quartile is .54, the upper quartile .99, so that the latter
exceeds the former by about 80 per cent. The upper quartile of the At exceeds
the lower quartile by almost 240 per cent, which is a much larger difference.
This reflects the fact that one obtains a more homogeneous picture of the
prediction performance with respect to the individual variables when their
mean square errors are divided by the mean square successive differences.
77.7. An Informational Appraisal of Stage 1
We conclude this chapter with a comparison of naive extrapolations and

stage 1 forecasts in informational terms. One type of naive extrapolation was
7.82
coOfNCNr'-r-osrôsO-^rôso
7.79
—.55
—.45
—.91
o ^ ^ ^ ^ 9 ^ ‘'l ^ ^ ^ ^
o *nr4vd«ni-iOr4vdr4vdorôas
(7.1)
AND
OO Tf OS
1.04
OS oo r^ <N m r- oo m O oo
.82
.99
1.78
1.19
r- m m © so <N (N os O OS SO nC ITS
(3.1), (6.1)
jH OF NINETEEN VARIABLES FOR THE MEAN-SQUARE SPECIFICATIONS
Os rf m (N r- oo ro OO (N so
.50
^ M OS M
.72
.96
.99
.63
*r) u~> tH lO so r- (N OO </-} rf o os m
'
^}- m m m <N r- so r— (N
2.57
<N o
3.22
2.26
4.57
2.95
X
O 00 V) V) h (N TS I—1 </“> o (N q OO q
O SO (N 00* OO CN (N m* so oo so
m — <N */A OS V~> r- Os r- m

9.00
8.02
3.82
2.82
TABLE 11.6
2.71
X q so o cn 00 Os q oo Os (N
o <N 00 iri l-H rn oo so rn oo* o
o Tt —
o r- (N r- r- so so o X- <N
2.70
5.75
2.68
4.51
2.41
CO 5 o r- <N CO o q os q so
o CO os (N o o (N o
Ai, A[, A",
c
o
a
d
|
THE COEFFICIENTS
o
g (u
*c
 d d ^
O
P
1)
>
<D
X) oP g
a a> <D -*-» OO
<D o 2 CJ
17. Import price level
>
18. Export price level
OX) d w
19. World price level
a> E •5 g <L> 

a _o
2 o o E 6 .5
d u O 3 O £3 o
o
d
a
« *
.2 a .g X C0) £ ~
16. World trade
a ju i >
■*-*
r—c
g
a 6
ctf <L> 03 £ §
a <5 ^
e . 6
<d
d
4—*
P <D
z s 60 c
o £
15. Exports
£
P 2 <D ° £
00 C/3 £ T3 0)
S > £5 oy O
£ «
a a > > OX) (L> •
o o "X P > ’ a
> d 2 °
U U £ k> »—< i £ O O 0
r-’ OO Os « <N m Tj-
considered in the previous section: no change. Under the assumption of

normality we need the variance of the error to compute the entropy of the
true change, given the realization. One should be careful with respect to the
identification of mean square successive differences with the variance of these
differences, because most of the variables are subject to strong upward
trends. For example, the mean of the 10 log-changes xitl of the Wage rate
(7 = 8) is .092, the second moment is .0192; hence the variance is
.0192 - (.092)2 = .0108
which is more than 40 per cent smaller than the second moment around zero.
One can avoid this difficulty by replacing no-change extrapolation by con¬
stant-change extrapolation such that the errors are zero on the average. This
will be our second naive method. We shall postulate that xitl is a normal
variate with mean and variance A"2B"2, or equivalently that
Xjg ~ Hi
(7.1)
A"B"
is a normal variate with zero mean and unit variance. It will be noted that
this constant-change method gives the benefit of the doubt to the extrap-
olator and, therefore, discriminates against the stage 1 forecaster: The extrap-
olator is supposed to predict xitl by fii in every year, and we assume that he
knows these 19 numbers /<;, 7=1, ..., 19. We shall have to take this feature
into account when evaluating the results.
The adjustment procedure for A", B"t, ^ is identical to that of A ■ and B'
except for the /r’s. We generalize (6.5) and (6.6) as follows:
y'(X/l7 ili) ^,2

(7.2) i= 1, • ..,19
L B't'2 ~ ;
t '
y (xul - nf b„2
(7.3) 1 = 1,. ..,10
L A'!2
To specify the /Ts we proceed in a way which is basically the same as that of
At, B„ Ch in Section 11.4. Consider the ratio (7.1), which is supposed to have
zero expectation for each (/, t) combination. It follows that the 10 ratios
(xity — Hi)/Bt', t= 1, ..., 10, also have zero expectation. In addition, they have
all the same variance A'-2. We adjust /zf such that the average of these
10 ratios is equal to zero:
(7'4)
t
The numerical specification is based on three-step rounds:

Step 1. Take for the average of the 10 log-changes xitl, t= 1, 10.
[This amounts to (7.4) with B" = 1.] Then compute the left-hand side of (7.2)
with B” — 1, which gives the first-round values of A!'2, i= 1, ..., 19.
Step 2. Take the same and the A'-2 found in Step 1. Compute the left-
hand side of (7.3) and divide the outcomes by their geometric mean. This
leads to the first-round values of B"2, t= 1, ..., 10, normalized such that their
geometric mean is 1.
Step 3. Use the B” of Step 2 to obtain the first-round values from (7.4).
Again, the second round consists of the same steps but it uses the first-
round /t; and B\\ values in (7.2), and so on. The results for 5" are given in the
third column of Table 11.5, and for Bt\B” in the last column. The outcomes
for the latter ratio indicate that there is more variability than in the case of
Bt/B't, but the two sets of ratios are similar to the extent that there is no clear
evidence of an upward or a downward trend. Table 11.6 contains A'/,
Ail A” and The majority of the ratios AÂ'I exceeds 1 and the median is
1.27. Hence the adjustment by means of the average changes (/i;) does succeed
in reducing the constant-change extrapolation errors substantially below the
no-change extrapolation errors, at least on the average. Figure 11.3 gives the
Fig. 11.3. Frequency distribution of 190 standardized constant-change extrapolation

errors
frequency distribution of the former errors, the number of which is

19 x 10 = 190. There is no evidence of skewness, but the picture suggests that
the distribution is slightly platykurtic. The skewness coefficient is — .09 and
the kurtosis coefficient 2.31.
We now consider the constant-change extrapolation as “stage 0” prior to
stage 1: xit0 = fj,t. What then is the information gain of stage 1 ? Answer (under
the assumption of normality):
(7.5) H(xitl\nt) - H(xitl\xin)

= i\og2neA-2B”2 - ±\og2ne(-Â?B?)
A" B"
= log -J- + log ' - \ log fi-
Ai Bt
where use is made of (4.12). It follows that the information gain of stage 1,
contrary to that of the later stages, is not independent of the variable pre¬
dicted and the year for which it is predicted. The gain is thus different for
each (i, t) combination, but it can be decomposed in a simple additive manner.
Note that the constant term in (7.5), — \ log -j-f, owes its existence to the non¬
zero mean of the forecast errors. Its presence is required by the entropy
formula (1.3), which is formulated in terms of the variance, not the second
moment around zero. Its value is positive and thus contributes to a larger
value of the information gain. It is of course quite reasonable to correct for
the nonzero mean of the forecast errors, because we employ as many as
19 parameters for the same purpose in the extrapolation case. Given this large
difference - 1 parameter in the forecast case versus 19 in the extrapolation
case - it is instructive to compare (7.5) with the same formula in which no
such corrections for nonzero means are made:
4 B’t
(7.6) log + log
A; b]
Here A\ and B[ are the coefficients of the second-moment specification (6.1)

for the simple no-change extrapolation.
The expressions (7.5) and (7.6) are specified numerically in Table 11.7 for
each variable in each year. The table contains also the average over time for
each variable:
A". B" A[ 1 5
(7.7) log +ToZ !°g ~ i log II = log 2 log 1 6
A-i t &t A;
, 4 B; A[
(7.8) log — + To L lo§ D = lQg T
Ai t B, At
D
, ^H
c3 k-.
D
D
«n
vo
<N
ON
<N
m
O
ON
of <N
vo
t-H
<N
VO
of
O
in
O
(N
o”i
Of ON
of
CO
vo
m
O
«n
in
CO
<N
of
o
ON
D E
> O 1 1 i I ( 1 i 1 1 1
i i 1 l 1 1
<
O VO c4 OO ON oo O of oo O <N O vo •n
<N 00 OO VO O' r- r- ’—1 in »n oo O o o m CO r- in
On T-H 1 1 1 T-H i i 1—1 i 1 1 i—i i—i i T-H 1 i i i

1 1 I i
1 1
r- oo <N CO oo of vo r- of VO 00 vo CO
TWO ALTERNATIVE FORMS OF THE INFORMATION GAIN OF STAGE 1 FORECASTS, BY VARIABLES AND YEARS
vO r- Of •n <n r- of CO oo CN l/^ vo r- of t-H O CO (N Of

On
i 1 i i 1 1 1 i i 1 1 i 1 i 1 1 1
1
© o in o of vo t—H r~- O t— ON or OO ON Q oo of
VO 4 CO O O of fN fN CO r- CO o o
t-H 1 1 i 1 1 | 1 1
1 1 i 1
io 1 CO
iH | i—1
bl) ,
Os o co vo co r- c-* VO O OO r- o (N »n vo CO in oo CO
*n <N o O n T-H T-HI H CO <N CO o —H 1 ON CO <N »n co T-H
ON rH |<M
1 1 1 1 1 i 1
|
uq | kq
00 r- vo (N r- CO vo O 00 r- O OO CO in vo r- m oo CO
m 00 o co CO in or of O in vo CO 1— T-H <N vo o oo of VO Of
l
-1-
* | -s.
r- 'N T"C oo H
r- <N VO O vo CO of <N 04 O oo ON (N vo r-
on to o (N Of O (N CO Of in O O m O VO CO of <N
1 1 1 1 i
VO vO OO r-~ o <N oo or m vo Of in o (N r- VO <N vo (N

in VO co of VO CO <N r- o in VO in o vo CO
1 i i i 1 1 1 i i i I i 1 1 1 1 1 1
co *n oo of r- ON in ,
(N CO
l <N r"~ m of m CO ON oo
»n ON vo vo or oo in <n o of CO r- OO oo (N CO ON in CO >n
ON | 1 1 i 1 i i 1 1 i 1 1 1 1 1 1 1 i
'r‘H i
<N CO t"- n vo 00 CO ON , <N ON l vo VO CO <N oo vo

of O r- m ON vo vo o in of ON ON T—1 Of <N VO
o vo of
ON T-H | T-H T-H
i i 1 1 1 i 1 i | 1 1 1 1 1
T“H
1 1 1
CO m vo o of ON co CO vo >n CO of ON CO in CO in ON O
co o <N o O of 1—1 (N iH (N (N oo <N CO of O o
ON
I I I I I I
G
o
a
£
G
o CO g
0) G D
CD > o £
> o
Average over variables
•Boo e^>>
<D D . c/5
D C X D bl) c iD D D
X> D £ D co ctf > 0) >
g G o G ^D >
.2 .a O 3 O -5 o £ £ D
"C
Dh q
00 CJ
D
a a) D D D
0< ■*-* ><
5 u S "O JJ O
a d> d S 8 g 1 d c 5 03 kn
Q< D c3 D £ o £ J3
V-, a a
G s 23 ^ | ? D D D
_ D
g kn £ t> CMC to
In
—
T3 Jh -a
D £
w
G
w
C £
^ c3 c3 c3 bS)
V
D -G D 2 ° ~ o
O O ^ w
> >
*C «
a > -o X ft ft o u,
X o
E >< ■
u u £ PH Ph Ch o£ W £
(N co Of IT) VO oo ON O (N co of in 0> t— oo ON
T—'
CD
bo
a
■— <co On co no go -h
<D ^NOT-HVq^fSjfSGOCNGOOOOO^—iNOcOTj-OO
>
<
(N
vD ^^^^^^^(NGONoor-oNrgr-fNoi-cooN GO
ON ^NoohôF-iVOMNooNOonhîrirtO
NO £! 2 GO CO (N
O
CO
ON
1 1 1 -
t 1 i
O i- (N OO NO o NO o GO O ON CO ON o NO r-
NO IN O <N O oo NO <N ON NO ON CO
(N q- GO q- r-
ON
T—^ T—(
ON 'I o oo (N co OO r- r- ON fN
go l'- O r- CO <N go
q- NO NO On o o <N GO
ON ■q-
1 l
CQ loq
bo
oo o 0 NO 00 o TT O NO OO NO CO r- r- GO OO ro
go q- o q- <N <o CO GO (N <N oo 4- o
On
I I I I
OD
o
GO rjL,£it'~0'00,'0<^>a\r^CT\<NTtOTJ-^D
'Oh:q^»in«iOmiOffiO"(N|co4?i r-
ON q-
NO
O°0'O<N'OfNTtOV0'+0\lor-O^0O(N —i oo co
GO
ON
®^9^t7"'ori'O00OONh?in — © q-
GO
'^-(NC'IÔO'OOU'IOO'NCO ON O of NO GO OO r-
GO
ON
VTipl^COr-riTl''—
O O no <N CO O O <N
I I
GO
oovoooOTtoNOooNocor-r^Gooo (N o < q-
^ M Tt o O ,— GO-—'— (N 7f CO o h co O ’—i co co
ON
I I I I
CO © ->3- © ON(Noooor^r^»—(Go NO (N NO
GO
ON
GO NO T— NCof-<N — goc^nooo o o • r^» co
ON
^ o o
h> Nt On
co
CJ
_o
-4—*
a
q
eq
o =3 q
<D
> q 4-
o
q 8
Average over variables
<D
2 q CD
bD +-»
q >
-O
q
£ s CD
o
o £
- +->? q
<d I <D
CD x q ft a
> E *q •- a) RJ <D P OJ XI
q
g & a <d "3 Dh Oh
q q
3 5
q Ctf CD <D <D 1-1 S I S> I £ 2 _ S-< X) £ »-h *0
C/J ~J ^ 3 3 3
C/5 0)
bfi 2 ° t: g ° ^
, r
§
I—y
> > >
t-H
aj T3
^ U TO B
q ft o & ftft O
v« kT
U U Z su eu &. £ S O O U £ w £ £ w >
MCO^GONOhOOONO fN CO GO NO t"~ OO ON
and the average over all variables for each year:
(7.9) iV X log y + log B‘ - \ log

i At Dt
(7.10) tV X lo§ ~T + lo8 n'

i Ai
it will be noted that the sums over t on the left of (7.7) and (7.8) vanish due to
the fact that Bt, B't, B" are normalized such that they have unit geometric
mean.1 In particular, the average over time of the expressions (7.6) gives simply
the logarithm of the reciprocal of the weighted inequality coefficient, see (7.8).
Equivalently, the reciprocal of this coefficient is equal to the antilog of the
information gain of the stage 1 forecast over the no-change extrapolation -
provided, of course, the errors of both are normally distributed with zero
mean.
The results of Table 11.7 indicate that our appraisal of the stage 1 forecasts
should be rather negative when the constant-change extrapolations form our
yardstick. Almost two thirds of the 190 information gains are negative.
Clearly, the extrapolations’ advantage obtained from adjusting 19 means /q is
more than the stage 1 forecasts can stand. As the lower part of the table shows,
the picture is different when the mean square errors of no-change extrap¬
olations are compared with those of the forecasts; the percentage of
negative values in that part of the table is only 20. Needless to say, these
figures are gains in the sense of continuous information theory only if the
errors have zero mean, which is actually not the case.

The data on predicted and estimated log-changes are all taken from
Tables 5.1 and 5.10 of Applied Economic Forecasting with the following
exceptions. First, the stage 7 data for 1961 and the stage 6 and 7 data for 1962
were not available when that book was written; they are given in Table 11.8,
together with a more detailed description of the 19 variables. Second,
revised stage 7 estimates were made for World trade and the World price
level; the relevant changes are given in Table 11.1, which replaces the corre¬
sponding parts of Table 5.10 of the book just quoted.
It was mentioned in the text that the forecasts and estimates were not al¬
ways presented in the systematic and carefully timed fashion of the seven
successive stages. For example, there are several stage 5 and 6 estimates
which were originally formulated in values whereas volumes are needed. An
adjustment was then made by means of price indices as these were estimated
at the relevant moment. Also, it should be mentioned that World trade and
World price level are variables whose values (including those of the later stages)
are collected by the Central Planning Bureau, not by the Central Bureau
of Statistics.
421
422
DESCRIPTION OF VARIABLES AND ESTIMATED LOG-CHANGES IN 1961 AND 1962*

CONTINUOUS INFORMATION THEORY
All figures are to be multiplied by 1(U4.

Weighted according to the importance for Dutch exports.
BIBLIOGRAPHY
Allen, R. G. D. Mathematical Analysis for Economists. London: MacMillan and Co.,

1947.
Anderson, O., and R. K. Bauer. “t)ber eine K. T. Untersuchung unternehmerischer
ex ante-Ansichten in Abhangigkeit von den ex ante- und ex post-Ansichten des Vor-
monats.” Mimeographed report of the Institute of Statistics of the Mannheim School of
Economics read at the CIRET Conference in Vienna, April 1963.
Anderson, O., R. K. Bauer, H. Fuhrer and J. P. Petersen. “On Short-Term Entre¬
preneurial Reaction Patterns.” Weltwirtschaftliches Archiv, Vol. 81 (1958), pp. 243-
264.
Arrow, K. J., H. B. Chenery, B. S. Minhas and R. M. Solow. “Capital-Labor
Substitution and Economic Efficiency.” The Review of Economics and Statistics, Vol. 43
(1961), pp. 225-250.
Attneave, F. Applications of Information Theory to Psychology. New York: Henry Holt
and Company, 1959.
Barna, T., (editor). The Structural Interdependence of the Economy. Proceedings of an
International Conference on Input-Output Analysis, held in Varenna, June 27-July 10,
1954. New York-Milan: John Wiley and Sons, Inc., and A. Giuffre, Editore.
Barten, A. P., “Consumer Demand Functions under Conditions of Almost Additive
Preferences.” Econometrica, Vol. 32 (1964), pp. 1-38.
Barten, A. P. “Evidence on the Slutsky Conditions for Demand Equations.” Report
6504 of the Econometric Institute of the Netherlands School of Economics, 1965.
Barten, A. P. Theorie en empirie van een volledig stelsel van vraagvergelijkingen. Un¬
published Ph. D. dissertation, Rotterdam 1966.
Barten, A. P., and S. J. Turnovsky. “Some Aspects of the Aggregation Problem for
Composite Demand Equations.” Report 6415 of the Econometric Institute of the Nether¬
lands School of Economics, 1964.
Berkson, J. “Application of the Logistic Function to Bio-assay.” Journal of the American
Statistical Association, Vol. 39 (1944), pp. 357-365.
Brillouin, L. Science and Information Theory. Second Edition. New York: Academic
Press, 1962.
Census, U. S. Bureau of the -. Historical Statistics of the United States, Colonial Times
to 1957. Washington, D.C., 1960.
Coleman, E. J. “Personal Income by States in 1961.” Survey of Current Business, Vol. 42,
No. 8 (August 1962), pp. 8-17.
Cramer, S. J. A Statistical Model of the Ownership of Major Consumer Durables. New
York: Cambridge University Press, 1962.
De Jongh, B. H., see Jongh, B. H. de.
Debreu, G. “Representation of a Preference Ordering by a Numerical Function.”
Chapter XI of Decision Processes, edited by R. M. Thrall, C. H. Coombs and R. L.
423
424 BIBLIOGRAPHY
Davis. New York-London: John Wiley and Sons, Inc., and Chapman and Hall, Ltd.,
1954.
Deming, W. E. Statistical Adjustment of Data. New York: John Wiley and Sons, Inc.,
1944.
Dosser, D. “Allocating the Burden of International Aid for Underdeveloped Countries.”
The Review of Economics and Statistics, Vol. 45 (1963), pp. 207-209.
Dosser, D., and A. T. Peacock. “The International Distribution of Income with
‘Maximum’ Aid.” The Review of Economics and Statistics, Vol. 46 (1964), pp. 432-434.
Emanuel, H., L. H. Klaassen and H. Theil. “On the Interaction of Purchasing Motives
and the Optimal Programming of Their Activation.” Management Science, Vol. 7 (1960),
pp. 62-79.
Finney, D. J. Probit Analysis. Second Edition. New York: Cambridge University Press,
1962.
Fisher, I. The Making of Index Numbers. Boston and New York: Houghton Mifflin
Company, 1922.
Frisch, R. New Methods of Measuring Marginal Utility. Volume 3 of Beitrage zur
Okonomischen Theorie. Tubingen: J. C. B. Mohr, 1932.
Frisch, R. “Annual Survey of General Economic Theory: The Problem of Index
Numbers.” Econometrica, Vol. 4 (1936), pp. 1-38.
Frisch, R. “AComplete Scheme for Computing all Direct and Cross Demand Elasticities
in a Model with Many Sectors.” Econometrica, Vol. 27 (1959), pp. 177-196.
Garner, W. R. and W. J. McGill. “The Relation between Information and Variance
Analyses.” Psychometrika, Vol. 21 (1956), pp. 219-228.
Goldman, S. Information Theory. New York: Prentice-Hall, Inc., 1953.
Hartley, R. V. L. “Transmission of Information.” Bell System Technical Journal,
Vol. 7 (1928), pp. 535-563.
Herfindahl, O. C. “Concentration in the Steel Industry.” Unpublished Ph. D. disser¬
tation, Columbia University, 1950.
Hicks, J. R. Value and Capital. Second Edition. Oxford: Oxford University Press, 1948.
Hirschman, A. O. National Power and the Structure of Foreign Trade. Berkeley and Los
Angeles: University of California Press, 1945.
Hooper, J. W., and H. Theil. “The Information Approach to the Measurement of
Income Inequality.” Report 6501 of the Econometric Institute of the Netherlands School
of Economics (1965).
Houthakker, H. S. “Sur la forme des courbes d’Engel.” Cahiers du Seminaire d'Econo¬
metric, No. 2 (1953), pp. 59-66.
Houthakker, H. S. “Additive Preferences.” Econometrica, Vol. 28 (1960), pp. 244-257.
Houthakker, H. S. “A Note on Self-dual Preferences.” Econometrica, Vol. 33 (1965),
pp. 797-801.
Jongh, B. H. DE. Egalisation, Disparity and Entropy. Utrecht: A. W. Bruna en Zoons -
Uitgevers Maatschappij, 1952.
Kendall, M. G., and A. Stuart. The Advanced Theory of Statistics. Volume 1. London:
Charles Griffin & Co., Ltd., 1958.
Khinchin, A. I. Mathematical Foundations of Information Theory. New York: Dover
Publications, Inc., 1957.
Kloek, T. Indexcijfers: enige methodologische aspecten. Unpublished Ph. D. disser¬
tation, Rotterdam 1966.
BIBLIOGRAPHY 425
Kloek, T., and H. Theil. “International Comparisons of Prices and Quantities Con¬
sumed.” Econometrica, Vol. 33 (1965), pp. 535-556.
Klijn, N., and W. J. Zwezerijnen. “Reliability and Information Content of Weather
Forecasts.” Report 6323 of the Econometric Institute of the Netherlands School of Eco¬
nomics (1963).
Koopman, B. O., and G. E. Kimball. “Information Theory.” Chapter 9 of Notes on
Operations Research 1959, assembled by the Operations Research Center, M.I.T. Cam¬
bridge: The Technology Press, M.I.T., 1959.
Kullback, S. Information Theory and Statistics. New York: John Wiley and Sons, Inc.,
1959.
Leser, C. E. V. “Family Budget Data and Price Elasticities of Demand.” Review of
Economic Studies, Vol. 9 (1941-42), pp. 40-57.
Lisman, J. H. C. “Econometrics and Thermodynamics: A Remark on Davis’ Theory
of Budgets.” Econometrica, Vol. 17 (1949), pp. 59-62.
MacDonald, D. K. C. “Information Theory and Its Application to Taxonomy.”
Journal of Applied Physics, Vol. 23 (1952), pp. 529-531.
Malinvaud, E. “Aggregation Problems in Input-Output Models.” Chapter 8 of The
Structural Interdependence of the Economy, edited by T. Barna (Proceedings of an Inter¬
national Conference on Input-Output Analysis held in Varenna, June 27-July 10, 1954.
New York-Milan: John Wiley and Sons, Inc., and A. Giuffre, Editore.)
Markowitz, H. “The Utility of Wealth.” The Journal of Political Economy, Vol. 60
(1952), pp. 151-158.
Michaeli, M. “Concentration of Exports and Imports: An International Comparison.”
The Economic Journal, Vol. 68 (1958), pp. 722-736.
Middleton, D. An Introduction to Statistical Communication Theory. New York:
McGraw-Hill Book Company, Inc., 1960.
Munksgaard, H. “Einige Resultate des Danischen Konjunkturtests 1956-1959.”
Mimeographed report of the Odense School of Economics read at the CIRET Conference
in Rome, April 1965.
Nataf, A. Theorie des choix et fonctions de demande. Paris: Centre National de la
Recherche Scientifique, 1964.
Pearce, I. F. “An Exact Method of Consumer Demand Analysis.” Econometrica,
Vol. 29 (1961), pp. 499-516.
Pearce, I. F. A Contribution to Demand Analysis. Oxford: Clarendon Press, 1964.
Pikler, A. “Optimum Allocation in Econometrics and Physics.” Weltwirtschaftliches
Archiv, Vol. 66 (1951), pp. 97M32.
Quastler, H. (editor). Information Theory in Biology. Urbana, Ill.: The University of
Illinois Press, 1953.
Quastler, H. (editor). Information Theory in Psychology. Glencoe, Ill.: The Free Press,
1955.
Rajaoja, V. A Study in the Theory of Demand Functions and Price Indexes. Helsinki:
Societas Scientiarum Fennica, 1958.
Rosenbluth, G. “Measures of Concentration.” Pages 57-95 of Business Concentration
and Price Policy, Special Conference Series of the National Bureau of Economic Research.
Princeton: Princeton University Press, 1955.
Rosenstein-Rodan, P. N. “International Aid for Underdeveloped Countries.” The
Review of Economics and Statistics, Vol. 43 (1961), pp. 107-138.
426 BIBLIOGRAPHY
Roy, R. De I’utilite; contribution a la theorie des choix. Paris: Hermann et Cie, 1942.
Samuelson, P. A. Foundations of Economic Analysis. Cambridge: Harvard University
Press, 1947.
Samuelson, P. A. “Using Full Duality to Show that Simultaneously Additive Direct
and Indirect Utilities Implies Unitary Price Elasticity of Demand.” Econometrica, Vol. 33
(1965), pp. 781-796.
Schwartz, C. F., and R. E. Graham, Jr. Personal Income by States since 1929. A Supple¬
ment to the Survey of Current Business (Washington, D.C., 1956), pp. 57-65.
Schwartz, L. S. Principles of Coding, Filtering, and Information Theory. Baltimore-
London: Spartan Books, Inc., and Cleaver-Hume Press, Ltd., 1963.
Shannon, C. E. “A Mathematical Theory of Communication.” Bell System Technical
Journal, Vol. 27 (1948), pp. 379-423, 623-656.
Shannon, C. E., and W. Weaver. The Mathematical Theory of Communication. Urbana,
Ill.: The University of Illinois Press, 1949.
Skolka, J. The Aggregation Problem in Input-Output Analysis (Czech with English
summary). Prague: Czechoslovakian Academy of Sciences, 1964.
Soest, J. L. van. “A Contribution of Information Theory to Sociology.” Synthese,
Vol. 9 (1954), pp. 265-273.
Somermeyer, W. H., and J. W. W. A. Wit. “Een ‘verdeel’-model.” Mimeographed
memorandum of the Netherlands Central Bureau of Statistics (June 1956).
Stone, R. “Linear Expenditure Systems and Demand Analysis: An Application to the
Pattern of British Demand.” The Economic Journal, Vol. 64 (1954), pp. 511-527.
Stone, R., and A. Brown. A Computable Model of Economic Growth. Volume 1 of the
series “A Programme for Growth.” London: Chapman and Hall, Ltd., 1964.
Strotz, R. H. “The Empirical Implications of a Utility Tree.” Econometrica, Vol. 25
(1957), pp. 269-280.
Theil, H. “Linear Aggregation in Input-Output Analysis.” Econometrica, Vol. 25
(1957), pp. 111-122.
Theil, H. “Some Developments of Economic Thought in the Netherlands.” American
Economic Review, Vol. 54 (1964), pp. 34-55.
Theil, H. “The Information Approach to Demand Analysis.” Econometrica, Vol. 33
(1965), pp. 67-87.
Theil, H. Applied Economic Forecasting. Volume 4 of the series “Studies in Mathe¬
matical and Managerial Economics.” Amsterdam-Chicago: North-Holland Publishing
Company and Rand McNally & Company, 1966.
Theil, H., and R. H. Mnookin. “The Information Value of Demand Equations and
Predictions.” Journal of Political Economy, Vol. 74 (1966), pp. 34-45.
Theil, H., and H. Neudecker. “Substitution, Complementarity, and the Residual
Variation around Engel Curves. Review of Economic Studies, Vol. 25 (1957), pp. 114—
123.
Theil, H., and M. Scholes. “Forecast Evaluation Based on a Multiplicative Decompo¬
sition of Mean Square Errors.” Report 6511 of the Center for Mathematical Studies in
Business and Economics, The University of Chicago, 1965.
Theil, H., M. Scholes and P. Uribe. “An Informational Approach to the Measurement
of Industrial Concentration.” Report 6512 of the Center for Mathematical Studies in
Business and Economics, The University of Chicago, 1965.
Theil, H., and C. B. Tilanus. “The Demand for Production Factors and the Price
BIBLIOGRAPHY 427
Sensitivity of Input-Output Predictions.” International Economic Review, Vol. 5 (1964),

pp. 258-272.
Theil, H., and P. Uribe. “The Information Approach to the Aggregation of Input-
Output Tables.” Report 6503 of the Center for Mathematical Studies in Business and
Economics, The University of Chicago, 1965.
Thrall, R. M., C. H. Coombs and R. L. Davis (editors). Decision Processes. New York-
London: John Wiley and Sons, Inc., and Chapman and Hall, Ltd., 1954.
Tilanus, C. B. “Thirteen Aggregated Input-Output Tables, The Netherlands 1948-
1960.” Report 6423 of the Econometric Institute of the Netherlands School of Economics,
1964.
Tilanus, C. B., and H. Theil. “The Information Approach to the Evaluation of Input-
Output Forecasts.” Econometrica, Vol. 32 (1965), pp. 847-862.
Tobin, J. “Estimation of Relationships for Limited Dependent Variables.” Econometrica,
Vol. 26 (1958), pp. 24-36.
Tornqvist, L. “The Bank of Finland’s Consumption Price Index.” Bank of Finland
Monthly Bulletin, No. 10 (1936), pp. 1-8.
Uribe, P., C. G. de Leeuw and H. Theil. “The Information Approach to the Prediction
of Interregional Trade Flows.” Report 6507 of the Econometric Institute of the Nether¬
lands School of Economics, 1965.
Uzawa, H. “Production Functions with Constant Elasticities of Substitution.” Review
of Economic Studies, Vol. 29 (1962), pp. 291-299.
Van Soest, J. L., see Soest, J. L. van.
Waelbroeck, J. “Une nouvelle methode d’analyse des matrices d’echanges inter-
nationaux.” Cahiers economiques de Bruxelles, No. 21 (1964), pp. 93-114.
Ward's Automotive Yearbook. Detroit, Michigan; annual publication.
Wold, H. Demand Analysis. New York: John Wiley and Sons, Inc., 1953.
Woodward, P. M. Probability and Information Theory with Applications to Radar.
New York-London: McGraw Hill Book Company and Pergamon Press, Ltd., 1953.
Yaglom, A. M., et I. M. Yaglom. Probability et information. Paris: Monographies
Dunod, 1959.
Zellner, A., and Tong Hun Lee. “Joint Estimation of Relationships for Limited
Dependent Variables.” Econometrica, Vol. 33 (1965), pp. 382-394.
TABLES
Tables A, B, and C are both in logarithms to the base 2 and in natural

logarithms. Table A contains the logarithms of all integers from 1 to 1000 as
well as of all multiples of 10 from 1000 to 10 000. Table B contains the
logarithm of the probability p, specified as a multiple of .001. Table C gives
the logit corresponding to p, also specified as a multiple of .001, but for
values ^ .5 only. The logit of a p smaller than .5 is found by replacing p by
1 —p and adding a minus sign to the resulting logit value.
Table D consists of two parts, D1 and D2. The former deals with the
transformation of relative changes to log-changes (the change in the natural
logarithm). Let the relative change be x (hence the percentage change lOOx),
then the new value of the variable is A (l + x) if A is the old value. Hence the
log-change is
log* 4(1 + x) - logeA = log*(l + x)
which is specified in Table D1 for positive and negative multiples of .001

from 0 to 1. For example, an increase of 10.1 per cent (x = .101) corresponds
to a log-change of .096 219 and a decrease of the same percentage to a log-
change of —.106 472.
Table D2 specifies the transformation of log-changes to relative changes.
Let the new logarithmic value of the variable by B + y, where B is the old
logarithmic value, so that y is the log-change. The relative change is then
e ' — e eJ(ey-l)
= ey-l
which is given in Table D2 for positive and negative multiples of .001 from
0 to 1. For example, a log-change of .101 corresponds to a percentage in¬
crease of 10.6277 and a log-change of -.101 to a percentage decrease of
9.6067.
428
TABLES 429
TABLE A
LOGARITHMS TO THE BASE 2 AND NATURAL LOGARITHMS OF INTEGERS FROM 1 TO 10,000
n logs n log* n n log2 n loge n
i 0 0 51 5.672 43 3.931 83
2 1.000 00 .693 147 52 5.700 44 3.951 24
3 1.584 96 1.098 61 53 5.727 92 3.970 29
4 2.000 00 1.386 29 54 5.754 89 3.988 98
5 2.321 93 1.609 44 55 5.781 36 4.007 33
6 2.584 96 1.791 76 56 5.807 35 4.025 35
7 2.807 35 1.945 91 57 5.832 89 4.043 05
8 3.000 00 2.079 44 58 5.857 98 4.060 44
9 3.169 92 2.197 22 59 5.882 64 4.077 54
10 3.321 93 2.302 59 60 5.906 89 4.094 34
11 3.459 43 2.397 90 61 5.930 74 4.110 87

12 3.584 96 2.484 91 62 5.954 20 4.127 13
13 3.700 44 2.564 95 63 5.977 28 4.143 13
14 3.807 35 2.639 06 64 6.000 00 4.158 88
15 3.906 89 2.708 05 65 6.022 37 4.174 39
16 4.000 00 2.772 59 66 6.044 39 4.189 65
17 4.087 46 2.833 21 67 6.066 09 4.204 69
18 4.169 92 2.890 37 68 6.087 46 4.219 51
19 4.247 93 2.944 44 69 6.108 52 4.234 11
20 4.321 93 2.995 73 70 6.129 28 4.248 50
21 4.392 32 3.044 52 71 6.149 75 4.262 68

22 4.459 43 3.091 04 72 6.169 92 4.276 67
23 4.523 56 3.135 49 73 6.189 82 4.290 46
24 4.584 96 3.178 05 74 6.209 45 4.304 07
25 4.643 86 3.218 88 75 6.228 82 4.317 49
26 4.700 44 3.258 10 76 6.247 93 4.330 73
27 4.754 89 3.295 84 77 6.266 79 4.343 81
28 4.807 35 3.332 20 78 6.285 40 4.356 71
29 4.857 98 3.367 30 79 6.303 78 4.369 45
30 4.906 89 3.401 20 80 6.321 93 4.382 03
31 4.954 20 3.433 99 81 6.339 85 4.394 45

32 5.000 00 3.465 74 82 6.357 55 4.406 72
33 5.044 39 3.496 51 83 6.375 04 4.418 84
34 5.087 46 3.526 36 84 6.392 32 4.430 82
35 5.129 28 3.555 35 85 6.409 39 4.442 65
36 5.169 93 3.583 52 86 6.426 26 4.454 35
37 5.209 45 3.610 92 87 6.442 94 4.465 91
38 5.247 93 3.637 59 88 6.459 43 4.477 34
39 5.285 40 3.663 56 89 6.475 73 4.488 64
40 5.321 93 3.688 88 90 6.491 85 4.499 81
41 5.357 55 3.713 57 91 6.507 79 4.510 86

42 5.392 32 3.737 67 92 6.523 56 4.521 79
43 5.426 26 3.761 20 93 6.539 16 4.532 60
44 5.459 43 3.784 19 94 6.554 59 4.543 29
45 5.491 85 3.806 66 95 6.569 86 4.553 88
46 5.523 56 3.828 64 96 6.584 96 4.564 35
47 5.554 59 3.850 15 97 6.599 91 4.574 71
48 5.584 96 3.871 20 98 6.614 71 4.584 97
49 5.614 71 3.891 82 99 6.629 36 4.595 12
50 5.643 86 3.912 02 100 6.643 86 4.605 17
430 TABLES
TABLE A (continued)
n log2 n loge n n log2 n loge n
101 6.658 21 4.615 12 151 7.238 40 5.017 28

102 6.672 43 4.624 97 152 7.247 93 5.023 88
103 6.686 50 4.634 73 153 7.257 39 5.030 44
104 6.700 44 4.644 39 154 7.266 79 5.036 95
105 6.714 25 4.653 96 155 7.276 12 5.043 43
106 6.727 92 4.663 44 156 7.285 40 5.049 86
107 6.741 47 4.672 83 157 7.294 62 5.056 25
108 6.754 89 4.682 13 158 7.303 78 5.062 60
109 6.768 18 4.691 35 159 7.312 88 5.068 90
110 6.781 36 4.700 48 160 7.321 93 5.075 17
111 6.794 42 4.709 53 161 7.330 92 5.081 40

112 6.807 35 4.718 50 162 7.339 85 5.087 60
113 6.820 18 4.727 39 163 7.348 73 5.093 75
114 6.832 89 4.736 20 164 7.357 55 5.099 87
115 6.845 49 4.744 93 165 7.366 32 5.105 95
116 6.857 98 4.753 59 166 7.375 04 5.111 99
117 6.870 36 4.762 17 167 7.383 70 5.117 99
118 6.882 64 4.770 68 168 7.392 32 5.123 96
119 6.894 82 4.779 12 169 7.400 88 5.129 90
120 6.906 89 4.787 49 170 7.409 39 5.135 80
121 6.918 86 4.795 79 171 7.417 85 5.141 66

122 6.930 74 4.804 02 172 7.426 26 5.147 49
123 6.942 51 4.812 18 173 7.434 63 5.153 29
124 6.954 20 4.820 28 174 7.442 94 5.159 06
125 6.965 78 4.828 31 175 7.451 21 5.164 79
126 6.977 28 4.836 28 176 7.459 43 5.170 48
127 6.988 68 4.844 19 177 7.467 61 5.176 15
128 7.000 00 4.852 03 178 7.475 73 5.181 78
129 7.011 23 4.859 81 179 7.483 82 5.187 39
130 7.022 37 4.867 53 180 7.491 85 5.192 96
131 7.033 42 4.875 20 181 7.499 85 5.198 50

132 7.044 39 4.882 80 182 7.507 79 5.204 01
133 7.055 28 4.890 35 183 7.515 70 5.209 49
134 7.066 09 4.897 84 184 7.523 56 5.214 94
135 7.076 82 4.905 27 185 7.531 38 5.220 36
136 7.087 46 4.912 65 186 7.539 16 5.225 75
137 7.098 03 4.919 98 187 7.546 89 5.231 11
138 7.108 52 4.927 25 188 7.554 59 5.236 44
139 7.118 94 4.934 47 189 7.562 24 5.241 75
140 7.129 28 4.941 64 190 7.569 86 5.247 02
141 7.139 55 4.948 76 191 7.577 43 5.252 27
142 7.149 75 4.955 83 192 7.584 96 5.257 50
143 7.159 87 4.962 84 193 7.592 46 5.262 69
144 7.169 92 4.969 81 194 7.599 91 5.267 86
145 7.179 91 4.976 73 195 7.607 33 5.273 00
146 7.189 82 4.983 61 196 7.614 71 5.278 11
147 7.199 67 4.990 43 197 7.622 05 5.283 20
148 7.209 45 4.997 21 198 7.629 36 5.288 27
149 7.219 17 5.003 95 199 7.636 62 5.293 30
150 7.228 82 5.010 64 200 7.643 86 5.298 32
TABLES 431
TABLE A (continued)
n log2 n log« n n log2 n loge n
201 7.651 05 5.303 30 251 7.971 54 5.525 45

202 7.658 21 5.308 27 252 7.977 28 5.529 43
203 7.665 34 5.313 21 253 7.982 99 5.533 39
204 7.672 43 5.318 12 254 7.988 68 5.537 33
205 7.679 48 5.323 01 255 7.994 35 5.541 26
206 7.686 50 5.327 88 256 8.000 00 5.545 18
207 7.693 49 5.332 72 257 8.005 62 5.549 08
208 7.700 44 5.337 54 258 8.011 23 5.552 96
209 7.707 36 5.342 33 259 8.016 81 5.556 83
210 7.714 25 5.347 11 260 8.022 37 5.560 68
211 7.721 10 5.351 86 261 8.027 91 5.564 52

212 7.727 92 5.356 59 262 8.033 42 5.568 34
213 7.734 71 5.361 29 263 8.038 92 5.572 15
214 7.741 47 5.365 98 264 8.044 39 5.575 95
215 7.748 19 5.370 64 265 8.049 85 5.579 73
216 7.754 89 5.375 28 266 8.055 28 5.583 50
217 7.761 55 5.379 90 267 8.060 70 5.587 25
218 7.768 18 5.384 50 268 8.066 09 5.590 99
219 7.774 79 5.389 07 269 8.071 46 5.594 71
220 7.781 36 5.393 63 270 8.076 82 5.598 42
221 7.787 90 5.398 16 271 8.082 15 5.602 12

222 7.794 42 5.402 68 272 8.087 46 5.605 80
223 7.800 90 5.407 17 273 8.092 76 5.609 47
224 7.807 35 5.411 65 274 8.098 03 5.613 13
225 7.813 78 5.416 10 275 8.103 29 5.616 77
226 7.820 18 5.420 54 276 8.108 52 5.620 40
227 7.826 55 5.424 95 277 8.113 74 5.624 02
228 7.832 89 5.429 35 278 8.118 94 5.627 62
229 7.839 20 5.433 72 279 8.124 12 5.631 21
230 7.845 49 5.438 08 280 8.129 28 5.634 79
231 7.851 75 5.442 42 281 8.134 43 5.638 35

232 7.857 98 5.446 74 282 8.139 55 5.641 91
233 7.864 19 5.451 04 283 8.144 66 5.645 45
234 7.870 36 5.455 32 284 8.149 75 5.648 97
235 7.876 52 5.459 59 285 8.154 82 5.652 49
236 7.882 64 5.463 83 286 8.159 87 5.655 99
237 7.888 74 5.468 06 287 8.164 91 5.659 48
238 7.894 82 5.472 27 288 8.169 92 5.662 96
239 7.900 87 5.476 46 289 8.174 93 5.666 43
240 7.906 89 5.480 64 290 8.179 91 5.669 88
241 7.912 89 5.484 80 291 8.184 88 5.673 32

242 7.918 86 5.488 94 292 8.189 82 5.676 75
243 7.924 81 5.493 06 293 8.194 76 5.680 17
244 7.930 74 5.497 17 294 8.199 67 5.683 58
245 7.936 64 5.501 26 295 8.204 57 5.686 98
246 7.942 51 5.505 33 296 8.209 45 5.690 36
247 7.948 37 5.509 39 297 8.214 32 5.693 73
248 7.954 20 5.513 43 298 8.219 17 5.697 09
249 7.960 00 5.517 45 299 8.224 00 5.700 44
250 7.965 78 5.521 46 300 8.228 82 5.703 78
432 TABLES
TABLE A (continued)
n Iog2 n log„ n n log2 n loge n
301 8.233 62 5.707 11 351 8.455 33 5.860 79

302 8.238 40 5.710 43 352 8.459 43 5.863 63
303 8.243 17 5.713 73 353 8.463 52 5.866 47
304 8.247 93 5.717 03 354 8.467 61 5.869 30
305 8.252 67 5.720 31 355 8.471 68 5.872 12
306 8.257 39 5.723 59 356 8.475 73 5.874 93
307 8.262 09 5.726 85 357 8.479 78 5.877 74
308 8.266 79 5.730 10 358 8.483 82 5.880 53
309 8.271 46 5.733 34 359 8.487 84 5.883 32
310 8.276 12 5.736 57 360 8.491 85 5.886 10
311 8.280 77 5.739 79 361 8.495 86 5.888 88

312 8.285 40 5.743 00 362 8.499 85 5.891 64
313 8.290 02 5.746 20 363 8.503 83 5.894 40
314 8.294 62 5.749 39 364 8.507 79 5.897 15
315 8.299 21 5.752 57 365 8.511 75 5.899 90
316 8.303 78 5.755 74 366 8.515 70 5.902 63
317 8.308 34 5.758 90 367 8.519 64 5.905 36
318 8.312 88 5.762 05 368 8.523 56 5.908 08
319 8.317 41 5.765 19 369 8.527 48 5.910 80
320 8.321 93 5.768 32 370 8.531 38 5.913 50
321 8.326 43 5.771 44 371 8.535 28 5.916 20

322 8.330 92 5.774 55 372 8.539 16 5.918 89
323 8.335 39 5.777 65 373 8.543 03 5.921 58
324 8.339 85 5.780 74 374 8.546 89 5.924 26
325 8.344 30 5.783 83 375 8.550 75 5.926 93
326 8.348 73 5.786 90 376 8.554 59 5.929 59
327 8.353 15 5.789 96 377 8.558 42 5.932 25
328 8.357 55 5.793 01 378 8.562 24 5.934 89
329 8.361 94 5.796 06 379 8.566 05 5.937 54
330 8.366 32 5.799 09 380 8.569 86 5.940 17
331 8.370 69 5.802 12 381 8.573 65 5.942 80
332 8.375 04 5.805 13 382 8.577 43 5.945 42
333 8.379 38 5.808 14 383 8.581 20 5.948 03
334 8.383 70 5.811 14 384 8.584 96 5.950 64
335 8.388 02 5.814 13 385 8.588 71 5.953 24
336 8.392 32 5.817 11 386 8.592 46 5.955 84
337 8.396 60 5.820 08 387 8.596 19 5.958 42
338 8.400 88 5.823 05 388 8.599 91 5.961 01
339 8.405 14 5.826 00 389 8.603 63 5.963 58
340 8.409 39 5.828 95 390 8.607 33 5.966 15
341 8.413 63 5.831 88 391 8.611 02 5.968 71
342 8.417 85 5.834 8! 392 8.614 71 5.971 26
343 8.422 06 5.837 73 393 8.618 39 5.973 81
344 8.426 26 5.840 64 394 8.622 05 5.976 35
345 8.430 45 5.843 54 395 8.625 71 5.978 89
346 8.434 63 5.846 44 396 8.629 36 5.981 41
347 8.438 79 5.849 32 397 8.633 00 5.983 94
348 8.442 94 5.852 20 398 8.636 62 5.986 45
349 8.447 08 5.855 07 399 8.640 24 5.988 96
350 8.451 21 5.857 93 400 8.643 86 5.991 46
TABLES 433
TABLE A (continued)
n log2 n logc n n log2 n loge n
401 8.647 46 5.993 96 451 8.816 98 6.111 47

402 8.651 05 5.996 45 452 8.820 18 6.113 68
403 8.654 64 5.998 94 453 8.823 37 6.115 89
404 8.658 21 6.001 41 454 8.826 55 6.118 10
405 8.661 78 6.003 89 455 8.829 72 6.120 30
406 8.665 34 6.006 35 456 8.832 89 6.122 49
407 8.668 88 6.008 81 457 8.836 05 6.124 68
408 8.672 43 6.011 27 458 8.839 20 6.126 87
409 8.675 96 6.013 72 459 8.842 35 6.129 05
410 8.679 48 6.016 16 460 8.845 49 6.131 23
411 8.682 99 6.018 59 461 8.848 62 6.133 40

412 8.686 50 6.021 02 462 8.851 75 6.135 56
413 8.690 00 6.023 45 463 8.854 87 6.137 73
414 8.693 49 6.025 87 464 8.857 98 6.139 88
415 8.696 97 6.028 28 465 8.861 09 6.142 04
416 8.700 44 6.030 69 466 8.864 19 6.144 19
417 8.703 90 6.033 09 467 8.867 28 6.146 33
418 8.707 36 6.035 48 468 8.870 36 6.148 47
419 8.710 81 6.037 87 469 8.873 44 6.150 60
420 8.714 25 6.040 25 470 8.876 52 6.152 73
421 8.717 68 6.042 63 471 8.879 58 6.154 86

422 8.721 10 6.045 01 472 8.882 64 6.156 98
423 8.724 51 6.047 37 473 8.885 70 6.159 10
424 8.727 92 6.049 73 474 8.888 74 6.161 21
425 8.731 32 6.052 09 475 8.891 78 6.163 31
426 8.734 71 6.054 44 476 8.894 82 6.165 42
427 8.738 09 6.056 78 477 8.897 85 6.167 52
428 8.741 47 6.059 12 478 8.900 87 6.169 61
429 8.744 83 6.061 46 479 8.903 88 6.171 70
430 8.748 19 6.063 79 480 8.906 89 6.173 79
431 8.751 54 6.066 11 481 8.909 89 6.175 87

432 8.754 89 6.068 43 482 8.912 89 6.177 94
433 8.758 22 6.070 74 483 8.915 88 6.180 02
434 8.761 55 6.073 04 484 8.918 86 6.182 08
435 8.764 87 6.075 35 485 8.921 84 6.184 15
436 8.768 18 6.077 64 486 8.924 81 6.186 21
437 8.771 49 6.079 93 487 8.927 78 6.188 26
438 8.774 79 6.082 22 488 8.930 74 6.190 32
439 8.778 08 6.084 50 489 8.933 69 6.192 36
440 8.781 36 6.086 77 490 8.936 64 6.194 41
441 8.784 63 6.089 04 491 8.939 58 6.196 44

442 8.787 90 6.091 31 492 8.942 51 6.198 48
443 8.791 16 6.093 57 493 8.945 44 6.200 51
444 8.794 42 6.095 82 494 8.948 37 6.202 54
445 8.797 66 6.098 07 495 8.951 28 6.204 56
446 8.800 90 6.100 32 496 8.954 20 6.206 58
447 8.804 13 6.102 56 497 8.957 10 6.208 59
448 8.807 35 6.104 79 498 8.960 00 6.210 60
449 8.810 57 6.107 02 499 8.962 90 6.212 61
450 8.813 78 6.109 25 500 8.965 78 6.214 61
434 TABLES
TABLE A (continued)
n log2 n loge n n log2 n log? n
501 8.968 67 6.216 61 551 9.105 91 6.311 73

502 8.971 54 6.218 60 552 9.108 52 6.313 55
503 8.974 41 6.220 59 553 9.111 14 6.315 36
504 8.977 28 6.222 58 554 9.113 74 6.317 16
505 8.980 14 6.224 56 555 9.116 34 6.318 97
506 8.982 99 6.226 54 556 9.118 94 6.320 77
507 8.985 84 6.228 51 557 9.121 53 6.322 57
508 8.988 68 6.230 48 558 9.124 12 6.324 36
509 8.991 52 6.232 45 559 9.126 70 6.326 15
510 8.994 35 6.234 41 560 9.129 28 6.327 94
511 8.997 18 6.236 37 561 9.131 86 6.329 72

512 9.000 00 6.238 32 562 9.134 43 6.331 50
513 9.002 82 6.240 28 563 9.136 99 6.333 28
514 9.005 62 6.242 22 564 9.139 55 6.335 05
515 9.008 43 6.244 17 565 9.142 11 6.336 83
516 9.011 23 6.246 11 566 9.144 66 6.338 59
517 9.014 02 6.248 04 567 9.147 20 6.340 36
518 9.016 81 6.249 98 568 9.149 75 6.342 12
519 9.019 59 6.251 90 569 9.152 28 6.343 88
520 9.022 37 6.253 83 570 9.154 82 6.345 64
521 9.025 14 6.255 75 571 9.157 35 6.347 39

522 9.027 91 6.257 67 572 9.159 87 6.349 14
523 9.030 67 6.259 58 573 9.162 39 6.350 89
524 9.033 42 6.261 49 574 9.164 91 6.352 63
525 9.036 17 6.263 40 575 9.167 42 6.354 37
526 9.038 92 6.265 30 576 9.169 92 6.356 11
527 9.041 66 6.267 20 577 9.172 43 6.357 84
528 9.044 39 6.269 10 578 9.174 93 6.359 57
529 9.047 12 6.270 99 579 9.177 42 6.361 30
530 9.049 85 6.272 88 580 9.179 91 6.363 03
531 9.052 57 6.274 76 581 9.182 39 6.364 75

532 9.055 28 6.276 64 582 9.184 88 6.366 47
533 9.057 99 6.278 52 583 9.187 35 6.368 19
534 9.060 70 6.280 40 584 9.189 82 6.369 90
535 9.063 40 6.282 27 585 9.192 29 6.371 61
536 9.066 09 6.284 13 586 9.194 76 6.373 32
537 9.068 78 6.286 00 587 9.197 22 6.375 02
538 9.071 46 6.287 86 588 9.199 67 6.376 73
539 9.074 14 6.289 72 589 9.202 12 6.378 43
540 9.076 82 6.291 57 590 9.204 57 6.380 12
541 9.079 48 6.293 42 591 9.207 01 6.381 82

542 9.082 15 6.295 27 592 9.209 45 6.383 51
543 9.084 81 6.297 11 593 9.211 89 6.385 19
544 9.087 46 6.298 95 594 9.214 32 6.386 88
545 9.090 11 6.300 79 595 9.216 75 6.388 56
546 9.092 76 6.302 62 596 9.219 17 6.390 24
547 9.095 40 6.304 45 597 9.221 59 6.391 92
548 9.098 03 6.306 28 598 9.224 00 6.393 59
549 9.100 66 6.308 10 599 9.226 41 6.395 26
550 9.103 29 6.309 92 600 9.228 82 6.396 93
TABLES 435
TABLE A (continued)
n log2 n loge n n log2 n log? n
601 9.231 22 6.398 59 651 9.346 51 6.478 51

602 9.233 62 6.400 26 652 9.348 73 6.480 04
603 9.236 01 6.401 92 653 9.350 94 6.481 58
604 9.238 40 6.403 57 654 9.353 15 6.483 11
605 9.240 79 6.405 23 655 9.355 35 6.484 64
606 9.243 17 6.406 88 656 9.357 55 6.486 16
607 9.245 55 6.408 53 657 9.359 75 6.487 68
608 9.247 93 6.410 17 658 9.361 94 6.489 20
609 9.250 30 6.411 82 659 9.364 13 6.490 72
610 9.252 67 6.413 46 660 9.366 32 6.492 24
611 9.255 03 6.415 10 661 9.368 51 6.493 75

612 9.257 39 6.416 73 662 9.370 69 6.495 27
613 9.259 74 6.418 36 663 9.372 87 6.496 77
614 9.262 09 6.419 99 664 9.375 04 6.498 28
615 9.264 44 6.421 62 665 9.377 21 6.499 79
616 9.266 79 6.423 25 666 9.379 38 6.501 29
617 9.269 13 6.424 87 667 9.381 54 6.502 79
618 9.271 46 6.426 49 668 9.383 70 6.504 29
619 9.273 80 6.428 11 669 9.385 86 6.505 78
620 9.276 12 6.429 72 670 9.388 02 6.507 28
621 9.278 45 6.431 33 671 9.390 17 6.508 77

622 9.280 77 6.432 94 672 9.392 32 6.510 26
623 9.283 09 6.434 55 673 9.394 46 6.511 75
624 9.285 40 6.436 15 674 9.396 60 6.513 23
625 9.287 71 6.437 75 675 9.398 74 6.514 71
626 9.290 02 6.439 35 676 9.400 88 6.516 19
627 9.292 32 6.440 95 677 9.403 01 6.517 67
628 9.294 62 6.442 54 678 9.405 14 6.519 15
629 9.296 92 6.444 13 679 9.407 27 6.520 62
630 9.299 21 6.445 72 680 9.409 39 6.522 09
631 9.301 50 6.447 31 681 9.411 51 6.523 56

632 9.303 78 6.448 89 682 9.413 63 6.525 03
633 9.306 06 6.450 47 683 9.415 74 6.526 49
634 9.308 34 6.452 05 684 9.417 85 6.527 96
635 9.310 61 6.453 62 685 9.419 96 6.529 42
636 9.312 88 6.455 20 686 9.422 06 6.530 88
637 9.315 15 6.456 77 687 9.424 17 6.532 33
638 9.317 41 6.458 34 688 9.426 26 6.533 79
639 9.319 67 6.459 90 689 9.428 36 6.535 24
640 9.321 93 6.461 47 690 9.430 45 6.536 69
641 9.324 18 6.463 03 691 9.432 54 6.538 14

642 9.326 43 6.464 59 692 9.434 63 6.539 59
643 9.328 67 6.466 14 693 9.436 71 6.541 03
644 9.330 92 6.467 70 694 9.438 79 6.542 47
645 9.333 16 6.469 25 695 9.440 87 6.543 91
646 9.335 39 6.470 80 696 9.442 94 6.545 35
647 9.337 62 6.472 35 697 9.445 01 6.546 79
648 9.339 85 6.473 89 698 9.447 08 6.548 22
649 9.342 07 6.475 43 699 9.449 15 6.549 65
650 9.344 30 6.476 97 700 9.451 21 6.551 08
436 TABLES
TABLE A (continued)
n log2 « loge n n log2 n loge «
701 9.453 27 6.552 51 751 9.552 67 6.621 41

702 9.455 33 6.553 93 752 9.554 59 6.622 74
703 9.457 38 6.555 36 753 9.556 51 6.624 07
704 9.459 43 6.556 78 754 9.558 42 6.625 39
705 9.461 48 6.558 20 755 9.560 33 6.626 72
706 9.463 52 6.559 62 756 9.562 24 6.628 04
707 9.465 57 6.561 03 757 9.564 15 6.629 36
708 9.467 61 6.562 44 758 9.566 05 6.630 68
709 9.469 64 6.563 86 759 9.567 96 6.632 00
710 9.471 68 6.565 26 760 9.569 86 6.633 32
711 9.473 71 6.566 67 761 9.571 75 6.634 63

712 9.475 73 6.568 08 762 9.573 65 6.635 95
713 9.477 76 6.569 48 763 9.575 54 6.637 26
714 9.479 78 6.570 88 764 9.577 43 6.638 57
715 9.481 80 6.572 28 765 9.579 32 6.639 88
716 9.483 82 6.573 68 766 9.581 20 6.641 18
717 9.485 83 6.575 08 767 9.583 08 6.642 49
718 9.487 84 6.576 47 768 9.584 96 6.643 79
719 9.489 85 6.577 86 769 9.586 84 6.645 09
720 9.491 85 6.579 25 770 9.588 71 6.646 39
721 9.493 86 6.580 64 771 9.590 59 6.647 69

722 9.495 86 6.582 03 772 9.592 46 6.648 98
723 9.497 85 6.583 41 773 9.594 32 6.650 28
724 9.499 85 6.584 79 774 9.596 19 6.651 57
725 9.501 84 6.586 17 775 9.598 05 6.652 86
726 9.503 83 6.587 55 776 9.599 91 6.654 15
727 9.505 81 6.588 93 777 9.601 77 6.655 44
728 9.507 79 6.590 30 778 9.603 63 6.656 73
729 9.509 77 6.591 67 779 9.605 48 6.658 01
730 9.511 75 6.593 04 780 9.607 33 6.659 29
731 9.513 73 6.594 41 781 9.609 18 6.660 58

732 9.515 70 6.595 78 782 9.611 02 6.661 85
733 9.517 67 6.597 15 783 9.612 87 6.663 13
734 9.519 64 6.598 51 784 9.614 71 6.664 41
735 9.521 60 6.599 87 785 9.616 55 6.665 68
736 9.523 56 6.601 23 786 9.618 39 6.666 96
737 9.525 52 6.602 59 787 9.620 22 6.668 23
738 9.527 48 6.603 94 788 9.622 05 6.669 50
739 9.529 43 6.605 30 789 9.623 88 6.670 77
740 9.531 38 6.606 65 790 9.625 71 6.672 03
741 9.533 33 6.608 00 791 9.627 53 6.673 30
742 9.535 28 6.609 35 792 9.629 36 6.674 56
743 9.537 22 6.610 70 793 9.631 18 6.675 82
744 9.539 16 6.612 04 794 9.633 00 6.677 08
745 9.541 10 6.613 38 795 9.634 81 6.678 34
746 9.543 03 6.614 73 796 9.636 62 6.679 60
747 9.544 96 6.616 07 797 9.638 44 6.680 85
748 9.546 89 6.617 40 798 9.640 24 6.682 11
749 9.548 82 6.618 74 799 9.642 05 6.683 36
750 9.550 75 6.620 07 800 9.643 86 6.684 61
TABLES 437
TABLE A (continued)
801 9.645 66 6.685 86 851 9.733 02 6.746 41

802 9.647 46 6.687 11 852 9.734 71 6.747 59
803 9.649 26 6.688 35 853 9.736 40 6.748 76
804 9.651 05 6.689 60 854 9.738 09 6.749 93
805 9.652 84 6.690 84 855 9.739 78 6.751 10
806 9.654 64 6.692 08 856 9.741 47 6.752 27
807 9.656 42 6.693 32 857 9.743 15 6.753 44
808 9.658 21 6.694 56 858 9.744 83 6.754 60
809 9.660 00 6.695 80 859 9.746 51 6.755 77
810 9.661 78 6.697 03 860 9.748 19 6.756 93
811 9.663 56 6.698 27 861 9.749 87 6.758 09

812 9.665 34 6.699 50 862 9.751 54 6.759 26
813 9.667 11 6.700 73 863 9.753 22 6.760 41
814 9.668 88 6.701 96 864 9.754 89 6.761 57
815 9.670 66 6.703 19 865 9.756 56 6.762 73
816 9.672 43 6.704 41 866 9.758 22 6.763 88
817 9.674 19 6.705 64 867 9.759 89 6.765 04
818 9.675 96 6.706 86 868 9.761 55 6.766 19
819 9.677 72 6.708 08 869 9.763 21 6.767 34
820 9.679 48 6.709 30 870 9.764 87 6.768 49
821 9.681 24 6.710 52 871 9.766 53 6.769 64

822 9.682 99 6.711 74 872 9.768 18 6.770 79
823 9.684 75 6.712 96 873 9.769 84 6.771 94
824 9.686 50 6.714 17 874 9.771 49 6.773 08
825 9.688 25 6.715 38 875 9.773 14 6.774 22
826 9.690 00 6.716 59 876 9.774 79 6.775 37
827 9.691 74 6.717 80 877 9.776 43 6.776 51
828 9.693 49 6.719 01 878 9.778 08 6.777 65
829 9.695 23 6.720 22 879 9.779 72 6.778 78
830 9.696 97 6.721 43 880 9.781 36 6.779 92
831 9.698 70 6.722 63 881 9.783 00 6.781 06

832 9.700 44 6.723 83 882 9.784 63 6.782 19
833 9.702 17 6.725 03 883 9.786 27 6.783 33
834 9.703 90 6.726 23 884 9.787 90 6.784 46
835 9.705 63 6.727 43 885 9.789 53 6.785 59
836 9.707 36 6.728 63 886 9.791 16 6.786 72
837 9.709 08 6.729 82 887 9.792 79 6.787 84
838 9.710 81 6.731 02 888 9.794 42 6.788 97
839 9.712 53 6.732 21 889 9.796 04 6.790 10
840 9.714 25 6.733 40 890 9.797 66 6.791 22
841 9.715 96 6.734 59 891 9.799 28 6.792 34

842 9.717 68 6.735 78 892 9.800 90 6.793 47
843 9.719 39 6.736 97 893 9.802 52 6.794 59
844 9.721 10 6.738 15 894 9.804 13 6.795 71
845 9.722 81 6.739 34 895 9.805 74 6.796 82
846 9.724 51 6.740 52 896 9.807 35 6.797 94
847 9.726 22 6.741 70 897 9.808 96 6.799 06
848 9.727 92 6.742 88 898 9.810 57 6.800 17
849 9.729 62 6.744 06 899 9.812 18 6.801 28
850 9.731 32 6.745 24 900 9.813 78 6.802 39
438 TABLES
TABLE A (continued)
901 9.815 38 6.803 51 951 9.893 30 6.857 51

902 9.816 98 6.804 61 952 9.894 82 6.858 57
903 9.818 58 6.805 72 953 9.896 33 6.859 61
904 9.820 18 6.806 83 954 9.897 85 6.860 66
905 9.821 77 6.807 93 955 9.899 36 6.861 71
906 9.823 37 6.809 04 956 9.900 87 6.862 76
907 9.824 96 6.810 14 957 9.902 38 6.863 80
908 9.826 55 6.811 24 958 9.903 88 6.864 85
909 9.828 14 6.812 35 959 9.905 39 6.865 89
910 9.829 72 6.813 44 960 9.906 89 6.866 93
911 9.831 31 6.814 54 961 9.908 39 6.867 97

912 9.832 89 6.815 64 962 9.909 89 6.869 01
913 9.834 47 6.816 74 963 9.911 39 6.870 05
914 9.836 05 6.817 83 964 9.912 89 6.871 09
915 9.837 63 6.818 92 965 9.914 39 6.872 13
916 9.839 20 6.820 02 966 9.915 88 6.873 16
917 9.840 78 6.821 11 967 9.917 37 6.874 20
918 9.842 35 6.822 20 968 9.918 86 6.875 23
919 9.843 92 6.823 29 969 9.920 35 6.876 26
920 9.845 49 6.824 37 970 9.921 84 6.877 30
921 9.847 06 6.825 46 971 9.923 33 6.878 33

922 9.848 62 6.826 55 972 9.924 81 6.879 36
923 9.850 19 6.827 63 973 9.926 30 6.880 38
924 9.851 75 6.828 71 974 9.927 78 6.881 41
925 9.853 31 6.829 79 975 9.929 26 6.882 44
926 9.854 87 6.830 87 976 9.930 74 6.883 46
927 9.856 43 6.831 95 977 9.932 21 6.884 49
928 9.857 98 6.833 03 978 9.933 69 6.885 51
929 9.859 53 6.834 11 979 9.935 17 6.886 53
930 9.861 09 6.835 18 980 9.936 64 6.887 55
931 9.862 64 6.836 26 981 9.938 11 6.888 57

932 9.864 19 6.837 33 982 9.939 58 6.889 59
933 9.865 73 6.838 41 983 9.941 05 6.890 61
934 9.867 28 6.839 48 984 9.942 51 6.891 63
935 9.868 82 6.840 55 985 9.943 98 6.892 64
936 9.870 36 6.841 62 986 9.945 44 6.893 66
937 9.871 91 6.842 68 987 9.946 91 6.894 67
938 9.873 44 6.843 75 988 9.948 37 6.895 68
939 9.874 98 6.844 82 989 9.949 83 6.896 69
940 9.876 52 6.845 88 990 9.951 28 6.897 70
941 9.878 05 6.846 94 991 9.952 74 6.898 71
942 9.879 58 6.848 01 992 9.954 20 6.899 72
943 9.881 11 6.849 07 993 9.955 65 6.900 73
944 9.882 64 6.850 13 994 9.957 10 6.901 74
945 9.884 17 6.851 18 995 9.958 55 6.902 74
946 9.885 70 6.852 24 996 9.960 00 6.903 75
947 9.887 22 6.853 30 997 9.961 45 6.904 75
948 9.888 74 6.854 35 998 9.962 90 6.905 75
949 9.890 26 6.855 41 999 9.964 34 6.906 75
950 9.891 78 6.856 46 1 000 9.965 78 6.907 76
TABLES 439
TABLE A (continued)
n log2 n loge n n loga n loge n
1 010 9.980 14 6.917 71 1 510 10.560 3 7.319 86

1 020 9.994 35 6.927 56 1 520 10.569 9 7.326 47
1 030 10.008 4 6.937 31 1 530 10.579 3 7.333 02
1 040 10.022 4 6.946 98 1 540 10.588 7 7.339 54
1 050 10.036 2 6.956 55 1 550 10.598 1 7.346 01
1 060 10.049 8 6.966 02 1 560 10.607 3 7.352 44
1 070 10.063 4 6.975 41 1 570 10.616 5 7.358 83
1 080 10.076 8 6.984 72 l 580 10.625 7 7.365 18
1 090 10.090 1 6.993 93 1 590 10.634 8 7.371 49
1 100 10.103 3 7.003 07 1 600 10.643 9 7.377 76
1 110 10.116 3 7.012 12 1 610 10.652 8 7.383 99

1 120 10.129 3 7.021 08 1 620 10.661 8 7.390 18
1 130 10.142 1 7.029 97 1 630 10.670 7 7.396 34
1 140 10.154 8 7.038 78 1 640 10.679 5 7.402 45
1 150 10.167 4 7.047 52 1 650 10.688 3 7.408 53
1 160 10.179 9 7.056 18 1 660 10.697 0 7.414 57
1 170 10.192 3 7.064 76 1 670 10.705 6 7.420 58
1 180 10.204 6 7.073 27 1 680 10.714 2 7.426 55
1 190 10.216 7 7.081 71 1 690 10.722 8 7.432 48
1 200 10.228 8 7.090 08 1 700 10.731 3 7.438 38
1 210 10.240 8 7.098 38 1 710 10.739 8 7.444 25

1 220 10.252 7 7.106 61 1 720 10.748 2 7.450 08
1 230 10.264 4 7.114 77 1 730 10.756 6 7.455 88
1 240 10.276 1 7.122 87 1 740 10.764 9 7.461 64
1 250 10.287 7 7.130 90 1 750 10.773 1 7.467 37
1 260 10.299 2 7.138 87 1 760 10.781 4 7.473 07
1 270 10.310 6 7.146 77 1 770 10.789 5 7.478 73
1 280 10.321 9 7.154 62 1 780 10.797 7 7.484 37
1 290 10.333 2 7.162 40 1 790 10.805 7 7.489 97
1 300 10.344 3 7.170 12 1 800 10.813 8 7.495 54
1 310 10.355 4 7.177 78 1 810 10.821 8 7.501 08

1 320 10.366 3 7.185 39 1 820 10.829 7 7.506 59
1 330 10.377 2 7.192 93 1 830 10.837 6 7.512 07
1 340 10.388 0 7.200 42 1 840 10.845 5 7.517 52
1 350 10.398 7 7.207 86 1 850 10.853 3 7.522 94
1 360 10.409 4 7.215 24 1 860 10.861 1 7.528 33
1 370 10.420 0 7.222 57 1 870 10.868 8 7.533 69
1 380 10.430 5 7.229 84 1 880 10.876 5 7.539 03
1 390 10.440 9 7.237 06 1 890 10.884 2 7.544 33
1 400 10.451 2 7.244 23 1 900 10.891 8 7.549 61
1 410 10.461 5 7.251 34 1 910 10.899 4 7.554 86

1 420 10.471 7 7.258 41 1 920 10.906 9 7.560 08
1 430 10.481 8 7.265 43 1 930 10.914 4 7.565 28
1 440 10.491 9 7.272 40 1 940 10.921 8 7.570 44
1 450 10.501 8 7.279 32 1 950 10.929 3 7.575 58
1 460 10.511 8 7.286 19 1 960 10.936 6 7.580 70
1 470 10.521 6 7.293 02 1 970 10.944 0 7.585 79
1 480 10.531 4 7.299 80 1 980 10.951 3 7.590 85
1 490 10.541 1 7.306 53 1 990 10.958 6 7.595 89
1 500 10.550 7 7.313 22 2 000 10.965 8 7.600 90
440 TABLES
TABLE A (continued)
2010 10.973 0 7.605 89 2510 11.293 5 7.828 04

2 020 10.980 1 7.610 85 2 520 11.299 2 7.832 01
2 030 10.987 3 7.615 79 2 530 11.304 9 7.835 97
2 040 10.994 4 7.620 71 2 540 11.3106 7.839 92
2 050 11.001 4 7.625 60 2 550 11.3163 7.843 85
2 060 11.008 4 7.630 46 2 560 11.321 9 7.847 76
2 070 11.015 4 7.635 30 2 570 11.327 6 7.851 66
2 080 11.022 4 7.640 12 2 580 11.333 2 7.855 54
2 090 11.029 3 7.644 92 2 590 11.338 7 7.859 41
2 100 11.036 2 7.649 69 2 600 11.3443 7.863 27
2 110 11.043 0 7.654 44 2 610 11.349 8 7.867 11

2 120 11.049 8 7.659 17 2 620 11.355 4 7.870 93
2 130 11.056 6 7.663 88 2 630 11.360 8 7.874 74
2 140 11.063 4 7.668 56 2 640 11.366 3 7.878 53
2 150 11.070 1 7.673 22 2 650 11.371 8 7.882 31
2 160 11.076 8 7.677 86 2 660 11.377 2 7.886 08
2 170 11.083 5 7.682 48 2 670 11.382 6 7.889 83
2 180 11.090 1 7.687 08 2 680 11.388 0 7.893 57
2 190 11.096 7 7.691 66 2 690 11.393 4 7.897 30
2 200 11.103 3 7.696 21 2 700 11.398 7 7.901 01
2 210 11.109 8 7.700 75 2 710 11.404 1 7.904 70

2 220 11.116 3 7.705 26 2 720 11.409 4 7.908 39
2 230 11.1228 7.709 76 2 730 11.414 7 7.912 06
2 240 11.129 3 7.714 23 2 740 11.420 0 7.915 71
2 250 11.135 7 7.718 69 2 750 11.425 2 7.919 36
2 260 11.142 1 7.723 12 2 760 11.430 5 7.922 99
2 270 11.148 5 7.727 54 2 770 11.435 7 7.926 60
2 280 11.1548 7.731 93 2 780 11.440 9 7.930 21
2 290 11.161 1 7.736 31 2 790 11.446 0 7.933 80
2 300 11.1674 7.740 66 2 800 11.451 2 7.937 37
2310 11.173 7 7.745 00 2 810 11.456 4 7.940 94

2 320 11.179 9 7.749 32 2 820 11.461 5 7.944 49
2 330 11.186 1 7.753 62 2 830 11.466 6 7.948 03
2 340 11.192 3 7.757 91 2 840 11.471 7 7.951 56
2 350 11.1984 7.762 17 2 850 11.476 7 7.955 07
2 360 11.204 6 7.766 42 2 860 11.481 8 7.958 58
2 370 11.2107 7.770 65 2 870 11.486 8 7.962 07
2 380 11.216 7 7.774 86 2 880 11.491 9 7.965 55
2 390 11.222 8 7.779 05 2 890 11.496 9 7.969 01
2 400 11.228 8 7.783 22 2 900 11.501 8 7.972 47
2 410 11.234 8 7.787 38 2 910 11.506 8 7.975 91
2 420 11.240 8 7.791 52 2 920 11.511 8 7.979 34
2 430 11.246 7 7.795 65 2 930 11.516 7 7.982 76
2 440 11.252 7 7.799 75 2 940 11.521 6 7.986 16
2 450 11.258 6 7.803 84 2 950 11.526 5 7.989 56
2 460 11.264 4 7.807 92 2 960 11.531 4 7.992 94
2 470 11.270 3 7.811 97 2 970 11.536 2 7.996 32
2 480 11.276 1 7.816 01 2 980 11.541 1 7.999 68
2 490 11.281 9 7.820 04 2 990 11.545 9 8.003 03
2 500 11.287 7 7.824 05 3 000 11.550 7 8.006 37
TABLES 441
TABLE A (continued)
n log2 n log? n n log2 n loge n
3 010 11.555 5 8.009 70 3 510 11.777 3 8.163 37

3 020 11.560 3 8.013 01 3 520 11.781 4 8.166 22
3 030 11.565 1 8.016 32 3 530 11.785 5 8.169 05
3 040 11.569 9 8.019 61 3 540 11.789 5 8.171 88
3 050 11.574 6 8.022 90 3 550 11.793 6 8.174 70
3 060 11.579 3 8.026 17 3 560 11.797 7 8.177 52
3 070 11.5840 8.029 43 3 570 11.801 7 8.180 32
3 080 11.588 7 8.032 68 3 580 11.805 7 8.183 12
3 090 11.593 4 8.035 93 3 590 11.809 8 8.185 91
3 100 11.598 1 8.039 16 3 600 11.813 8 8.188 69
3 110 11.602 7 8.042 38 3 610 11.817 8 8.191 46

3 120 11.607 3 8.045 59 3 620 11.821 8 8.194 23
3 130 11.611 9 8.048 79 3 630 11.825 8 8.196 99
3 140 11.616 5 8.051 98 3 640 11.829 7 8.199 74
3 150 11.621 1 8.055 16 3 650 11.833 7 8.202 48
3 160 11.625 7 8.058 33 3 660 11.837 6 8.205 22
3 170 11.630 3 8.061 49 3 670 11.841 6 8.207 95
3 180 11.634 8 8.064 64 3 680 11.845 5 8.210 67
3 190 11.639 3 8.067 78 3 690 11.849 4 8.213 38
3 200 11.643 9 8.070 91 3 700 11.853 3 8.216 09
3 210 11.648 4 8.074 03 3 710 11.857 2 8.218 79

3 220 11.652 8 8.077 14 3 720 11.861 1 8.221 48
3 230 11.657 3 8.080 24 3 730 11.865 0 8.224 16
3 240 11.661 8 8.083 33 3 740 11.868 8 8.226 84
3 250 11.666 2 8.086 41 3 750 11.8727 8.229 51
3 260 11.670 7 8.089 48 3 760 11.876 5 8.232 17
3 270 11.675 1 8.092 55 3 770 11.880 3 8.234 83
3 280 11.679 5 8.095 60 3 780 11.884 2 8.237 48
3 290 11.683 9 8.098 64 3 790 11.888 0 8.240 12
3 300 11.688 3 8.101 68 3 800 11.891 8 8.242 76
3 310 11.692 6 8.104 70 3 810 11.895 6 8.245 38

3 320 11.697 0 8.107 72 3 820 11.899 4 8.248 01
3 330 11.701 3 8.11073 3 830 11.903 1 8.250 62
3 340 11.705 6 8.113 73 3 840 11.906 9 8.253 23
3 350 11.709 9 8.116 72 3 850 11.9106 8.255 83
3 360 11.7142 8.119 70 3 860 11.9144 8.258 42
3 370 11.718 5 8.122 67 3 870 11.918 1 8.261 01
3 380 11.722 8 8.125 63 3 880 11.921 8 8.263 59
3 390 11.727 1 8.128 59 3 890 11.925 6 8.266 16
3 400 11.731 3 8.131 53 3 900 11.929 3 8.268 73
3 410 11.735 6 8.134 47 3 910 11.933 0 8.271 29

3 420 11.739 8 8.137 40 3 920 11.936 6 8.273 85
3 430 11.7440 8.140 32 3 930 11.940 3 8.276 39
3 440 11.748 2 8.143 23 3 940 11.944 0 8.278 94
3 450 11.752 4 8.146 13 3 950 11.947 6 8.281 47
3 460 11.756 6 8.149 02 3 960 11.951 3 8.284 00
3 470 11.760 7 8.151 91 3 970 11.954 9 8.286 52
3 480 11.7649 8.154 79 3 980 11.958 6 8.289 04
3 490 11.769 0 8.157 66 3 990 11.962 2 8.291 55
3 500 11.773 1 8.160 52 4 000 11.965 8 8.294 05
442 TABLES
TABLE A (continued)
n log2 n log« n n log2 n logs n
4 010 11.969 4 8.296 55 4 510 12.138 9 8.414 05

4 020 11.973 0 8.299 04 4 520 12.142 1 8.416 27
4 030 11.976 6 8.301 52 4 530 12.145 3 8.418 48
4 040 11.980 1 8.304 00 4 540 12.148 5 8.420 68
4 050 11.983 7 8.306 47 4 550 12.151 7 8.422 88
4 060 11.987 3 8.308 94 4 560 12.154 8 8.425 08
4 070 11.990 8 8.311 40 4 570 12.158 0 8.427 27
4 080 11.994 4 8.313 85 4 580 12.161 1 8.429 45
4 090 11.997 9 8.316 30 4 590 12.164 3 8.431 64
4 100 12.001 4 8.318 74 4 600 12.167 4 8.433 81
4 110 12.004 9 8.321 18 4 610 12.170 6 8.435 98

4 120 12.008 4 8.323 61 4 620 12.173 7 8.438 15
4 130 12.011 9 8.326 03 4 630 12.176 8 8.440 31
4 140 12.015 4 8.328 45 4 640 12.179 9 8.442 47
4 150 12.018 9 8.330 86 4 650 12.183 0 8.444 62
4 160 12.022 4 8.333 27 4 660 12.186 1 8.446 77
4 170 12.025 8 8.335 67 4 670 12.189 2 8.448 91
4180 12.029 3 8.338 07 4 680 12.192 3 8.451 05
4 190 12.032 7 8.340 46 4 690 12.195 4 8.453 19
4 200 12.036 2 8.342 84 4 700 12.198 4 8.455 32
4 210 12.039 6 8.345 22 4 710 12.201 5 8.457 44

4 220 12.043 0 8.347 59 4 720 12.204 6 8.459 56
4 230 12.046 4 8.349 96 4 730 12.207 6 8.461 68
4 240 12.049 8 8.352 32 4 740 12.210 7 8.463 79
4 250 12.053 2 8.354 67 4 750 12.213 7 8.465 90
4 260 12.056 6 8.357 02 4 760 12.216 7 8.468 00
4 270 12.060 0 8.359 37 4 770 12.219 8 8.470 10
4 280 12.063 4 8.361 71 4 780 12.222 8 8.472 20
4 290 12.066 8 8.364 04 4 790 12.225 8 8.474 29
4 300 12.070 1 8.366 37 4 800 12.228 8 8.476 37
4 310 12.073 5 8.368 69 4 810 12.231 8 8.478 45

4 320 12.076 8 8.371 01 4 820 12.234 8 8.480 53
4 330 12.080 2 8.373 32 4 830 12.237 8 8.482 60
4 340 12.083 5 8.375 63 4 840 12.240 8 8.484 67
4 350 12.086 8 8.377 93 4 850 12.243 8 8.486 73
4 360 12.090 1 8.380 23 4 860 12.246 7 8.488 79
4 370 12.093 4 8.382 52 4 870 12.249 7 8.490 85
4 380 12.096 7 8.384 80 4 880 12.252 7 8.492 90
4 390 12.100 0 8.387 08 4 890 12.255 6 8.494 95
4 400 12.103 3 8.389 36 4 900 12.258 6 8.496 99
4410 12.106 6 8.391 63 4910 12.261 5 8.499 03
4 420 12.109 8 8.393 89 4 920 12.264 4 8.501 06
4 430 12.113 1 8.396 15 4 930 12.267 4 8.503 09
4 440 12.116 3 8.398 41 4 940 12.270 3 8.505 12
4 450 12.119 6 8.400 66 4 950 12.273 2 8.507 14
4 460 12.122 8 8.402 90 4 960 12.276 1 8.509 16
4 470 12.126 1 8.405 14 4 970 12.279 0 8.511 18
4 480 12.129 3 8.407 38 4 980 12.281 9 8.513 19
4 490 12.132 5 8.409 61 4 990 12.284 8 8.515 19
4 500 12.135 7 8.411 83 5 000 12.287 7 8.517 19
TABLES 443
TABLE A (continued)
n log-2 n loge n n log-2 n loge n
5 010 12.290 6 8.519 19 5 510 12.427 8 8.614 32

5 020 12.293 5 8.521 19 5 520 12.430 5 8.616 13
5 030 12.296 3 8.523 18 5 530 12.433 1 8.617 94
5 040 12.299 2 8.525 16 5 540 12.435 7 8.619 75
5 050 12.302 1 8.527 14 5 550 12.438 3 8.621 55
5 060 12.304 9 8.529 12 5 560 12.440 9 8.623 35
5 070 12.307 8 8.531 10 5 570 12.443 5 8.625 15
5 080 12.3106 8.533 07 5 580 12.446 0 8.626 94
5 090 12.313 4 8.535 03 5 590 12.448 6 8.628 73
5 100 12.316 3 8.537 00 5 600 12.451 2 8.630 52
5 110 12.319 1 8.538 95 5 610 12.453 8 8.632 31

5 120 12.321 9 8.540 91 5 620 12.456 4 8.634 09
5 130 12.324 7 8.542 86 5 630 12.458 9 8.635 86
5 140 12.327 6 8.544 81 5 640 12.461 5 8.637 64
5 150 12.330 4 8.546 75 5 650 12.464 0 8.639 41
5 160 12.333 2 8.548 69 5 660 12.466 6 8.641 18
5 170 12.335 9 8.550 63 5 670 12.469 1 8.642 94
5 180 12.338 7 8.552 56 5 680 12.471 7 8.644 71
5 190 12.341 5 8.554 49 5 690 12.474 2 8.646 47
5 200 12.344 3 8.556 41 5 700 12.476 7 8.648 22
5 210 12.347 1 8.558 34 5 710 12.479 3 8.649 97

5 220 12.349 8 8.560 25 5 720 12.481 8 8.651 72
5 230 12.352 6 8.562 17 5 730 12.484 3 8.653 47
5 240 12.355 4 8.564 08 5 740 12.486 8 8.655 21
5 250 12.358 1 8.565 98 5 750 12.489 3 8.656 96
5 260 12.360 8 8.567 89 5 760 12.491 9 8.658 69
5 270 12.363 6 8.569 79 5 770 12.494 4 8.660 43
5 280 12.366 3 8.571 68 5 780 12.496 9 8.662 16
5 290 12.369 1 8.573 57 5 790 12.499 3 8.663 89
5 300 12.371 8 8.575 46 5 800 12.501 8 8 665 61
5 310 12.374 5 8.577 35 5 810 12.504 3 8.667 34

5 320 12.377 2 8.579 23 5 820 12.506 8 8.669 06
5 330 12.379 9 8.581 11 5 830 12.509 3 8.670 77
5 340 12.382 6 8.582 98 5 840 12.511 8 8.672 49
5 350 12.385 3 8.584 85 5 850 12.5142 8.674 20
5 360 12.388 0 8.586 72 5 860 12.516 7 8.675 90
5 370 12.390 7 8.588 58 5 870 12.519 1 8.677 61
5 380 12.393 4 8.590 44 5 880 12.521 6 8.679 31
5 390 12.396 1 8.592 30 5 890 12.524 1 8.681 01
5 400 12.398 7 8.594 15 5 900 12.526 5 8.682 71
5 410 12.401 4 8.596 00 5 910 12.528 9 8.684 40

5 420 12.404 1 8.597 85 5 920 12.531 4 8.686 09
5 430 12.406 7 8.599 69 5 930 12.533 8 8.687 78
5 440 12.409 4 8.601 53 5 940 12.536 2 8.689 46
5 450 12.412 0 8.603 37 5 950 12.538 7 8.691 15
5 460 12.414 7 8.605 20 5 960 12.541 1 8.692 83
5 470 12.417 3 8.607 03 5 970 12.543 5 8.694 50
5 480 12.420 0 8.608 86 5 980 12.545 9 8.696 18
5 490 12.422 6 8.610 68 5 990 12.548 3 8.697 85
5 500 12.425 2 8.612 50 6 000 12.550 7 8.699 51
444 TABLES
TABLE A (continued)
n log2 n logc n n log2 n loge n
6 010 12.553 1 8.701 18 6 510 12.668 4 8.781 09

6 020 12.555 5 8.702 84 6 520 12.670 7 8.782 63
6 030 12.557 9 8.704 50 6 530 12.672 9 8.784 16
6 040 12.560 3 8.706 16 6 540 12.675 1 8.785 69
6 050 12.562 7 8.707 81 6 550 12.677 3 8.787 22
6 060 12.565 1 8.709 47 6 560 12.679 5 8.788 75
6 070 12.567 5 8.711 11 6 570 12.681 7 8.790 27
6 080 12.569 9 8.712 76 6 580 12.683 9 8.791 79
6 090 12.572 2 8.714 40 6 590 12.686 1 8.793 31
6 100 12.574 6 8.716 04 6 600 12.688 3 8.794 82
6 110 12.577 0 8.717 68 6 610 12.690 4 8.796 34

6 120 12.579 3 8.719 32 6 620 12.692 6 8.797 85
6 130 12.581 7 8.720 95 6 630 12.694 8 8.799 36
6 140 12.584 0 8.722 58 6 640 12.697 0 8.800 87
6 150 12.586 4 8.724 21 6 650 12.699 1 8.802 37
6 160 12.588 7 8.725 83 6 660 12.701 3 8.803 87
6 170 12.591 1 8.727 45 6 670 12.703 5 8.805 38
6 180 12.593 4 8.729 07 6 680 12.705 6 8.806 87
6 190 12.595 7 8.730 69 6 690 12.707 8 8.808 37
6 200 12.598 1 8.732 30 6 700 12.709 9 8.809 86
6210 12.600 4 8.733 92 6 710 12.712 1 8.811 35

6 220 12.602 7 8.735 53 6 720 12.7142 8.812 84
6 230 12.605 0 8.737 13 6 730 12.7164 8.814 33
6 240 12.607 3 8.738 74 6 740 12.718 5 8.815 82
6 250 12.609 6 8.740 34 6 750 12.720 7 8.817 30
6 260 12.611 9 8.741 94 6 760 12.722 8 8.818 78
6 270 12.614 2 8.743 53 6 770 12.724 9 8.820 26
6 280 12.616 5 8.745 13 6 780 12.727 1 8.821 73
6 290 12.618 8 8.746 72 6 790 12.729 2 8.823 21
6 300 12.621 1 8.748 30 6 800 12.731 3 8.824 68
6310 12.623 4 8.749 89 6 810 12.733 4 8.826 15

6 320 12.625 7 8.751 47 6 820 12.735 6 8.827 61
6 330 12.628 0 8.753 06 6 830 12.737 7 8.829 08
6 340 12.630 3 8.754 63 6 840 12.739 8 8.830 54
6 350 12.632 5 8.756 21 6 850 12.741 9 8.832 00
6 360 12.634 8 8.757 78 6 860 12.744 0 8.833 46
6 370 12.637 1 8.759 35 6 870 12.746 1 8.834 92
6 380 12.639 3 8.760 92 6 880 12.748 2 8.836 37
6 390 12.641 6 8.762 49 6 890 12.750 3 8.837 83
6 400 12.643 9 8.764 05 6 900 12.752 4 8.839 28
6410 12.646 1 8.765 61 i 6 910 12.754 5 8.840 72
6 420 12.648 4 8.767 17 6 920 12.756 6 8.842 17
6 430 12.650 6 8.768 73 6 930 12.758 6 8.843 62
6.440 12.652 8 8.770 28 6 940 12.760 7 8.845 06
6 450 12.655 1 8.771 84 6 950 12.762 8 8.846 50
6 460 12.657 3 8.773 38 6 960 12.764 9 8.847 93
6 470 12.659 5 8.774 93 6 970 12.766 9 8.849 37
6 480 12.661 8 8.776 48 6 980 12.769 0 8.850 80
6 490 12.664 0 8.778 02 6 990 12.771 1 8.852 24
6 500 12.666 2 8.779 56 7 000 12.773 1 8.853 67
TABLES 445
TABLE A (continued)
n log2 « loge« n log2 /7 loge n
7 010 12.775 2 8.855 09 7 510 12.874 6 8.923 99

7 020 12.777 3 8.856 52 7 520 12.876 5 8.925 32
7 030 12.779 3 8.857 94 7 530 12.878 4 8.926 65
7 040 12.781 4 8.859 36 7 540 12.880 3 8.927 98
7 050 12.783 4 8.860 78 7 550 12.882 3 8.929 30
7 060 12.785 5 8.862 20 7 560 12.884 2 8.930 63
7 070 12.787 5 8.863 62 7 570 12.886 1 8.931 95
7 080 12.789 5 8.865 03 7 580 12.888 0 8.933 27
7 090 12.791 6 8.866 44 7 590 12.889 9 8.934 59
7 100 12.793 6 8.867 85 7 600 12.891 8 8.935 90
7 110 12.795 6 8.869 26 7 610 12.893 7 8.937 22

7 120 12.797 7 8.870 66 7 620 12.895 6 8.938 53
7 130 12.799 7 8.872 07 7 630 12.897 5 8.939 84
7 140 12.801 7 8.873 47 7 640 12.899 4 8.941 15
7 150 12.803 7 8.874 87 7 650 12.901 2 8.942 46
7 160 12.805 7 8.876 27 7 660 12.903 1 8.943 77
7 170 12.807 8 8.877 66 7 670 12.905 0 8.945 07
7 180 12.809 8 8.879 05 7 680 12.906 9 8.946 37
7 190 12.811 8 8.880 45 7 690 12.908 8 8.947 68
7 200 12.813 8 8.881 84 7 700 12.910 6 8.948 98
7 210 12.815 8 8.883 22 7 710 12.912 5 8.950 27

7 220 12.817 8 8.884 61 7 720 12.914 4 8.951 57
7 230 12.819 8 8.885 99 7 730 12.916 3 8.952 86
7 240 12.821 8 8.887 38 7 740 12.918 1 8.954 16
7 250 12.823 8 8.888 76 7 750 12.920 0 8.955 45
7 260 12.825 8 8.890 14 7 760 12.921 8 8.956 74
7 270 12.827 7 8.891 51 7 770 12.923 7 8.958 03
7 280 12.829 7 8.892 89 7 780 12.925 6 8.959 31
7 290 12.831 7 8.894 26 7 790 12.927 4 8.960 60
7 300 12.833 7 8.895 63 7 800 12.929 3 8.961 88
7 310 12.835 7 8.897 00 7 810 12.931 1 8.963 16

7 320 12.837 6 8.898 37 7 820 12.933 0 8.964 44
7 330 12.839 6 8.899 73 7 830 12.934 8 8.965 72
7 340 12.841 6 8.901 09 7 840 12.936 6 8.966 99
7 350 12.843 5 8.902 46 7 850 12.938 5 8.968 27
7 360 12.845 5 8.903 82 7 860 12.940 3 8.969 54
7 370 12.847 4 8.905 17 7 870 12.942 1 8.970 81
7 380 12.849 4 8.906 53 7 880 12.944 0 8.972 08
7 390 12.851 4 8.907 88 7 890 12.945 8 8.973 35
7 400 12.853 3 8.909 24 7 900 12.947 6 8.974 62
7 410 12.855 3 8.910 59 7 910 12.949 5 8.975 88

7 420 12.857 2 8.911 93 7 920 12.951 3 8.977 15
7 430 12.859 1 8.913 28 7 930 12.953 1 8.978 41
7 440 12.861 1 8.914 63 7 940 12.954 9 8.979 67
7 450 12.863 0 8.915 97 7 950 12.956 7 8.980 93
7 460 12.865 0 8.917 31 7 960 12.958 6 8.982 18
7 470 12.866 9 8.918 65 7 970 12.960 4 8.983 44
7 480 12.868 8 8.919 99 7 980 12.962 2 8.984 69
7 490 12.870 7 8.921 32 7 990 12.964 0 8.985 95
7 500 12.872 7 8.922 66 8 000 12.965 8 8.987 20
446 TABLES
TABLE A (continued)
n log2 n \oge n n log2 n loge n
8010 12.967 6 8.988 45 8 510 13.054 9 9.049 00

8 020 12.969 4 8.989 69 8 520 13.056 6 9.050 17
8 030 12.971 2 8.990 94 8 530 13.058 3 9.051 34
8 040 12.973 0 8.992 18 8 540 13.060 0 9.052 52
8 050 12.974 8 8.993 43 8 550 13.061 7 9.053 69
8 060 12.976 6 8.994 67 8 560 13.063 4 9.054 86
8 070 12.978 4 8.995 91 8 570 13.065 1 9.056 02
8 080 12.980 1 8.997 15 8 580 13.066 8 9.057 19
8 090 12.981 9 8.998 38 8 590 13.068 4 9.058 35
8 100 12.983 7 8.999 62 8 600 13.070 1 9.059 52
8 110 12.985 5 9.000 85 8 610 13.071 8 9.060 68

8 120 12.987 3 9.002 09 8 620 13.073 5 9.061 84
8 130 12.989 0 9.003 32 8 630 13.075 1 9.063 00
8 140 12.990 8 9.004 55 8 640 13.076 8 9.064 16
8 150 12.992 6 9.005 77 8 650 13.078 5 9.065 31
8 160 12.994 4 9.007 00 8 660 13.080 2 9.066 47
8 170 12.996 1 9.008 22 8 670 13.081 8 9.067 62
8 180 12.997 9 9.009 45 8 680 13.083 5 9.068 78
8 190 12.999 6 9.010 67 8 690 13.085 1 9.069 93
8 200 13.001 4 9.011 89 8 700 13.086 8 9.071 08
8 210 13.003 2 9.013 11 8 710 13.088 5 9.072 23

8 220 13.004 9 9.014 33 8 720 13.090 1 9.073 37
8 230 13.006 7 9.015 54 8 730 13.091 8 9.074 52
8 240 13.008 4 9.016 76 8 740 13.093 4 9.075 67
8 250 13.0102 9.017 97 8 750 13.095 1 9.076 81
8 260 13.011 9 9.019 18 8 760 13.096 7 9.077 95
8 270 13.013 7 9.020 39 8 770 13.098 4 9.079 09
8 280 13.015 4 9.021 60 8 780 13.100 0 9.080 23
8 290 13.017 2 9.022 81 8 790 13.101 6 9.081 37
8 300 13.018 9 9.024 01 8 800 13.103 3 9.082 51
8 310 13.020 6 9.025 21 8 810 13.1049 9.083 64
8 320 13.022 4 9.026 42 8 820 13.106 6 9.084 78
8 330 13.024 1 9.027 62 8 830 13.108 2 9.085 91
8 340 13.025 8 9.028 82 8 840 13.109 8 9.087 04
8 350 13.027 6 9.030 02 8 850 13.111 5 9.088 17
8 360 13.029 3 9.031 21 8 860 13.113 1 9.089 30
8 370 13.031 0 9.032 41 8 870 13.1147 9.090 43
8 380 13.032 7 9.033 60 8 880 13.116 3 9.091 56
8 390 13.034 5 9.034 80 8 890 13.118 0 9.092 68
8 400 13.036 2 9.035 99 8 900 13.119 6 9.093 81
8 410 13.037 9 9.037 18 8 910 13.121 2 9.094 93
8 420 13.039 6 9.038 37 8 920 13.1228 9.096 05
8 430 13.041 3 9.039 55 8 930 13.1244 9.097 17
8 440 13.043 0 9.040 74 8 940 13.126 1 9.098 29
8 450 13.044 7 9.041 92 8 950 13.127 7 9.099 41
8 460 13.046 4 9.043 10 8 960 13.129 3 9.100 53
8 470 13.048 1 9.044 29 8 970 13.1309 9.101 64
8 480 13.049 8 9.045 47 8 980 13.132 5 9.102 76
8 490 13.051 5 9.046 64 8 990 13.134 1 9.103 87
8 500 13.053 2 9.047 82 9 000 13.135 7 9.104 98
TABLES 447
TABLE A (concluded)
n log2 n logc n n log2 n log,, n
9 010 13.137 3 9.106 09 9 510 13.215 2 9.160 10

9 020 13.138 9 9.107 20 9 520 13.2167 9.161 15
9 030 13.140 5 9.108 31 9 530 13.218 3 9.162 20
9 040 13.142 1 9.109 41 9 540 13.219 8 9.163 25
9 050 13.143 7 9.110 52 9 550 13.221 3 9.164 30
9 060 13.145 3 9.111 62 9 560 13.222 8 9.165 34
9 070 13.146 9 9.112 73 9 570 13.224 3 9.166 39
9 080 13.148 5 9.113 83 9 580 13.225 8 9.167 43
9 090 13.150 1 9.11493 9 590 13.227 3 9.168 48
9 100 13.151 7 9.11603 9 600 13.228 8 9.169 52
9 110 13.153 2 9.117 13 9 610 13.230 3 9.170 56

9 120 13.1548 9.118 23 9 620 13.231 8 9.171 60
9 130 13.1564 9.119 32 9 630 13.233 3 9.172 64
9 140 13.1580 9.120 42 9 640 13.234 8 9.173 68
9 150 13.159 6 9.121 51 9 650 13.236 3 9.174 71
9 160 13.161 1 9.122 60 9 660 13.237 8 9.175 75
9 170 13.162 7 9.123 69 9 670 13.239 3 9.176 78
9 180 13.164 3 9.124 78 9 680 13.240 8 9.177 82
9 190 13.165 8 9.125 87 9 690 13.242 3 9.178 85
9 200 13.167 4 9.126 96 9 700 13.243 8 9.179 88
9 210 13.169 0 9.128 05 9 710 13.245 3 9.180 91

9 220 13.170 6 9.129 13 9 720 13.246 7 9.181 94
9 230 13.172 1 9.130 21 9 730 13.248 2 9.182 97
9 240 13.173 7 9.131 30 9 740 13.249 7 9.184 00
9 250 13.175 2 9.132 38 9 750 13.251 2 9.185 02
9 260 13.176 8 9.133 46 9 760 13.252 7 9.186 05
9 270 13.178 4 9.134 54 9 770 13.254 1 9.187 07
9 280 13.179 9 9.135 62 9 780 13.255 6 9.188 09
9 290 13.181 5 9.136 69 9 790 13.257 1 9.189 12
9 300 13.183 0 9.137 77 9 800 13.258 6 9.190 14
9 310 13.184 6 9.138 84 9 810 13.260 0 9.191 16

9 320 13.186 1 9.139 92 9 820 13.261 5 9.192 18
9 330 13.187 7 9.140 99 9 830 13.263 0 9.193 19
9 340 13.1892 9.142 06 9 840 13.264 4 9.194 21
9 350 13.190 8 9.143 13 9 850 13.265 9 9.195 23
9 360 13.192 3 9.144 20 9 860 13.267 4 9.196 24
9 370 13.193 8 9.145 27 9 870 13.268 8 9.197 26
9 380 13.195 4 9.146 34 9 880 13.270 3 9.198 27
9 390 13.196 9 9.147 40 9 890 13.271 8 9.199 28
9 400 13.198 4 9.148 46 9 900 13.273 2 9.200 29
9 410 13.200 0 9.149 53 9 910 13.274 7 9.201 30

9 420 13.201 5 9.150 59 9 920 13.276 1 9.202 31
9 430 13.203 0 9.151 65 9 930 13.277 6 9.203 32
9 440 13.204 6 9.152 71 9 940 13.279 0 9.204 32
9 450 13.206 1 9.153 77 9 950 13.280 5 9.205 33
9 460 13.207 6 9.154 83 9 960 13.281 9 9.206 33
9 470 13.209 1 9.155 88 9 970 13.283 4 9.207 34
9 480 13.210 7 9.156 94 9 980 13.284 8 9.208 34
9 490 13.2122 9.157 99 9 990 13.286 3 9.209 34
9 500 13.213 7 9.159 05 10 000 13.287 7 9.210 34
448 TABLES
TABLE B
LOGARITHMS TO THE BASE 2 AND NATURAL LOGARITHMS OF RECIPROCALS OF PROBABILITIES
1 1 1 1
p log2 logc P l0g2 loge -
P P P P
.001 9.965 78 6.907 76 .051 4.293 36 2.975 93

.002 8.965 78 6.214 61 .052 4.265 34 2.956 51
.003 8.380 82 5.809 14 .053 4.237 86 2.937 46
.004 7.965 78 5.521 46 .054 4.210 90 2.918 77
.005 7.643 86 5.298 32 .055 4.184 42 2.900 42
.006 7.380 82 5.11600 .056 4.158 43 2.882 40
.007 7.158 43 4.961 85 .057 4.132 89 2.864 70
.008 6.965 78 4.828 31 .058 4.107 80 2.847 31
.009 6.795 86 4.710 53 .059 4.083 14 2.830 22
.010 6.643 86 4.605 17 .060 4.058 89 2.813 41
.011 6.506 35 4.509 86 .061 4.035 05 2.796 88

.012 6.380 82 4.422 85 .062 4.011 59 2.780 62
.013 6.265 34 4.342 81 .063 3.988 50 2.764 62
.014 6.158 43 4.268 70 .064 3.965 78 2.748 87
.015 6.058 89 4.199 71 .065 3.943 42 2.733 37
.016 5.965 78 4.135 17 .066 3.921 39 2.718 10
.017 5.878 32 4.074 54 .067 3.899 70 2.703 06
.018 5.795 86 4.017 38 .068 3.878 32 2.688 25
.019 5.717 86 3.963 32 .069 3.857 26 2.673 65
.020 5.643 86 3.912 02 .070 3.836 50 2.659 26
.021 5.573 47 3.863 23 .071 3.816 04 2.645 08

.022 5.506 35 3.816 71 .072 3.795 86 2.631 09
.023 5.442 22 3.772 26 .073 3.775 96 2.617 30
.024 5.380 82 3.729 70 .074 3.756 33 2.603 69
.025 5.321 93 3.688 88 .075 3.736 97 2.590 27
.026 5.265 34 3.649 66 .076 3.717 86 2.577 02
.027 5.210 90 3.611 92 .077 3.699 00 2.563 95
.028 5.158 43 3.575 55 .078 3.680 38 2.551 05
.029 5.107 80 3.540 46 .079 3.662 00 2.538 31
.030 5.058 89 3.506 56 .080 3.643 86 2.525 73
.031 5.011 59 3.473 77 .081 3.625 93 2.513 31

.032 4.965 78 3.442 02 .082 3.608 23 2.501 04
.033 4.921 39 3.411 25 .083 3.590 74 2.488 91
.034 4.878 32 3.381 39 .084 3.573 47 2.476 94
.035 4.836 50 3.352 41 .085 3.556 39 2.465 10
.036 4.795 86 3.324 24 .086 3.539 52 2.453 41
.037 4.756 33 3.296 84 .087 3.522 84 2.441 85
.038 4.717 86 3.270 17 .088 3.506 35 2.430 42
.039 4.680 38 3.244 19 .089 3.490 05 2.419 12
.040 4.643 86 3.218 88 .090 3.473 93 2.407 95
.041 4.608 23 3.194 18 .091 3.457 99 2.396 90

.042 4.573 47 3.170 09 .092 3.442 22 2.385 97
.043 4.539 52 3.146 56 .093 3.426 63 2.375 16
.044 4.506 35 3.123 57 .094 3.411 20 2.364 46
.045 4.473 93 3.101 09 .095 3.395 93 2.353 88
.046 4.442 22 3.079 11 .096 3.380 82 2.343 41
.047 4.411 20 3.057 61 .097 3.365 87 2.333 04
.048 4.380 82 3.036 55 .098 3.351 07 2.322 79
.049 4.351 07 3.015 93 .099 3.336 43 2.312 64
.050 4.321 93 2.995 73 .100 3.321 93 2.302 59
TABLES 449
TABLE B (continued)
p log2 1 loge 1 P log2 — loge 1

P p P P
.101 3.307 57 2.292 63 .151 2.727 38 1.890 48

.102 3.293 36 2.282 78 .152 2.717 86 1.883 87
.103 3.279 28 2.273 03 .153 2.708 40 1.877 32
.104 3.265 34 2.263 36 .154 2.699 00 1.870 80
.105 3.251 54 2.253 79 .155 2.689 66 1.864 33
.106 3.237 86 2.244 32 .156 2.680 38 1.857 90
.107 3.224 32 2.234 93 .157 2.671 16 1.851 51
.108 3.210 90 2.225 62 .158 2.662 00 1.845 16
.109 3.197 60 2.216 41 .159 2.652 90 1.838 85
.110 3.184 42 2.207 27 .160 2.643 86 1.832 58
.111 3.171 37 2.198 23 .161 2.634 87 1.826 35

.112 3.158 43 2.189 26 .162 2.625 93 1.820 16
.113 3.145 61 2.180 37 .163 2.617 06 1.81401
.114 3.132 89 2.171 56 .164 2.608 23 1.807 89
.115 3.120 29 2.162 82 .165 2.599 46 1.801 81
.116 3.107 80 2.15417 .166 2.590 74 1.795 77
.117 3.095 42 2.145 58 .167 2.582 08 1.789 76
.118 3.083 14 2.137 07 .168 2.573 47 1.783 79
.119 3.070 97 2.128 63 .169 2.564 90 1.777 86
.120 3.058 89 2.120 26 .170 2.556 39 1.771 96
.121 3.046 92 2.111 96 .171 2.547 93 1.766 09

.122 3.035 05 2.103 73 .172 2.539 52 1.760 26
.123 3.023 27 2.095 57 .173 2.531 16 1.754 46
.124 3.011 59 2.087 47 .174 2.522 84 1.748 70
.125 3.000 00 2.079 44 .175 2.514 57 1.742 97
.126 2.988 50 2.071 47 .176 2.506 35 1.737 27
.127 2.977 10 2.063 57 .177 2.498 18 1.731 61
.128 2.965 78 2.055 73 .178 2.490 05 1.725 97
.129 2.954 56 2.047 94 .179 2.481 97 1.720 37
.130 2.943 42 2.040 22 .180 2.473 93 1.714 80
.131 2.932 36 2.032 56 .181 2.465 94 1.709 26

.132 2.921 39 2.024 95 .182 2.457 99 1.703 75
.133 2.910 50 2.017 41 .183 2.450 08 1.698 27
.134 2.899 70 2.009 92 .184 2.442 22 1.692 82
.135 2.888 97 2.002 48 .185 2.434 40 1.687 40
.136 2.878 32 1.995 10 .186 2.426 63 1.682 01
.137 2.867 75 1.987 77 .187 2.418 89 1.676 65
.138 2.857 26 1.980 50 .188 2.411 20 1.671 31
.139 2.846 84 1.973 28 .189 2.403 54 1.666 01
.140 2.836 50 1.966 11 .190 2.395 93 1.660 73
.141 2.826 23 1.959 00 .191 2.388 36 1.655 48

.142 2.816 04 1.951 93 .192 2.380 82 1.650 26
.143 2.805 91 1.944 91 .193 2.373 33 1.645 07
.144 2.795 86 1.937 94 .194 2.365 87 1.639 90
.145 2.785 88 1.931 02 .195 2.358 45 1.634 76
.146 2.775 96 1.924 15 .196 2.351 07 1.629 64
.147 2.766 11 1.917 32 .197 2.343 73 1.624 55
.148 2.756 33 1.910 54 .198 2.336 43 1.619 49
.149 2.746 62 1.903 81 .199 2.329 16 1.614 45
.150 2.736 97 1.897 12 .200 2.321 93 1.609 44
450 TABLES
TABLE B (continued)
P loga 1 loge 1 P log2 1 loge -

P P P P
.201 2.314 73 1.604 45 .251 1.994 24 1.382 30

.202 2.307 57 1.599 49 .252 1.988 50 1.378 33
.203 2.300 45 1.594 55 .253 1.982 79 1.374 37
.204 2.293 36 1.589 64 .254 1.977 10 1.370 42
.205 2.286 30 1.584 75 .255 1.971 43 1.366 49
.206 2.279 28 1.579 88 .256 1.965 78 1.362 58
.207 2.272 30 1.575 04 .257 1.960 16 1.358 68
.208 2.265 34 1.570 22 .258 1.954 56 1.354 80
.209 2.258 43 1.565 42 .259 1.948 98 1.350 93
.210 2.251 54 1.560 65 .260 1.943 42 1.347 07
.211 2.244 69 1.555 90 .261 1.937 88 1.343 23

.212 2.237 86 1.551 17 .262 1.932 36 1.339 41
.213 2.231 07 1.546 46 .263 1.926 87 1.335 60
.214 2.224 32 1.541 78 .264 1.921 39 1.331 81
.215 2.217 59 1.537 12 .265 1.915 94 1.328 03
.216 2.210 90 1.532 48 .266 1.910 50 1.324 26
.217 2.204 23 1.527 86 .267 1.905 09 1.320 51
.218 2.197 60 1.523 26 .268 1.899 70 1.31677
.219 2.191 00 1.518 68 .269 1.894 32 1.313 04
.220 2.184 42 1.514 13 .270 1.888 97 1.309 33
.221 2.177 88 1.509 59 .271 1.883 64 1.305 64

.222 2.171 37 1.505 08 .272 1.878 32 1.301 95
.223 2.164 88 1.500 58 .273 1.873 03 1.298 28
.224 2.158 43 1.496 11 .274 1.867 75 1.294 63
.225 2.152 00 1.491 65 .275 1.862 50 1.290 98
.226 2.145 61 1.487 22 .276 1.857 26 1.287 35
.227 2.139 24 1.482 81 .277 1.852 04 1.283 74
.228 2.132 89 1.478 41 .278 1.846 84 1.280 13
.229 2.126 58 1.474 03 .279 1.841 66 1.276 54
.230 2.120 29 1.469 68 .280 1.836 50 1.272 97
.231 2.114 04 1.465 34 .281 1.831 36 1.269 40

.232 2.107 80 1.461 02 .282 1.826 23 1.265 85
.233 2.101 60 1.456 72 .283 1.821 13 1.262 31
.234 2.095 42 1.452 43 .284 1.816 04 1.258 78
.235 2.089 27 1.448 17 .285 1.81097 1.255 27
.236 2.083 14 1.443 92 .286 1.805 91 1.251 76
.237 2.077 04 1.439 70 .287 1.800 88 1.248 27
.238 2.070 97 1.435 48 .288 1.795 86 1.244 79
.239 2.064 92 1.431 29 .289 1.790 86 1.241 33
.240 2.058 89 1.427 12 .290 1.785 88 1.237 87
.241 2.052 89 1.422 96 .291 1.780 91 1.234 43

.242 2.046 92 1.418 82 .292 1.775 96 1.231 00
.243 2.040 97 1.414 69 .293 1.771 03 1.227 58
.244 2.035 05 1.410 59 .294 1.766 11 1.224 18
.245 2.029 15 1.406 50 .295 1.761 21 1.220 78
.246 2.023 27 1.402 42 .296 1.756 33 1.217 40
.247 2.017 42 1.398 37 .297 1.751 47 1.214 02
.248 2.011 59 1.394 33 .298 1.746 62 1.210 66
.249 2.005 78 1.390 30 .299 1.741 78 1.207 31
.250 2.000 00 1.386 29 .300 1.736 97 1.203 97
TABLES 451
TABLE B ( continued)
p 10g2 lOge 1 P log2 ' logc 1

P P P P
.301 1.732 16 1.200 65 .351 1.51046 1.046 97

.302 1.727 38 1.197 33 .352 1.506 35 1.044 12
.303 1.722 61 1.194 02 .353 1.502 26 1.041 29
.304 1.717 86 1.190 73 .354 1.498 18 1.038 46
.305 1.713 12 1.187 44 .355 1.494 11 1.035 64
.306 1.708 40 1.184 17 .356 1.490 05 1.032 82
.307 1.703 69 1.180 91 .357 1.486 00 1.030 02
.308 1.699 00 1.177 66 .358 1.481 97 1.027 22
.309 1.694 32 1.174 41 .359 1.477 94 1.024 43
.310 1.689 66 1.171 18 .360 1.473 93 1.021 65
.311 1.685 01 1.167 96 .361 1.469 93 1.018 88

.312 1.680 38 1.164 75 .362 1.465 94 1.016 11
.313 1.675 77 1.161 55 .363 1.461 96 1.013 35
.314 1.671 16 1.158 36 .364 1.457 99 1.010 60
.315 1.666 58 1.155 18 .365 1.454 03 1.007 86
.316 1.662 00 1.152 01 .366 1.450 08 1.005 12
.317 1.657 45 1.148 85 .367 1.446 15 1.002 39
.318 1.652 90 1.145 70 .368 1.442 22 .999 672
.319 1.648 37 1.142 56 .369 1.438 31 .996 959
.320 1.643 86 1.139 43 .370 1.434 40 .994 252
.321 1.639 35 1.136 31 .371 1.430 51 .991 553

.322 1.634 87 1.133 20 .372 1.426 63 .988 861
.323 1.630 39 1.130 10 .373 1.422 75 .986 177
.324 1.625 93 1.127 01 .374 1.418 89 .983 499
.325 1.621 49 1.123 93 .375 1.415 04 .980 829
.326 1.617 06 1.120 86 .376 1.411 20 .978 166
.327 1.612 64 1.117 80 .377 1.407 36 .975 510
.328 1.608 23 1.114 74 .378 1.403 54 .972 861
.329 1.603 84 1.111 70 .379 1.399 73 .970 219
.330 1.599 46 1.108 66 .380 1.395 93 .967 584
.331 1.595 10 1.105 64 .381 1.392 14 .964 956

.332 1.590 74 1.102 62 .382 1.388 36 .962 335
.333 1.586 41 1.099 61 .383 1.384 58 .959 720
.334 1.582 08 1.096 61 .384 1.380 82 .957 113
.335 1.577 77 1.093 62 .385 1.377 07 .954 512
.336 1.573 47 1.090 64 .386 1.373 33 .951 918
.337 1.569 18 1.087 67 .387 1.369 59 .949 331
.338 1.564 90 1.084 71 .388 1.365 87 .946 750
.339 1.560 64 1.081 76 .389 1.362 16 .944 176
.340 1.556 39 1.078 81 .390 1.358 45 .941 609
.341 1.552 16 1.075 87 .391 1.354 76 .939 048

.342 1.547 93 1.072 94 .392 1.351 07 .936 493
.343 1.543 72 1.070 02 .393 1.347 40 .933 946
.344 1.539 52 1.067 11 .394 1.343 73 .931 404
.345 1.535 33 1.064 21 .395 1.340 08 .928 870
.346 1.531 16 1.061 32 .396 1.336 43 .926 341
.347 1.526 99 1.058 43 .397 1.332 79 .923 819
.348 1.522 84 1.055 55 .398 1.329 16 .921 303
.349 1.518 70 1.052 68 .399 1.325 54 .918 794
.350 1.514 57 1.049 82 .400 1.321 93 .916 291
452 TABLES
TABLE B (continued)
1 1 1 1
p 10g2 loge — P l0g2 — loge
P P P P
.401 1.318 33 .913 794 .451 1.148 80 .796 288

.402 1.314 73 .911 303 .452 1.145 61 .794 073
.403 1.311 15 .908 819 .453 1.142 42 .791 863
.404 1.307 57 .906 340 .454 1.139 24 .789 658
.405 1.304 01 .903 868 .455 1.136 06 .787 458
.406 1.300 45 .901 402 .456 1.132 89 .785 262
.407 1.296 90 .898 942 .457 1.129 73 .783 072
.408 1.293 36 .896 488 .458 1.126 58 .780 886
.409 1.289 83 .894 040 .459 1.123 43 .778 705
.410 1.286 30 .891 598 .460 1.120 29 .776 529
.411 1.282 79 .889 162 .461 1.117 16 .774 357

.412 1.279 28 .886 732 .462 1.11404 .772 190
.413 1.275 79 .884 308 .463 1.11092 .770 028
.414 1.272 30 .881 889 .464 1.107 80 .767 871
.415 1.268 82 .879 477 .465 1.104 70 .765 718
.416 1.265 34 .877 070 .466 1.101 60 .763 570
.417 1.261 88 .874 669 .467 1.098 51 .761 426
.418 1.258 43 .872 274 .468 1.095 42 .759 287
.419 1.254 98 .869 884 .469 1.092 34 .757 153
.420 1.251 54 .867 501 .470 1.089 27 .755 023
.421 1.248 11 .865 122 .471 1.086 20 .752 897

.422 1.244 69 .862 750 .472 1.083 14 .750 776
.423 1.241 27 .860 383 .473 1.080 09 .748 660
.424 1.237 86 .858 022 .474 1.077 04 .746 548
.425 1.234 47 .855 666 .475 1.074 00 .744 440
.426 1.231 07 .853 316 .476 1.070 97 .742 337
.427 1.227 69 .850 971 .477 1.067 94 .740 239
.428 1.224 32 .848 632 .478 1.064 92 .738 145
.429 1.220 95 .846 298 .479 1.061 90 .736 055
.430 1.217 59 .843 970 .480 1.058 89 .733 969
.431 1.214 24 .841 647 .481 1.055 89 .731 888

.432 1.210 90 .839 330 .482 1.052 89 .729 811
.433 1.207 56 .837 018 .483 1.049 90 .727 739
.434 1.204 23 .834 711 .484 1.046 92 .725 670
.435 1.200 91 .832 409 .485 1.043 94 .723 606
.436 1.197 60 .830 113 .486 1.040 97 .721 547
.437 1.194 29 .827 822 .487 1.038 01 .719 491
.438 1.191 00 .825 536 .488 1.035 05 .717 440
.439 1.187 71 .823 256 .489 1.032 09 .715 393
.440 1.184 42 .820 981 .490 1.029 15 .713 350
.441 1.181 15 .818 710 .491 1.026 21 .711 311

.442 1.177 88 .816 445 .492 1.023 27 .709 277
.443 1.174 62 .814 186 .493 1.020 34 .707 246
.444 1.171 37 .811 931 .494 1.017 42 .705 220
.445 1.168 12 .809 681 .495 1.014 50 .703 198
.446 1.164 88 .807 436 .496 1.011 59 .701 179
.447 1.161 65 .805 197 .497 1.008 68 .699 165
.448 1.158 43 .802 962 .498 1.005 78 .697 155
.449 1.155 21 .800 732 .499 1.002 89 .695 149
.450 1.152 00 .798 508 .500 1.000 00 .693 147
TABLES 453
TABLE B (continued)
p log2 1 log* 1 P log2 1 loge 1

P P P P
.501 .997 117 .691 149 .551 .859 876 .596 020
.502 .994 241 .689 155 .552 .857 260 .594 207
.503 .991 370 .687 165 .553 .854 649 .592 397
.504 .988 504 .685 179 .554 .852 042 .590 591
.505 .985 645 .683 197 .555 .849 440 .588 787
.506 .982 791 .681 219 .556 .846 843 .586 987
.507 .979 942 .679 244 .557 .844 251 .585 190
.508 .977 100 .677 274 .558 .841 663 .583 396
.509 .974 262 .675 307 .559 .839 080 .581 606
.510 .971 431 .673 345 .560 .836 501 .579 818
.511 .968 605 .671 386 .561 .833 927 .578 034
.512 .965 784 .669 431 .562 .831 358 .576 253
.513 .962 969 .667 479 .563 .828 793 .574 476
.514 .960 160 .665 532 .564 .826 233 .572 701
.515 .957 356 .663 588 .565 .823 677 .570 930
.516 .954 557 .661 649 .566 .821 126 .569 161
.517 .951 764 .659 712 .567 .818 579 .567 396
.518 .948 976 .657 780 .568 .816 037 .565 634
.519 .946 194 .655 851 .569 .813 499 .563 875
.520 .943 416 .653 926 .570 .810 966 .562 119
.521 .940 645 .652 005 .571 .808 437 .560 366
.522 .937 878 .650 088 .572 .805 913 .558 616
.523 .935 117 .648 174 .573 .803 393 .556 870
.524 .932 361 .646 264 .574 .800 877 .555 126
.525 .929 611 .644 357 .575 .798 366 .553 385
.526 .926 865 .642 454 .576 .795 859 .551 648
.527 .924 125 .640 555 .577 .793 357 .549 913
.528 .921 390 .638 659 .578 .790 859 .548 181
.529 .918 660 .636 767 .579 .788 365 .546 453
.530 .915 936 .634 878 .580 .785 875 .544 727
.531 .913 216 .632 993 .581 .783 390 .543 005
.532 .910 502 .631 112 .582 .780 909 .541 285
.533 .907 793 .629 234 .583 .778 432 .539 568
.534 .905 088 .627 359 .584 .775 960 .537 854
.535 .902 389 .625 489 .585 .773 491 .536 143
.536 .899 695 .623 621 .586 .771 027 .534 435
.537 .897 006 .621 757 .587 .768 568 .532 730
.538 .894 322 .619 897 .588 .766 112 .531 028
.539 .891 643 .618 040 .589 .763 660 .529 329
.540 .888 969 .616 186 .590 .761 213 .527 633
.541 .886 299 .614 336 .591 .758 770 .525 939
.542 .883 635 .612 489 .592 .756 331 .524 249
.543 .880 976 .610 646 .593 .753 896 .522 561
.544 .878 321 .608 806 .594 .751 465 .520 876
.545 .875 672 .606 969 .595 .749 038 .519 194
.546 .873 027 .605 136 .596 .746 616 .517 515
.547 .870 387 .603 306 .597 .744 197 .515 838
.548 .867 752 .601 480 .598 .741 783 .514 165
.549 .865 122 .599 657 .599 .739 372 .512 494
.550 .862 496 .597 837 .600 .736 966 .510 826
454 TABLES
TABLE B (continued)
p log2 1 logc 1 P 10g2 ~ logs 1

P P P P
.601 .734 563 .509 160 .651 .619 271 .429 246
.602 .732 165 .507 498 .652 .617 056 .427 711
.603 .729 770 .505 838 .653 .614 845 .426 178
.604 .727 380 .504 181 .654 .612 637 .424 648
.605 .724 993 .502 527 .655 .610 433 .423 120
.606 .722 610 .500 875 .656 .608 232 .421 594
.607 .720 232 .499 226 .657 .606 035 .420 071
.608 .717 857 .497 580 .658 .603 841 .418 550
.609 .715 486 .495 937 .659 .601 650 .417 032
.610 .713 119 .494 296 .660 .599 462 .415 515
.611 .710 756 .492 658 .661 .597 278 .414 001
.612 .708 396 .491 023 .662 .595 097 .412 490
.613 .706 041 .489 390 .663 .592 919 .410 980
.614 .703 689 .487 760 .664 .590 745 .409 473
.615 .701 342 .486 133 .665 .588 574 .407 968
.616 .698 998 .484 508 .666 .586 406 .406 466
.617 .696 658 .482 886 .667 .584 241 .404 965
.618 .694 321 .481 267 .668 .582 080 .403 467
.619 .691 989 .479 650 .669 .579 922 .401 971
.620 .689 660 .478 036 .670 .577 767 .400 478
.621 .687 335 .476 424 .671 .575 615 .398 986
.622 .685 014 .474 815 .672 .573 467 .397 497
.623 .682 696 .473 209 .673 .571 322 .396 010
.624 .680 382 .471 605 .674 .569 179 .394 525
.625 .678 072 .470 004 .675 .567 041 .393 043
.626 .675 765 .468 405 .676 .564 905 .391 562
.627 .673 463 .466 809 .677 .562 772 .390 084
.628 .671 164 .465 215 .678 .560 643 .388 608
.629 .668 868 .463 624 .679 .558 517 .387 134
.630 .666 576 .462 035 .680 .556 393 .385 662
.631 .664 288 .460 449 .681 .554 273 .384 193
.632 .662 004 .458 866 .682 .552 156 .382 726
.633 .659 723 .457 285 .683 .550 043 .381 260
.634 .657 445 .455 706 .684 .547 932 .379 797
.635 .655 171 .454 130 .685 .545 824 .378 336
.636 .652 901 .452 557 .686 .543 720 .376 878
.637 .650 635 .450 986 .687 .541 618 .375 421
.638 .648 372 .449 417 .688 .539 520 .373 966
.639 .646 112 .447 851 .689 .537 424 .372 514
.640 .643 856 .446 287 .690 .535 332 .371 064
.641 .641 604 .444 726 .691 .533 242 .369 615
.642 .639 355 .443 167 .692 .531 156 .368 169
.643 .637 109 .441 611 .693 .529 073 .366 725
.644 .634 867 .440 057 .694 .526 992 .365 283
.645 .632 629 .438 505 .695 .524 915 .363 843
.646 .630 394 .436 956 .696 .522 841 .362 406
.647 .628 162 .435 409 .697 .520 769 .360 970
.648 .625 934 .433 865 .698 .518 701 .359 536
.649 .623 710 .432 323 .699 .516 636 .358 105
.650 .621 488 .430 783 .700 .514 573 .356 675
TABLES 455
TABLE B (continued)
p log2 — loge 1 P log2 1 logP 1

P P P P
.701 .512 514 .355 247 .751 .413 115 .286 350
.702 .510 457 .353 822 .752 .411 195 .285 019
.703 .508 403 .352 398 .753 .409 278 .283 690
.704 .506 353 .350 977 .754 .407 364 .282 363
.705 .504 305 .349 557 .755 .405 451 .281 038
.706 .502 260 .348 140 .756 .403 542 .279 714
.707 .500 218 .346 725 .757 .401 635 .278 392
.708 .498 179 .345 311 .758 .399 730 .277 072
.709 .496 142 .343 900 .759 .397 828 .275 753
.710 .494 109 .342 490 .760 .395 929 .274 437
.711 .492 079 .341 083 .761 .394 032 .273 122
.712 .490 051 .339 677 .762 .392 137 .271 809
.713 .488 026 .338 274 .763 .390 245 .270 497
.714 .486 004 .336 872 .764 .388 355 .269 187
.715 .483 985 .335 473 .765 .386 468 .267 879
.716 .481 969 .334 075 .766 .384 584 .266 573
.717 .479 955 .332 679 .767 .382 702 .265 268
.718 .477 944 .331 286 .768 .380 822 .263 966
.719 .475 936 .329 894 .769 .378 944 .262 664
.720 .473 931 .328 504 .770 .377 070 .261 365
.721 .471 929 .327 116 .771 .375 197 .260 067
.722 .469 929 .325 730 .772 .373 327 .258 771
.723 .467 932 .324 346 .773 .371 460 .257 476
.724 .465 938 .322 964 .774 .369 595 .256 183
.725 .463 947 .321 584 .775 .367 732 .254 892
.726 .461 959 .320 205 .776 .365 871 .253 603
.727 .459 973 .318 829 .777 .364 013 .252 315
.728 .457 990 .317 454 .778 .362 158 .251 029
.729 .456 009 .316 082 .779 .360 305 .249 744
.730 .454 032 .314 711 .780 .358 454 .248 461
.731 .452 057 .313 342 .781 .356 606 .247 180
.732 .450 084 .311 975 .782 .354 759 .245 901
.733 .448 115 .310610 .783 .352 916 .244 623
.734 .446 148 .309 246 .784 .351 074 .243 346
.735 .444 184 .307 885 .785 .349 235 .242 072
.736 .442 222 .306 525 .786 .347 399 .240 798
.737 .440 263 .305 167 .787 .345 564 .239 527
.738 .438 307 .303 811 .788 .343 732 .238 257
.739 .436 354 .302 457 .789 .341 903 .236 989
.740 .434 403 .301 105 .790 .340 075 .235 722
.741 .432 455 .299 755 .791 .338 250 .234 457
.742 .430 509 .298 406 .792 .336 428 .233 194
.743 .428 566 .297 059 .793 .334 607 .231 932
.744 .426 625 .295 714 .794 .332 789 .230 672
.745 .424 688 .294 371 .795 .330 973 .229 413
.746 .422 752 .293 030 .796 .329 160 .228 156
.747 .420 820 .291 690 .797 .327 348 .226 901
.748 .418 890 .290 352 .798 .325 539 .225 647
.749 .416 962 .289 016 .799 .323 733 .224 394
.750 .415 037 .287 682 .800 .321 928 .223 144
456 TABLES
TABLE B (continued)
1 1 1 1
p 10g2 log? P log2 - logs
P P P P
.801 .320 126 .221 894 .851 .232 769 .161 343
.802 .318 326 .220 647 .852 .231 075 .160 169
.803 .316 528 .219 401 .853 .229 382 .158 996
.804 .314 733 .218 156 .854 .227 692 .157 824
.805 .312 939 .216 913 .855 .226 004 .156 654
.806 .311 148 .215 672 .856 .224 317 .155 485
.807 .309 359 .214 432 .857 .222 633 .154 317
.808 .307 573 .213 193 .858 .220 950 .153 151
.809 .305 788 .211 956 .859 .219 270 .151 986
.810 .304 006 .210 721 .860 .217 591 .150 823
.811 .302 226 .209 487 .861 .215 915 .149 661
.812 .300 448 .208 255 .862 .214 240 .148 500
.813 .298 673 .207 024 .863 .212 568 .147 341
.814 .296 899 .205 795 .864 .210 897 .146 183
.815 .295 128 .204 567 .865 .209 228 .145 026
.816 .293 359 .203 341 .866 .207 561 .143 870
.817 .291 592 .202 116 .867 .205 896 .142 716
.818 .289 827 .200 893 .868 .204 233 .141 564
.819 .288 065 .199 671 .869 .202 572 .140412
.820 .286 304 .198 451 .870 .200 913 .139 262
.821 .284 546 .197 232 .871 .199 255 .138 113
.822 .282 790 .196015 .872 .197 600 .136 966
.823 .281 036 .194 799 .873 .195 946 .135 820
.824 .279 284 .193 585 .874 .194 295 .134 675
.825 .277 534 .192 372 .875 .192 645 .133 531
.826 .275 786 .191 161 .876 .190 997 .132 389
.827 .274 041 .189 951 .877 .189 351 .131 248
.828 .272 297 .188 742 .878 .187 707 .130 109
.829 .270 556 .187 535 .879 .186 065 .128 970
.830 .268 817 .186 330 .880 .184 425 .127 833
.831 .267 080 .185 125 .881 .182 786 .126 698
.832 .265 345 .183 923 .882 .181 149 .125 563
.833 .263 612 .182 722 .883 .179 515 .124 430
.834 .261 881 .181 522 .884 .177 882 .123 298
.835 .260 152 .180 324 .885 .176 251 .122 168
.836 .258 425 .179 127 .886 .174 621 .121 038
.837 .256 700 .177 931 .887 .172 994 .119 910
.838 .254 978 .176 737 .888 .171 368 .1 18 784
.839 .253 257 .175 545 .889 . 169 745 .117 658
.840 .251 539 .174 353 .890 .168 123 .116 534
.841 .249 822 .173 164 .891 .166 503 .115411

.842 .248 108 .171 975 .892 .164 884 .114 289
.843 .246 395 .170 788 .893 .163 268 .113 169
.844 .244 685 .169 603 .894 .161 653 .112 050
.845 .242 977 .168 419 .895 .160 040 .110 932
.846 .241 270 .167 236 .896 .158 429 .109 815
.847 .239 566 .166 055 .897 .156 820 .108 699
.848 .237 864 .164 875 .898 .155 213 .107 585
.849 .236 164 .163 696 .899 .153 607 .106 472
.850 .234 465 .162 519 .900 .152 003 .105 361
TABLES 457
TABLE B (concluded)
p log2 1 loge 1 P log2 1 logs 1

P P P P
.901 .150 401 .104 250 .951 .072 483 .050 241
.902 .148 801 .103 141 .952 .070 967 .049 190
.903 .147 202 .102 033 .953 .069 452 .048 140
.904 .145 605 .100 926 .954 .067 939 .047 092
.905 .144 010 .099 820 .955 .066 427 .046 044
.906 .142 417 .098 716 .956 .064 917 .044 997
.907 .140 826 .097 613 .957 .063 409 .043 952
.908 .139 236 .096 511 .958 .061 902 .042 907
.909 .137 648 .095 410 .959 .060 397 .041 864
.910 .136 062 .094 311 .960 .058 894 .040 822
.911 .134 477 .093 212 .961 .057 392 .039 781
.912 .132 894 .092 115 .962 .055 891 .038 741
.913 .131 313 .091 019 .963 .054 392 .037 702
.914 .129 734 .089 925 .964 .052 895 .036 664
.915 .128 156 .088 831 .965 .051 399 .035 627
.916 .126 580 .087 739 .966 .049 905 .034 591
.917 .125 006 .086 648 .967 .048 412 .033 557
.918 .123 434 .085 558 .968 .046 921 .032 523
.919 .121 863 .084 469 .969 .045 431 .031 491
.920 .120 294 .083 382 .970 .043 943 .030 459
.921 .118 727 .082 295 .971 .042 457 .029 429
.922 .117 161 .081 210 .972 .040 972 .028 399
.923 .115 597 .080 126 .973 .039 488 .027 371
.924 .114035 .079 043 .974 .038 006 .026 344
.925 .112 475 .077 962 .975 .036 526 .025 318
.926 .110916 .076 881 .976 .035 047 .024 293
.927 .109 359 .075 802 .977 .033 570 .023 269
.928 .107 803 .074 724 .978 .032 094 .022 246
.929 .106 249 .073 647 .979 .030 619 .021 224
.930 .104 697 .072 571 .980 .029 146 .020 203
.931 .103 147 .071 496 .981 .027 675 .019 183
.932 .101 598 .070 422 .982 .026 205 .018 164
.933 .100 051 .069 350 .983 .024 737 .017 146
.934 .098 506 .068 279 .984 .023 270 .016 129
.935 .096 962 .067 209 .985 .021 804 .015 114
.936 .095 420 .066 140 .986 .020 340 .014 099
.937 .093 879 .065 072 .987 .018 878 .013 085
.938 .092 340 .064 005 .988 .017417 .012 073
.939 .090 803 .062 940 .989 .015 958 .011 061
.940 .089 267 .061 875 .990 .014 500 .010 050
.941 .087 733 .060 812 .991 .013 043 .009 041
.942 .086 201 .059 750 .992 .011 588 .008 032
.943 .084 670 .058 689 .993 .010 134 .007 025
.944 .083 141 .057 629 .994 .008 682 .006 018
.945 .081 614 .056 570 .995 .007 232 .005 013
.946 .080 088 .055 513 .996 .005 782 .004 008
.947 .078 564 .054 456 .997 .004 335 .003 005
.948 .077 041 .053 401 .998 .002 888 .002 002
.949 .075 520 .052 346 .999 .001 443 .001 001
.950 .074 001 .051 293 1.000 0 0
458 TABLES
TABLE C
LOGITS TO THE BASE 2 AND NATURAL LOGITS
P
p log-2 P— Ion P P IOg2 1 loge . P
1 -p l -p l —p 1 —p
.500 0 0 .550 .289 507 .200 671

.501 .005 771 .004 000 .551 .295 337 .204 712
.502 .011 542 .008 000 .552 .301 170 .208 755
.503 .017 313 .012 000 .553 .307 005 .212 799
.504 .023 084 .016 000 .554 .312 842 .216 846
.505 .028 855 .020 001 .555 .318 682 .220 894
.506 .034 626 .024 001 .556 .324 525 .224 944
.507 .040 398 .028 002 .557 .330 371 .228 995
.508 .046 170 .032 003 .558 .336 219 .233 049
.509 .051 943 .036 004 .559 .342 070 .237 105
.510 .057 715 .040 005 .560 .347 923 .241 162
.511 .063 489 .044 007 .561 .353 780 .245 221
.512 .069 263 .048 009 .562 .359 639 .249 283
.513 .075 037 .052 012 .563 .365 502 .253 346
.514 .080 812 .056 015 .564 .371 367 .257 412
.515 .086 588 .060 018 .565 .377 235 .261 480
.516 .092 364 .064 022 .566 .383 107 .265 550
.517 .098 141 .068 026 .567 .388 982 .269 622
.518 .103 919 .072 031 .568 .394 860 .273 696
.519 .109 698 .076 037 .569 .400 741 .277 772
.520 .115 477 .080 043 .570 .406 625 .281 851
.521 .121 258 .084 049 .571 .412 513 .285 932
.522 .127 039 .088 057 .572 .418 404 .290 016
.523 .132 822 .092 065 .573 .424 299 .294 102
.524 .138 605 .096 074 .574 .430 197 .298 190
.525 .144 390 .100 083 .575 .436 099 .302 281
.526 .150 176 .104 094 .576 .442 005 .306 374
.527 .155 963 .108 105 .577 .447 914 .310 470
.528 .161 751 .112 117 .578 .453 826 .314 569
.529 .167 541 .116 130 .579 .459 743 .318 670
.530 .173 332 .120 144 .580 .465 664 .322 773
.531 .179 124 .124 159 .581 .471 588 .326 880
.532 .184918 .128 175 .582 .477 516 .330 989
.533 .190 713 .132 192 .583 .483 448 .335 101
.534 .196 510 .136 210 .584 .489 385 .339 216
.535 .202 308 .140 229 .585 .495 325 .343 333
.536 .208 108 .144 250 .586 .501 270 .347 454
.537 .213 910 .148 271 .587 .507 219 .351 577
.538 .219 713 . 152 294 .588 .513 172 .355 704
.539 .225 519 .156 318 .589 .519 129 .359 833
.540 .231 326 .160 343 .590 .525 091 .363 965
.541 .237 134 .164 369 .591 .531 057 .368 101
.542 .242 945 .168 397 .592 .537 028 .372 239
.543 .248 758 .172 426 .593 .543 003 .376 381
.544 .254 573 .176 456 .594 .548 983 .380 526
.545 .260 390 .180 488 .595 .554 968 .384 674
.546 .266 209 .184 522 .596 .560 957 .388 826
.547 .272 030 .188 557 .597 .566 951 .392 981
.548 .277 853 .192 593 .598 .572 950 .397 139
.549 .283 679 .196 631 .599 .578 954 .401 300
TABLES 459
TABLE C (continued)
p log2 ~— lOge — P log2 P loge P

1 -p 1 -p 1 -p 1 -p
.600 .584 963 .405 465 .650 .893 085 .619 039
.601 .590 976 .409 634 .651 .899 431 .623 438
.602 .596 995 .413 805 .652 .905 785 .627 842
.603 .603 019 .417 981 .653 .912 147 .632 252
.604 .609 048 .422 160 .654 .918 519 .636 669
.605 .615 082 .426 343 .655 .924 899 .641 091
.606 .621 122 .430 529 .656 .931 287 .645 519
.607 .627 167 .434 719 .657 .937 685 .649 954
.608 .633 218 .438 913 .658 .944 091 .654 394
.609 .639 274 .443 111 .659 .950 507 .658 841
.610 .645 335 .447 312 .660 .956 931 .663 294
.611 .651 402 .451 518 .661 .963 365 .667 754
.612 .657 475 .455 727 .662 .969 808 .672 220
.613 .663 554 .459 940 .663 .976 260 .676 692
.614 .669 638 .464 158 .664 .982 722 .681 171
.615 .675 728 .468 379 .665 .989 193 .685 657
.616 .681 824 .472 604 .666 .995 674 .690 149
.617 .687 926 .476 834 .667 1.002 16 .694 648
.618 .694 034 .481 068 .668 1.008 66 .699 153
.619 .700 148 .485 306 .669 1.015 17 .703 666
.620 .706 269 .489 548 .670 1.021 70 .708 185

.621 .712 395 .493 795 .671 1.028 23 .712 711
.622 .718 528 .498 046 .672 1.034 77 .717 245
.623 .724 668 .502 301 .673 1.041 32 .721 785
.624 .730 813 .506 561 .674 1.047 88 .726 333
.625 .736 966 .510 826 .675 1.054 45 .730 888
.626 .743 124 .515 095 .676 1.061 03 .735 450
.627 .749 290 .519 368 .677 1.067 62 .740 019
.628 .755 462 .523 646 .678 1.074 22 .744 596
.629 .761 641 .527 929 .679 1.080 84 .749 180
.630 .767 827 .532 217 .680 1.087 46 .753 772

.631 .774 019 .536 509 .681 1.094 10 .758 371
.632 .780 219 .540 806 .682 1.100 74 .762 978
.633 .786 425 .545 109 .683 1.107 40 .767 593
.634 .792 639 .549 416 .684 1.11407 .772 216
.635 .798 860 .553 728 .685 1.12075 .776 846
.636 .805 088 .558 045 .686 1.127 44 .781 485
.637 .811 324 .562 367 .687 1.134 15 .786 131
.638 .817 567 .566 694 .688 1.140 86 .790 786
.639 .823 817 .571 026 .689 1.147 59 .795 448
.640 .830 075 .575 364 .690 1.154 33 .800 119

.641 .836 341 .579 707 .691 1.161 08 .804 799
.642 .842 614 .584 055 .692 1.167 84 .809 486
.643 .848 895 .588 409 .693 1.174 62 .814 182
.644 .855 183 .592 768 .694 1.181 40 .818 887
.645 .861 480 .597 133 .695 1.188 20 .823 600
.646 .867 785 .601 503 .696 1.195 02 .828 322
.647 .874 098 .605 878 .697 1.201 84 .833 053
.648 .880 418 .610 260 .698 1.208 68 .837 792
.649 .886 747 .614 646 .699 1.215 53 .842 540
460 TABLES
TABLE C (continued)
1 P loir ^
p IOg2 1 lOge P log2 ~~~
l 1 -p 1 -/> l -p
.700 1.222 39 .847 298 .750 1.584 96 1.098 61

.701 1.229 27 .852 064 .751 1.592 67 1.103 95
.702 1.236 16 .856 840 .752 1.600 39 1.109 31
.703 1.243 06 .861 625 .753 1.608 14 1.11468
.704 1.249 98 .866 419 .754 1.615 91 1.120 06
.705 1.256 91 .871 222 .755 1.623 69 1.125 46
.706 1.263 85 .876 035 .756 1.631 51 1.130 87
.707 1.270 81 .880 858 .757 1.639 34 1.136 30
.708 1.277 78 .885 690 .758 1.647 19 1.141 75
.709 1.284 77 .890 532 .759 1.655 07 1.147 20
.710 1.291 77 .895 384 .760 1.662 96 1.152 68

.711 1.298 78 .900 246 .761 1.670 89 1.158 17
.712 1.305 81 .905 117 .762 1.678 83 1.163 68
.713 1.312 85 .909 999 .763 1.686 80 1.169 20
.714 1.31991 .914 891 .764 1.694 79 1.174 74
.715 1.326 98 .919 793 .765 1.702 80 1.180 29
.716 1.334 07 .924 706 .766 1.710 84 1.185 86
.717 1.341 17 .929 629 .767 1.718 90 1.191 45
.718 1.348 29 .934 562 .768 1.726 98 1.197 05
.719 1.355 42 .939 507 .769 1.735 09 1.202 67
.720 1.362 57 .944 462 .770 1.743 22 1.208 31

.721 1.369 73 .949 427 .771 1.751 38 1.213 97
.722 1.376 91 .954 404 .772 1.759 57 1.219 64
.723 1.384 11 .959 392 .773 1.767 78 1.225 33
.724 1.391 32 .964 391 .774 1.776 01 1.231 04
.725 1.398 55 .969 401 .775 1.784 27 1.236 76
.726 1.405 79 .974 422 .776 1.792 56 1.242 51
.727 1.413 05 .979 455 .777 1.800 87 1.248 27
.728 1.420 33 .984 499 .778 1.809 21 1.254 05
.729 1.427 63 .989 555 .779 1.817 58 1.259 85
.730 1.434 94 .994 623 .780 1.825 97 1.265 67

.731 1.442 27 .999 702 .781 1.834 39 1.271 50
.732 1.449 61 1.004 79 .782 1.842 84 1.277 36
.733 1.456 97 1.009 90 .783 1.851 32 1.283 24
.734 1.464 35 1.015 01 .784 1.859 82 1.289 13
.735 1.471 75 1.020 14 .785 1.868 36 1.295 05
.736 1.479 17 1.025 28 .786 1.876 92 1.300 98
.737 1.486 60 1.030 43 .787 1.885 51 1.306 94
.738 1.494 05 1.035 60 .788 1.894 13 1.31291
.739 1.501 52 1.040 78 .789 1.902 78 1.318 91
.740 1.509 01 1.045 97 .790 1.911 46 1.324 93

.741 1.516 52 1.051 17 .791 1.920 17 1.330 96
.742 1.524 05 1.056 39 .792 1.928 92 1.337 02
.743 1.531 59 1.061 62 .793 1.937 69 1.343 10
.744 1.539 16 1.066 86 .794 1.946 49 1.349 21
.745 1.546 74 1.072 12 .795 1.955 33 1.355 33
.746 1.554 35 1.077 39 .796 1.964 20 1.361 48
.747 1.561 97 1.082 68 .797 1.973 10 1.367 65
.748 1.569 61 1.087 97 .798 1.982 03 1.373 84
.749 1.577 28 1.093 29 .799 1.991 00 1.380 06
TABLES 461
TABLE C (continued)
p logs P P 10g2 — logs

82 1 -p 1 —p 1 -p 1 -p
.800 2.000 00 1.386 29 .850 2.502 50 1.734 60

.801 2.009 03 1.392 56 .851 2.513 85 1.742 47
.802 2.018 10 1.398 84 .852 2.525 26 1.750 37
.803 2.027 20 1.405 15 .853 2.536 73 1.758 33
.804 2.036 34 1.411 48 .854 2.548 27 1.766 32
.805 2.045 51 1.417 84 .855 2.559 87 1.774 37
.806 2.054 72 1.424 23 .856 2.571 54 1.782 46
.807 2.063 97 1.430 63 .857 2.583 28 1.790 59
.808 2.073 25 1.437 07 .858 2.595 09 1.798 78
.809 2.082 57 1.443 53 .859 2.606 96 1.807 01
.810 2.091 92 1.450 01 .860 2.618 91 1.815 29

.811 2.101 32 1.456 52 .861 2.630 93 1.823 62
.812 2.110 75 1.463 06 .862 2.643 02 1.832 00
.813 2.120 22 1.469 62 .863 2.655 18 1.840 43
.814 2.129 73 1.476 21 .864 2.667 42 1.848 92
.815 2.139 27 1.482 83 .865 2.679 74 1.857 45
.816 2.148 86 1.489 48 .866 2.692 13 1.866 05
.817 2.158 49 1.496 15 .867 2.704 61 1.874 69
.818 2.168 16 1.502 86 .868 2.717 16 1.883 39
.819 2.177 87 1.509 59 .869 2.729 79 1.892 15
.820 2.187 63 1.516 35 .870 2.742 50 1.900 96

.821 2.197 42 1.523 14 .871 2.755 30 1.909 83
.822 2.207 26 1.529 96 .872 2.768 18 1.918 76
.823 2.217 14 1.536 81 .873 2.781 15 1.927 75
.824 2.227 07 1.543 69 .874 2.794 21 1.936 80
.825 2.237 04 1.550 60 .875 2.807 35 1.945 91
.826 2.247 05 1.557 54 .876 2.820 59 1.955 08
.827 2.257 12 1.564 51 .877 2.833 92 1.964 32
.828 2.267 22 1.571 52 .878 2.847 34 1.973 63
.829 2.277 38 1.578 56 .879 2.860 86 1.982 99
.830 2.287 58 1.585 63 .880 2.874 47 1.992 43

.831 2.297 83 1.592 73 .881 2.888 18 2.001 93
.832 2.308 12 1.599 87 .882 2.901 99 2.011 51
.833 2.318 47 1.607 04 .883 2.915 90 2.021 15
.834 2.328 86 1.614 25 .884 2.929 92 2.030 87
.835 2.339 31 1.621 49 .885 2.944 04 2.040 66
.836 2.349 81 1.628 76 .886 2.958 27 2.050 52
.837 2.360 36 1.636 07 .887 2.972 61 2.060 46
.838 2.370 96 1.643 42 .888 2.987 06 2.070 47
.839 2.381 61 1.650 81 .889 3.001 62 2.080 57
.840 2.392 32 1.658 23 .890 3.016 30 2.090 74

.841 2.403 08 1.665 69 .891 3.031 10 2.101 00
.842 2.413 90 1.673 18 .892 3.046 01 2.111 33
.843 2.424 77 1.680 72 .893 3.061 05 2.121 76
.844 2.435 70 1.688 30 .894 3.076 21 2.132 27
.845 2.446 68 1.695 91 .895 3.091 50 2.142 86
.846 2.457 73 1.703 57 .896 3.106 92 2.153 55
.847 2.468 83 1.711 26 .897 3.122 46 2.164 33
.848 2.479 99 1.719 00 .898 3.138 15 2.175 20
.849 2.491 22 1.726 78 .899 3.153 97 2.186 16
462 TABLES
TABLE C (concluded)
p log-2 logc - P loe P

1 -p 1 -p 1082 i--p 1 p
.900 3.169 92 2.197 22 .950 4.247 93 2.944 44

.901 3.186 03 2.208 39 .951 4.278 59 2.965 69
.902 3.202 27 2.219 65 .952 4.309 86 2.987 36
.903 3.218 67 2.231 01 .953 4.341 74 3.009 47
.904 3.235 22 2.242 48 .954 4.374 28 3.032 02
.905 3.251 92 2.254 06 .955 4.407 50 3.055 05
.906 3.268 78 2.265 74 .956 4.441 44 3.078 57
.907 3.285 80 2.277 54 .957 4.476 11 3.102 60
.908 3.302 99 2.289 46 .958 4.511 56 3.127 18
.909 3.320 34 2.301 49 .959 4.547 83 3.152 32
.910 3.337 87 2.313 63 .960 4.584 96 3.178 05

.911 3.355 57 2.325 91 .961 4.622 99 3.204 41
.912 3.373 46 2.338 30 .962 4.661 97 3.231 43
.913 3.391 53 2.350 83 .963 4.701 94 3.259 14
.914 3.409 79 2.363 48 .964 4.742 96 3.287 57
.915 3.428 24 2.376 27 .965 4.785 10 3.316 78
.916 3.446 89 2.389 20 .966 4.828 42 3.346 80
.917 3.465 74 2.402 27 .967 4.872 98 3.377 69
.918 3.484 80 2.415 48 .968 4.918 86 3.409 50
.919 3.504 07 2.428 84 .969 4.966 16 3.442 28
.920 3.523 56 2.442 35 .970 5.014 95 3.476 10

.921 3.543 28 2.456 01 .971 5.065 35 3.511 03
.922 3.563 22 2.469 84 .972 5.117 46 3.547 15
.923 3.583 40 2.483 82 .973 5.171 41 3.584 55
.924 3.603 82 2.497 98 .974 5.227 34 3.623 31
.925 3.624 49 2.512 31 .975 5.285 40 3.663 56
.926 3.645 42 2.526 81 .976 5.345 77 3.705 41
.927 3.666 60 2.541 49 .977 5.408 65 3.748 99
.928 3.688 06 2.556 37 .978 5.474 26 3.794 47
.929 3.709 79 2.571 43 .979 5.542 85 3.842 01
.930 3.731 80 2.586 69 .980 5.614 71 3.891 82

.931 3.754 11 2.602 15 .981 5.690 18 3.944 13
.932 3.776 72 2.617 83 .982 5.769 65 3.999 22
.933 3.799 64 2.633 71 .983 5.853 58 4.057 40
.934 3.822 88 2.649 82 .984 5.942 51 4.119 04
.935 3.846 45 2.666 16 .985 6.037 09 4.184 59
.936 3.870 36 2.682 73 .986 6.138 09 4.254 60
.937 3.894 63 2.699 55 .987 6.246 47 4.329 72
.938 3.919 25 2.716 62 .988 6.363 40 4.410 78
.939 3.944 24 2.733 94 .989 6.490 40 4.498 80
.940 3.969 63 2.751 54 .990 6.629 36 4.595 12

.941 3.995 41 2.769 41 .991 6.782 82 4.701 49
.942 4.021 60 2.787 56 .992 6.954 20 4.820 28
.943 4.048 22 2.806 02 .993 7.148 29 4.954 82
.944 4.075 29 2.824 77 .994 7.372 14 5.109 98
.945 4.102 81 2.843 85 .995 7.636 62 5.293 30
.946 4.130 81 2.863 26 .996 7.960 00 5.517 45
.947 4.159 30 2.883 01 .997 8.376 49 5.806 14
.948 4.188 30 2.903 11 .998 8.962 90 6.212 61
.949 4.217 84 2.923 58 .999 9.964 34 6.906 75
TABLES 463
TABLE D1
TRANSFORMATION OF RELATIVE CHANGES TO LOG-CHANGES
loge (1 +x) loge (1 — x) X log* (1 +x) log* (1 —x)
.001 .001 000 -.001 001 .051 .049 742 -.052 346
.002 .001 998 -.002 002 .052 .050 693 -.053 401
.003 .002 996 -.003 005 .053 .051 643 -.054 456
.004 .003 992 -.004 008 .054 .052 592 -.055 513
.005 .004 988 -.005 013 .055 .053 541 -.056 570
.006 .005 982 -.006 018 .056 .054 488 -.057 629
.007 .006 976 -.007 025 .057 .055 435 -.058 689
.008 .007 968 -.008 032 .058 .056 380 -.059 750
.009 .008 960 -.009 041 .059 .057 325 -.060 812
.010 .009 950 -.010 050 .060 .058 269 -.061 875
.011 .010 940 -.011 061 .061 .059 212 -.062940

.012 .011 929 -.012 073 .062 .060 154 -.064 005
.013 .012 916 -.013 085 .063 .061 095 -.065 072
.014 .013 903 -.014 099 .064 .062 035 -.066 140
.015 .014 889 -.015 114 .065 .062 975 -.067 209
.016 .015 873 -.016 129 .066 .063 913 -.068 279
.017 .016 857 -.017 146 .067 .064 851 -.069 350
.018 .017 840 -.018 164 .068 .065 788 -.070 422
.019 .018 822 -.019 183 .069 .066 724 -.071 496
.020 .019 803 -.020 203 .070 .067 659 -.072 571
.021 .020 783 -.021 224 .071 .068 593 -.073 647
.022 .021 761 -.022 246 .072 .069 526 -.074 724
.023 .022 739 -.023 269 .073 .070 458 -.075 802
.024 .023 717 -.024 293 .074 .071 390 -.076 881
.025 .024 693 -.025 318 .075 .072 321 -.077 962
.026 .025 668 -.026 344 .076 .073 250 -.079 043
.027 .026 642 -.027 371 .077 .074 179 -.080 126
.028 .027 615 -.028 399 .078 .075 107 -.081 210
.029 .028 587 -.029 429 .079 .076 035 -.082 295
.030 .029 559 -.030 459 .080 .076 961 -.083 382
.031 .030 529 -.031 491 .081 .077 887 -.084 469
.032 .031 499 -.032 523 .082 .078 811 -.085 558
.033 .032 467 -.033 557 .083 .079 735 -.086 648
.034 .033 435 -.034 591 .084 .080 658 -.087 739
.035 .034 401 -.035 627 .085 .081 580 -.088 831
.036 .035 367 -.036 664 .086 .082 501 -.089 925
.037 .036 332 -.037 702 .087 .083 422 -.091 019
.038 .037 296 -.038 741 .088 .084 341 -.092 115
.039 .038 259 -.039 781 .089 .085 260 -.093 212
.040 .039 221 -.040 822 .090 .086 178 -.094 311
.041 .040 182 -.041 864 .091 .087 095 -.095 410
.042 .041 142 -.042 908 .092 .088 011 -.096 511
.043 .042 101 -.043 952 .093 .088 926 -.097 613
.044 .043 059 -.044 997 .094 .089 841 -.098 716
.045 .044 017 -.046 044 .095 .090 754 -.099 820
.046 .044 973 -.047 092 .096 .091 667 -.100 926
.047 .045 929 -.048 140 .097 .092 579 -.102 033
.048 .046 884 -.049 190 .098 .093 490 -.103 141
.049 .047 837 -.050 241 .099 .094 401 -.104 250
.050 .048 790 -.051 293 .100 .095 310 -.105 361
464 TABLES
TABLE D1 (continued)
X logs (1 +*) loge (1 —X) X log? (1 +x) loge (1 —x)
.101 .096 219 -.106 472 .151 .140 631 -.163 696
.102 .097 127 -.107 585 .152 .141 500 -.164 875
.103 .098 034 -.108 699 .153 .142 367 -.166 055
.104 .098 940 -.109 815 .154 .143 234 -.167 236
.105 .099 845 -.110932 .155 .144 100 -.168 419
.106 .100 750 -.112 050 .156 .144 966 -.169 603
.107 .101 654 -.113 169 .157 .145 830 -.170 788
.108 .102 557 -.114 289 .158 .146 694 -.171 975
.109 .103 459 -.115411 .159 .147 558 -.173 164
.110 .104 360 -.116 534 .160 .148 420 -.174 353
.111 .105 261 -.117 658 .161 .149 282 -.175 545
.112 .106 160 -.118 784 .162 .150 143 -.176 737
.113 .107 059 -.119910 .163 .151 003 -.177 931
.114 .107 957 -.121 038 .164 .151 862 -.179 127
.115 .108 854 -.122 168 .165 .152 721 -.180 324
.116 .109 751 -.123 298 .166 .153 579 -.181 522
.117 .110 647 -.124 430 .167 .154 436 -.182 722
.118 .111 541 -.125 563 .168 .155 293 -.183 923
.119 .112435 -.126 698 .169 .156 149 -.185 125
.120 .113 329 -.127 833 .170 .157 004 -.186 330
.121 .114 221 -.128 970 .171 .157 858 -.187 535
.122 .115 113 -.130 109 .172 .158 712 -.188 742
.123 .116 004 -.131 248 .173 .159 565 -.189 951
.124 .116 894 -.132 389 .174 .160417 -.191 161
.125 .117 783 -.133 531 .175 .161 268 -.192 372
.126 .118 672 -.134 675 .176 .162 119 .193 585
.127 .119 559 -.135 820 .177 .162 969 -.194 799
.128 .120 446 -.136 966 .178 .163 818 -.196015
.129 .121 332 -.138 113 .179 .164 667 -.197 232
.130 .122218 -.139 262 .180 .165 514 -.198 451
.131 .123 102 -.140412 .181 .166 362 -.199 671

.132 .123 986 -.141 564 .182 .167 208 -.200 893
.133 .124 869 -.142 716 .183 .168 054 -.202 116
.134 .125 751 -.143 870 .184 .168 899 -.203 341
.135 .126 633 -.145 026 .185 .169 743 -.204 567
.136 .127 513 -.146 183 .186 .170 586 -.205 795
.137 .128 393 -.147 341 .187 .171 429 -.207 024
.138 .129 272 -.148 500 .188 .172 271 -.208 255
.139 .130 151 -.149 661 .189 .173 113 -.209 487
.140 .131 028 -.150 823 .190 .173 953 -.210 721
.141 .131 905 -.151 986 .191 .174 793 -.211 956
.142 .132 781 -.153 151 .192 .175 633 -.213 193
.143 .133 656 -.154 317 .193 .176 471 -.214 432
.144 .134 531 -.155 485 .194 .177 309 -.215 672
.145 .135 405 -.156 654 .195 .178 146 -.216913
.146 .136 278 -.157 824 .196 .178 983 -.218 156
.147 .137 150 -.158 996 .197 .179 818 -.219 401
.148 .138 021 -.160 169 .198 .180 654 -.220 647
.149 .138 892 -.161 343 .199 .181 488 -.221 894
.150 .139 762 -.162 519 .200 .182 322 -.223 144
TABLES 465
X loge (1 + x) loge (1 — X) loge (1 +x) log,, (1 — x)
.201 .183 155 -.224 394 .251 .223 943 -.289 016
.202 .183 987 -.225 647 .252 .224 742 -.290 352
.203 .184 818 -.226 901 .253 .225 541 -.291 690
.204 .185 649 -.228 156 .254 .226 338 -.293 030
.205 .186 480 -.229 413 .255 .227 136 -.294 371
.206 .187 309 -.230 672 .256 .227 932 -.295 714
.207 .188 138 -.231 932 .257 .228 728 -.297 059
.208 .188 966 -.233 194 .258 .229 523 -.298 406
.209 .189 794 -.234 457 .259 .230 318 -.299 755
.210 .190 620 -.235 722 .260 .231 112 -.301 105
.211 .191 446 -.236 989 .261 .231 905 -.302 457
.212 .192 272 -.238 257 .262 .232 698 -.303 811
.213 .193 097 -.239 527 .263 .233 490 -.305 167
.214 .193 921 -.240 798 .264 .234 281 -.306 525
.215 .194 744 -.242 072 .265 .235 072 -.307 885
.216 .195 567 -.243 346 .266 .235 862 -.309 246
.217 .196 389 -.244 623 .267 .236 652 -.310610
.218 .197 210 -.245 901 .268 .237 441 -.311 975
.219 .198 031 -.247 180 .269 .238 229 -.313 342
.220 .198 851 -.248 461 .270 .239 017 -.314 711
.221 .199 670 -.249 744 .271 .239 804 -.316 082
.222 .200 489 -.251 029 .272 .240 590 -.317 454
.223 .201 307 -.252 315 .273 .241 376 -.318 829
.224 .202 124 -.253 603 .274 .242 162 -.320 205
.225 .202 941 -.254 892 .275 .242 946 -.321 584
.226 .203 757 -.256 183 .276 .243 730 -.322 964
.227 .204 572 -.257 476 .277 .244 514 -.324 346
.228 .205 387 -.258 771 .278 .245 296 -.325 730
.229 .206 201 -.260 067 .279 .246 079 -.327 116
.230 .207 014 -.261 365 .280 .246 860 -.328 504
.231 .207 827 -.262 664 .281 .247 641 -.329 894
.232 .208 639 -.263 966 .282 .248 421 -.331 286
.233 .209 450 -.265 268 .283 .249 201 -.332 679
.234 .210 261 -.266 573 .284 .249 980 -.334 075,
.235 .211 071 -.267 879 .285 .250 759 -.335 473
.236 .211 880 -.269 187 .286 .251 537 -.336 872
.237 .212 689 -.270 497 .287 .252 314 -.338 274
.238 .213 497 -.271 809 .288 .253 091 -.339 677
.239 .214 305 -.273 122 .289 .253 867 -.341 083
.240 .215 111 -.274 437 .290 .254 642 -.342 490
.241 .215 918 -.275 754 .291 .255 417 -.343 900
.242 .216 723 -.277 072 .292 .256 191 -.345 311
.243 .217 528 -.278 392 .293 .256 965 -.346 725
.244 .218 332 -.279 714 .294 .257 738 -.348 140
.245 .219 136 -.281 038 .295 .258 511 -.349 557
.246 .219 938 -.282 363 .296 .259 283 -.350 977
.247 .220 741 -.283 690 .297 .260 054 -.352 398
.248 .221 542 -.285 019 .298 .260 825 -.353 822
.249 .222 343 -.286 350 .299 .261 595 -.355 247
.250 .223 144 -.287 682 .300 .262 364 -.356 675
466 TABLES
X loge (1 +X) Ioge (1 — x) X loge (l+x) loge (1 — x)
.301 .263 133 -.358 105 .351 .300 845 -.432 323
.302 .263 902 -.359 536 .352 .301 585 -.433 865
.303 .264 669 -.360 970 .353 .302 324 -.435 409
.304 .265 436 -.362 406 .354 .303 063 -.436 956
.305 .266 203 -.363 843 .355 .303 801 -.438 505
.306 .266 969 -.365 283 .356 .304 539 -.440 057
.307 .267 734 -.366 725 .357 .305 276 -.441 611
.308 .268 499 -.368 169 .358 .306 013 -.443 167
.309 .269 263 -.369 615 .359 .306 749 -.444 726
.310 .270 027 -.371 064 .360 .307 485 -.446 287
.311 .270 790 -.372 514 .361 .308 220 -.447 851
.312 .271 553 -.373 966 .362 .308 954 -.449 417
.313 .272 315 -.375 421 .363 .309 688 -.450 986
.314 .273 076 -.376 878 .364 .310 422 -.452 557
.315 .273 837 -.378 336 .365 .311 154 -.454 130
.316 .274 597 -.379 797 .366 .311 887 -.455 706
.317 .275 356 -.381 260 .367 .312 619 -.457 285
.318 .276 115 -.382 726 .368 .313 350 -.458 866
.319 .276 874 -.384 193 .369 .314081 -.460 449
.320 .277 632 -.385 662 .370 .314811 -.462 035
.321 .278 389 -.387 134 .371 .315 540 -.463 624
.322 .279 146 -.388 608 .372 .316 270 -.465 215
.323 .279 902 -.390 084 .373 .316 998 -.466 809
.324 .280 657 -.391 562 .374 .317 726 -.468 405
.325 .281 412 -.393 043 .375 .318 454 -.470 004
.326 .282 167 -.394 525 .376 .319 181 -.471 605
.327 .282 921 -.396 010 .377 .319 907 -.473 209
.328 .283 674 -.397 497 .378 .320 633 -.474 815
.329 .284 427 -.398 986 .379 .321 359 -.476 424
.330 .285 179 -.400 478 .380 .322 083 -.478 036
.331 .285 931 -.401 971 .381 .322 808 - .479 650
.332 .286 682 -.403 467 .382 .323 532 -.481 267
.333 .287 432 -.404 965 .383 .324 255 -.482 886
.3.34 .288 182 -.406 466 .384 .324 978 -.484 508
.335 .288 931 -.407 968 .385 .325 700 -.486 133
.336 .289 680 -.409 473 .386 .326 422 -.487 760
.337 .290 428 .410 980 .387 .327 143 .489 390
.338 .291 176 -.412 490 .388 .327 864 .491 023
.339 .291 923 -.414 001 .389 .328 584 -.492 658
.340 .292 670 -.415 515 .390 .329 304 - .494 296
.341 .293 416 -.417 032 .391 .330 023 .495 937
.342 .294 161 -.418 550 .392 .330 742 -.497 580
.343 .294 906 -.420 071 .393 .331 460 -.499 226
.344 .295 650 -.421 594 .394 .332 177 -.500 875
.345 .296 394 -.423 120 .395 .332 894 -.502 527
.346 .297 137 -.424 648 .396 .333 611 .504 181
.347 .297 880 -.426 178 .397 .334 327 -.505 838
.348 .298 622 -.427 711 .398 .335 043 .507 498
.349 .299 364 -.429 246 .399 .335 758 -.509 160
350 .300 105 -.430 783 .400 .336 472 -.510 826
TABLES 467
X loge (1 +*) loge (1 - X) loge (1 + A) loge (1 —x)
.401 .337 186 -.512 494 .451 .372 253 -.599 657
.402 .337 900 -.514 165 .452 .372 942 -.601 480
.403 .338 613 -.515 838 .453 .373 630 -.603 306
.404 .339 325 -.517 515 .454 .374 318 -.605 136
.405 .340 037 -.519 194 .455 .375 006 -.606 969
.406 .340 749 -.520 876 .456 .375 693 -.608 806
.407 .341 460 -.522 561 .457 .376 380 -.610 646
.408 .342 170 -.524 249 .458 .377 066 -.612 489
.409 .342 880 -.525 939 .459 .377 751 -.614 336
.410 .343 590 -.527 633 .460 .378 436 -.616 186
.411 .344 299 -.529 329 .461 .379 121 -.618 040
.412 .345 007 -.531 028 .462 .379 805 -.619 897
.413 .345 715 -.532 730 .463 .380 489 -.621 757
.414 .346 423 -.534 435 .464 .381 172 -.623 621
.415 .347 130 -.536 143 .465 .381 855 -.625 489
.416 .347 836 -.537 854 .466 .382 538 -.627 359
.417 .348 542 -.539 568 .467 .383 219 -.629 234
.418 .349 247 -.541 285 .468 .383 901 -.631 112
.419 .349 952 -.543 005 .469 .384 582 -.632 993
.420 .350 657 -.544 727 .470 .385 262 -.634 878
.421 .351 361 -.546 453 .471 .385 942 -.636 767
.422 .352 064 -.548 181 .472 .386 622 -.638 659
.423 .352 767 -.549 913 .473 .387 301 -.640 555
.424 .353 470 -.551 648 .474 .387 980 -.642 454
.425 .354 172 -.553 385 .475 .388 658 -.644 357
.426 .354 873 -.555 126 .476 .389 336 -.646 264
.427 .355 574 -.556 870 .477 .390 013 -.648 174
.428 .356 275 -.558 616 .478 .390 690 -.650 088
.429 .356 975 -.560 366 .479 .391 366 -.652 005
.430 .357 674 -.562 119 .480 .392 042 -.653 926
.431 .358 374 -.563 875 .481 .392 718 -.655 851
.432 .359 072 -.565 634 .482 .393 393 -.657 780
.433 .359 770 -.567 396 .483 .394 067 -.659 712
.434 .360 468 -.569 161 .484 .394 741 -.661 649
.435 .361 165 -.570 930 .485 .395 415 -.663 588
.436 .361 861 -.572 701 .486 .396 088 -.665 532
.437 .362 558 -.574 476 .487 .396 761 -.667 479
.438 .363 253 -.576 253 .488 .397 433 -.669 431
.439 .363 948 -.578 034 .489 .398 105 -.671 386
.440 .364 643 -.579 818 .490 .398 776 -.673 345
.441 .365 337 -.581 606 .491 .399 447 -.675 307
.442 .366 031 -.583 396 .492 .400 118 -.677 274
.443 .366 724 -.585 190 .493 .400 788 -.679 244
.444 .367 417 -.586 987 .494 .401 457 -.681 219
.445 .368 109 -.588 787 .495 .402 126 -.683 197
.446 .368 801 -.590 591 .496 .402 795 -.685 179
.447 .369 492 -.592 397 .497 .403 463 -.687 165
.448 .370 183 -.594 207 .498 .404 131 -.689 155
.449 .370 874 -.596 020 .499 .404 798 -.691 149
.450 .371 564 -.597 837 .500 .405 465 -.693 147
468 TABLES
X log<>(l + *) log* (1 — x) X loge (\+x) loge (1 — X)
.501 .406 132 -.695 149 .551 .438 900 -.800 732
.502 .406 798 -.697 155 .552 .439 544 -.802 962
.503 .407 463 -.699 165 .553 .440 189 -.805 197
.504 .408 128 -.701 179 .554 .440 832 -.807 436
.505 .408 793 -.703 198 .555 .441 476 -.809 681
.506 .409 457 -.705 220 .556 .442 118 -.811 931
.507 .410 121 -.707 246 .557 .442 761 -.814 186
.508 .410 784 -.709 277 .558 .443 403 -.816445
.509 .411 447 -.711 311 .559 .444 045 -.818 710
.510 .412 110 -.713 350 .560 .444 686 -.820 981
.511 .412 772 -.715 393 .561 .445 327 -.823 256
.512 .413 433 -.717 440 .562 .445 967 -.825 536
.513 .414 094 -.719 491 .563 .446 607 -.827 822
.514 .414 755 -.721 547 .564 .447 247 -.830 113
.515 .415415 -.723 606 .565 .447 886 -.832 409
.516 .416 075 -.725 670 .566 .448 525 -.834 711
.517 .416 735 -.727 739 .567 .449 163 -.837 018
.518 .417 394 -.729 811 .568 .449 801 -.839 330
.519 .418 052 -.731 888 .569 .450 438 -.841 647
.520 .418 710 -.733 969 .570 .451 076 -.843 970
.521 .419 368 - 736 055 .571 .451 712 -.846 298
.522 .420 025 -.738 145 .572 .452 349 -.848 632
.523 .420 682 -.740 239 .573 .452 985 -.850 971
.524 .421 338 -.742 337 .574 .453 620 -.853 316
.525 .421 994 -.744 440 .575 .454 255 -.855 666
.526 .422 650 -.746 548 .576 .454 890 -.858 022
.527 .423 305 -.748 660 .577 .455 524 -.860 383
.528 .423 960 -.750 776 .578 .456 158 -.862 750
.529 .424 614 -.752 897 .579 .456 792 -.865 122
.530 .425 268 -.755 023 .580 .457 425 -.867 501
.531 .425 921 -.757 153 .581 .458 058 -.869 884
.532 .426 574 -.759 287 582 .458 690 -.872 274
.533 .427 227 -.761 426 .583 .459 322 -.874 669
.534 .427 879 -.763 570 .584 .459 953 -.877 070
.535 .428 530 -.765 718 .585 .460 584 -.879 477
.536 .429 182 -.767 871 .586 .461 215 -.881 889
.537 .429 832 -.770 028 .587 .461 845 -.884 308
.538 .430 483 -.772 190 .588 .462 475 -.886 732
.539 .431 133 -.774 357 .589 .463 105 -.889 162
.540 .431 782 -.776 529 .590 .463 734 -.891 598
.541 .432 432 -.778 705 .591 .464 363 -.894 040
.542 .433 080 -.780 886 .592 .464 991 -.896 488
.543 .433 729 -.783 072 .593 .465 619 -.898 942
.544 .434 376 -.785 262 .594 .466 247 -.901 402
.545 .435 024 -.787 458 .595 .466 874 -.903 868
.546 .435 671 -.789 658 .596 .467 500 -.906 340
.547 .436 318 -.791 863 .597 .468 127 -.908 819
.548 .436 964 -.794 073 .598 .468 753 -.911 303
.549 .437 610 -.796 288 .599 .469 378 -.913 794
.550 .438 255 -.798 508 .600 .470 004 -.916 291
TABLES 469
X loge (1 +x) logs (1 —JC) loge (1 +*) loge (1 — x)
.601 .470 628 -.918 794 .651 .501 381 -1.052 68

.602 .471 253 -.921 303 .652 .501 987 -1.055 55
.603 .471 877 -.923 819 .653 .502 592 -1.058 43
.604 .472 501 -.926 341 .654 .503 197 -1.061 32
.605 .473 124 -.928 870 .655 .503 801 -1.064 21
.606 .473 747 -.931 404 .656 .504 405 -1.067 11
.607 .474 369 -.933 946 .657 .505 009 -1.070 02
.608 .474 991 -.936 493 .658 .505 612 -1.072 94
.609 .475 613 -.939 048 .659 .506 215 -1.075 87
.610 .476 234 -.941 609 .660 .506 818 -1.078 81
.611 .476 855 — .944 176 .661 .507 420 -1.081 76

.612 .477 476 -.946 750 .662 .508 022 -1.084 71
.613 .478 096 -.949 331 .663 .508 623 -1.087 67
.614 .478 716 -.951 918 .664 .509 224 -1.090 64
.615 .479 335 -.954 512 .665 .509 825 -1.093 62
.616 .479 954 -.957 113 .666 .510 426 -1.096 61
.617 .480 573 -.959 720 .667 .511 026 -1.099 61
.618 .481 191 -.962 335 .668 .511 625 -1.102 62
.619 .481 809 -.964 956 .669 .512 225 -1.105 64
.620 .482 426 -.967 584 .670 .512 824 -1.108 66
.621 .483 043 -.970 219 .671 .513 422 -1.111 70

.622 .483 660 - 972 861 .672 .514 021 -1.114 74
.623 .484 276 -.975 510 .673 .514 618 -1.117 80
.624 .484 892 -.978 166 .674 .515 216 -1.120 86
.625 .485 508 -.980 829 .675 .515 813 -1.123 93
.626 .486 123 -.983 499 .676 .516410 -1.127 01
.627 .486 738 -.986 177 .677 .517 006 -1.130 10
.628 .487 352 -.988 861 .678 .517 603 -1.133 20
.629 .487 966 -.991 553 .679 .518 198 -1.136 31
.630 .488 580 -.994 252 .680 .518 794 -1.13943
.631 .489 193 -.996 959 .681 .519 389 -1.142 56

.632 .489 806 -.999 672 .682 .519 984 -1.145 70
.633 .490 419 -1.002 39 .683 .520 578 -1.148 85
.634 .491 031 -1.005 12 .684 .521 172 -1.152 01
.635 .491 643 -1.007 86 .685 .521 766 -1.155 18
.636 .492 254 -1.010 60 .686 .522 359 -1.158 36
.637 .492 865 -1.013 35 .687 .522 952 -1.161 55
.638 .493 476 1.016 11 .688 .523 544 -1.164 75
.639 .494 086 -1.018 88 .689 .524 137 -1.167 96
.640 .494 696 -1.021 65 .690 .524 729 -1.171 18
.641 .495 306 -1.024 43 .691 .525 320 -1.17441

.642 .495 915 -1.027 22 .692 .525 911 -1.177 66
.643 .496 524 -1.030 02 .693 .526 502 -1.180 91
.644 .497 132 -1.032 82 .694 .527 093 -1.184 17
.645 .497 740 -1.035 64 .695 .527 683 -1.187 44
.646 .498 348 -1.038 46 .696 .528 273 -1.190 73
.647 .498 955 -1.041 29 .697 .528 862 -1.194 02
.648 .499 562 -1.044 12 .698 .529 451 -1.197 33
.649 .500 169 -1.046 97 .699 .530 040 -1.200 65
.650 .500 775 -1.049 82 .700 .530 628 -1.203 97
470 TABLES
X loge (1 +*) loge (1 —jc) X loge (1 +*) loge (1 —X)
.701 .531 216 -1.207 31 .751 .560 187 -1.390 30

.702 .531 804 -1.21066 .752 .560 758 -1.394 33
.703 .532 391 1.214 02 .753 .561 329 -1.398 37
.704 .532 978 -1.217 40 .754 .561 899 -1.402 42
.705 .533 565 -1.220 78 .755 .562 469 -1.406 50
.706 .534 151 -1.224 18 .756 .563 038 -1.410 59
.707 .534 737 -1.227 58 .757 .563 608 -1.414 69
.708 .535 323 1.231 00 .758 .564 177 -1.418 82
.709 .535 908 -1.234 43 .759 .564 745 -1.422 96
.710 .536 493 -1.237 87 .760 .565 314 -1.427 12
.711 .537 078 -1.241 33 .761 .565 882 -1.431 29

.712 .537 662 -1.244 79 .762 .566 450 -1.435 48
.713 .538 246 -1.248 27 .763 .567 017 -1.439 70
.714 .538 830 1.251 76 .764 .567 584 -1.443 92
.715 .539 413 -1.255 27 .765 .568 151 -1.448 17
.716 .539 996 -1.258 78 .766 .568 717 -1.452 43
.717 .540 579 -1.262 31 .767 .569 283 -1.456 72
.718 .541 161 -1.265 85 .768 .569 849 -1.461 02
.719 .541 743 -1.269 40 .769 .570 414 -1.465 34
.720 .542 324 -1.272 97 .770 .570 980 -1.469 68
.721 .542 906 -1.276 54 .771 .571 544 -1.474 03

.722 .543 486 -1.280 13 .772 .572 109 -1.478 41
.723 .544 067 -1.283 74 .773 .572 673 -1.482 81
.724 .544 647 -1.287 35 .774 .573 237 -1.487 22
.725 .545 227 -1.290 98 .775 .573 800 -1.491 65
.726 .545 807 1.294 63 .776 .574 364 -1.496 11
.727 .546 386 -1.298 28 .777 .574 927 -1.500 58
.728 .546 965 -1.301 95 .778 .575 489 -1.505 08
.729 .547 543 -1.305 64 .779 .576 051 -1.509 59
.730 .548 121 -1.309 33 .780 .576 613 -1.514 13
.731 .548 699 -1.313 04 .781 .577 175 -1.518 68

.732 .549 277 -1.31677 .782 .577 736 -1.523 26
.733 .549 854 -1.320 51 .783 .578 297 -1.527 86
.734 .550 431 -1.324 26 .784 .578 858 -1.532 48
.735 .551 007 -1.328 03 .785 .579 418 -1.537 12
.736 .551 584 -1.331 81 .786 .579 978 1.541 78
.737 .552 159 -1.335 60 .787 .580 538 -1.546 46
.738 .552 735 -1.339 41 .788 .581 098 -1.551 17
.739 .553 310 -1.343 23 .789 .581 657 -1.555 90
.740 .553 885 -1.347 07 .790 .582 216 -1.560 65
.741 .554 460 - 1.350 93 .791 .582 774 -1.565 42

.742 .555 034 -1.354 80 .792 .583 332 -1.570 22
.743 .555 608 1.358 68 .793 .583 890 -1.575 04
.744 .556 181 -1.362 58 .794 .584 448 -1.579 88
.745 .556 755 -1.366 49 .795 .585 005 -1.584 75
.746 .557 327 -1.370 42 .796 .585 562 -1.589 64
.747 .557 900 -1.374 37 .797 .586 119 -1.594 55
.748 .558 472 -1.378 33 .798 .586 675 -1.599 49
.749 .559 044 -1.382 30 .799 .587 23! -1.604 45
.750 .559 616 1.386 29 .800 .587 787 -1.609 44
TABLES 471
X loge (1 +*) loge (1 — X) X loge (1 +*) loge (1 — X)
.801 .588 342 -1.61445 .851 .615 726 -1.903 81

.802 .588 897 -1.619 49 .852 .616 266 1.910 54
.803 .589 452 -1.624 55 .853 .616 806 -1.917 32
.804 .590 006 -1.629 64 .854 .617 345 -1.924 15
.805 .590 561 -1.634 76 .855 .617 885 -1.931 02
.806 .591 114 -1.639 90 .856 .618 424 -1.937 94
.807 .591 668 -1.645 07 .857 .618 962 -1.944 91
.808 .592 221 -1.650 26 .858 .619 501 -1.951 93
.809 .592 774 -1.655 48 .859 .620 039 -1.959 00
.810 .593 327 -1.660 73 .860 .620 576 -1.966 11
.811 .593 879 -1.666 01 .861 .621 114 -1.973 28

.812 .594 431 -1.671 31 .862 .621 651 -1.980 50
.813 .594 983 -1.676 65 .863 .622 188 -1.987 77
.814 .595 534 -1.682 01 .864 .622 725 -1.995 10
.815 .596 085 -1.687 40 .865 .623 261 -2.002 48
.816 .596 636 -1.692 82 .866 .623 797 -2.009 92
.817 .597 187 -1.698 27 .867 .624 333 -2.017 41
.818 .597 737 -1.703 75 .868 .624 868 -2.024 95
.819 .598 287 -1.709 26 .869 .625 404 -2.032 56
.820 .598 837 -1.714 80 .870 .625 938 -2.040 22
.821 .599 386 -1.720 37 .871 .626 473 -2.047 94

.822 .599 935 -1.725 97 .872 .627 007 -2.055 73
.823 .600 483 -1.731 61 .873 .627 541 -2.063 57
.824 .601 032 -1.737 27 .874 .628 075 -2.071 47
.825 .601 580 -1.742 97 .875 .628 609 -2.079 44
.826 .602 128 -1.748 70 .876 .629 142 -2.087 47
.827 .602 675 -1.754 46 .877 .629 675 -2.095 57
.828 .603 222 -1.760 26 .878 .630 207 -2.103 73
.829 .603 769 -1.766 09 .879 .630 740 -2.111 96
.830 .604 316 -1.771 96 .880 .631 272 -2.120 26
.831 .604 862 -1.777 86 .881 .631 804 -2.128 63

.832 .605 408 -1.783 79 .882 .632 335 -2.137 07
.833 .605 954 -1.789 76 .883 .632 866 -2.145 58
.834 .606 499 -1.795 77 .884 .633 397 -2.154 17
.835 .607 044 1.801 81 .885 .633 928 -2.162 82
.836 .607 589 -1.807 89 .886 .634 458 -2.171 56
.837 .608 134 -1.81401 .887 .634 988 -2.180 37
.838 .608 678 -1.820 16 .888 .635 518 -2.189 26
.839 .609 222 -1.826 35 .889 .636 048 -2.198 23
.840 .609 766 -1.832 58 .890 .636 577 -2.207 27
.841 .610 309 -1.838 85 .891 .637 106 -2.21641

.842 .610 852 -1.845 16 .892 .637 634 -2.225 62
.843 .611 395 -1.851 51 .893 .638 163 -2.234 93
.844 .611 937 -1.857 90 .894 .638 691 -2.244 32
.845 .612 479 -1.864 33 .895 .639 219 -2.253 79
.846 .613 021 -1.870 80 .896 .639 746 -2.263 36
.847 .613 563 -1.877 32 .897 .640 274 -2.273 03
.848 .614 104 -1.883 87 .898 .640 801 -2.282 78
.849 .614 645 -1.890 48 .899 .641 327 -2.292 63
.850 .615 186 -1.897 12 .900 .641 854 -2.302 59
472 TABLES
TABLE D1 (concluded)
X loge (1 +x) loge (1 — X) loge (1 +x) loge (1 — x)
.901 .642 380 -2.312 64 .951 .668 342 -3.015 93

.902 .642 906 -2.322 79 .952 .668 854 -3.036 55
.903 .643 432 -2.333 04 .953 .669 367 -3.057 61
.904 .643 957 -2.343 41 .954 .669 879 -3.079 11
.905 .644 482 -2.353 88 .955 .670 390 -3.101 09
.906 .645 007 -2.364 46 .956 .670 902 -3.123 57
.907 .645 531 -2.375 16 .957 .671 413 -3.146 56
.908 .646 056 -2.385 97 .958 .671 924 -3.170 09
.909 .646 580 -2.396 90 .959 .672 434 -3.194 18
.910 .647 103 -2.407 95 .960 .672 944 -3.218 88
.911 .647 627 -2.419 12 .961 .673 455 -3.244 19

.912 .648 150 -2.430 42 .962 .673 964 -3.270 17
.913 .648 673 -2.441 85 .963 .674 474 -3.296 84
.914 .649 195 -2.453 41 .964 .674 983 -3.324 24
.915 .649 718 -2.465 10 .965 .675 492 -3.352 41
.916 .650 240 -2.476 94 .966 .676 001 -3.381 39
.917 .650 761 -2.488 91 .967 .676 510 -3.411 25
.918 .651 283 -2.501 04 .968 .677 018 -3.442 02
.919 .651 804 -2.513 31 .969 .677 526 -3.473 77
.920 .652 325 -2.525 73 .970 .678 034 -3.506 56
.921 .652 846 -2.538 31 .971 .678 541 -3.540 46

.922 .653 366 -2.551 05 .972 .679 048 -3.575 55
.923 .653 886 -2.563 95 .973 .679 555 -3.611 92
.924 .654 406 -2.577 02 .974 .680 062 -3.649 66
.925 .654 926 -2.590 27 .975 .680 568 -3.688 88
.926 .655 445 -2.603 69 .976 .681 075 -3.729 70
.927 .655 964 -2.617 30 .977 .681 581 -3.772 26
.928 .656 483 -2.631 09 .978 .682 086 -3.816 71
.929 .657 002 -2.645 08 .979 .682 592 -3.863 23
.930 .657 520 -2.659 26 .980 .683 097 -3.912 02
.931 .658 038 -2.673 65 .981 .683 602 -3.963 32

.932 .658 556 -2.688 25 .982 .684 106 -4.017 38
.933 .659 073 -2.703 06 .983 .684 611 -4.074 54
.934 .659 590 -2.718 10 984 .685 115 -4.135 17
.935 .660 107 -2.733 37 .985 .685 619 -4.199 71
.936 .660 624 -2.748 87 .986 .686 123 -4.268 70
.937 .661 140 -2.764 62 .987 .686 626 -4.342 81
.938 .661 657 -2.780 62 .988 .687 129 -4.422 85
.939 .662 172 -2.796 88 .989 .687 632 -4.509 86
.940 .662 688 -2.813 41 .990 .688 135 -4.605 17
.941 .663 203 -2.830 22 .991 .688 637 -4.710 53

.942 .663 718 -2.847 31 .992 .689 139 -4.828 31
.943 .664 233 -2.864 70 .993 .689 641 -4.961 85
.944 .664 748 -2.882 40 .994 .690 143 -5.11600
.945 .665 262 -2.900 42 .995 .690 644 -5.298 32
.946 .665 776 -2.918 77 .996 .691 145 -5.521 46
.947 .666 290 -2.937 46 .997 .691 646 -5.809 14
.948 .666 803 -2.956 51 .998 .692 147 -6.214 61
.949 .667 316 -2.975 93 .999 .692 647 -6.907 76
.950 .667 819 -2.995 73 1.000 .693 147 — 00
TABLES 473
TABLE D2
TRANSFORMATION OF LOG-CHANGES TO RELATIVE CHANGES
y ev — l e-y — i y ey -\ e~y-1
.001 .001 001 -.001 000 .051 .052 323 -.049 721
.002 .002 002 -.001 998 .052 .053 376 -.050 671
.003 .003 005 -.002 996 .053 .054 430 -.051 620
.004 .004 008 -.003 992 .054 .055 485 -.052 568
.005 .005 013 -.004 988 .055 .056 541 -.053 515
.006 .006 018 -.005 982 .056 .057 598 -.054 461
.007 .007 025 -.006 976 .057 .058 656 -.055 406
.008 .008 032 -.007 968 .058 .059 715 -.056 350
.009 .009 041 -.008 960 .059 .060 775 -.057 293
.010 .010 050 -.009 950 .060 .061 837 -.058 235
.011 .011 061 -.010 940 .061 .062 899 -.059 177
.012 .012 072 -.011 928 .062 .063 962 -.060117
.013 .013 085 -.012 916 .063 .065 027 -.061 057
.014 .014 098 -.013 902 .064 .066 092 -.061 995
.015 .015 113 -.014 888 .065 .067 159 -.062 933
.016 .016 129 -.015 873 .066 .068 227 -.063 869
.017 .017 145 -.016 856 .067 .069 295 -.064 805
.018 .018 163 -.017 839 .068 .070 365 -.065 740
.019 .019 182 -.018 821 .069 .071 436 -.066 673
.020 .020 201 -.019 801 .070 .072 508 -.067 606
.021 .021 222 -.020 781 .071 .073 581 -.068 538
.022 .022 244 -.021 760 .072 .074 655 -.069 469
.023 .023 267 -.022 738 .073 .075 731 -.070 399
.024 .024 290 -.023 714 .074 .076 807 -.071 328
.025 .025 315 -.024 690 .075 .077 884 -.072 257
.026 .026 341 -.025 665 .076 .078 963 -.073 184
.027 .027 368 -.026 639 .077 .080 042 -.074 110
.028 .028 396 -.027 612 .078 .081 123 -.075 036
.029 .029 425 -.028 584 .079 .082 204 -.075 960
.030 .030 455 -.029 554 .080 .083 287 -.076 884
.031 .031 486 -.030 524 .081 .084 371 -.077 806
.032 .032 518 -.031 493 .082 .085 456 -.078 728
.033 .033 551 -.032 461 .083 .086 542 -.079 649
.034 .034 585 -.033 429 .084 .087 629 -.080 569
.035 .035 620 -.034 395 .085 .088 717 -.081 488
.036 .036 656 -.035 360 .086 .089 806 -.082 406
.037 .037 693 -.036 324 .087 .090 897 -.083 323
.038 .038 731 -.037 287 .088 .091 988 -.084 239
.039 .039 770 -.038 249 .089 .093 081 -.085 154
.040 .040 811 -.039 211 .090 .094 174 -.086 069
.041 .041 852 -.040171 .091 .095 269 -.086 982

.042 .042 894 -.041 130 .092 .096 365 -.087 895
.043 .043 938 -.042 089 .093 .097 462 -.088 807
.044 .044 982 -.043 046 .094 .098 560 -.089 717
.045 .046 028 -.044 003 .095 .099 659 -.090 627
.046 .047 074 -.044 958 .096 .100 759 -.091 536
.047 .048 122 -.045 913 .097 .101 860 -.092 444
.048 .049 171 -.046 866 .098 .102 963 -.093 351
.049 .050 220 -.047 819 .099 .104 066 -.094 257
.050 .051 271 -.048 771 .100 .105 171 -.095 163
474 TABLES
y ev — \ e~y - 1 y ev — \ e~y -\
.101 .106 277 -.096 067 .151 .162 997 -.140 152
.102 .107 383 -.096970 .152 .164 160 -.141 012
.103 .108 491 -.097 873 .153 .165 325 -.141 870
.104 .109 600 -.098 775 .154 .166 491 -.142 728
.105 .110711 -.099 675 .155 .167 658 -.143 585
.106 .111 822 -.100 575 .156 .168 826 -.144 441
.107 .112934 -.101 474 .157 .169 996 -.145 296
.108 .114 048 -.102 372 .158 .171 166 -.146 150
.109 .115 162 -.103 270 .159 .172 338 -.147 004
.110 .116 278 -.104 166 .160 .173 511 -.147 856
.111 .117 395 -.105 061 .161 .174 685 -.148 708
.112 .118 513 -.105 956 .162 .175 860 -.149 559
.113 .119 632 -.106 849 .163 .177 037 -.150 409
.114 .120 752 -.107 742 .164 .178 214 -.151 258
.115 .121 873 -.108 634 .165 .179 393 -.152 106
.116 .122 996 -.109 525 .166 .180 573 -.152 954
.117 .124 119 -.110415 .167 .181 754 -.153 800
.118 .125 244 -.111 304 .168 .182 937 -.154 646
.119 .126 370 -.112 192 .169 .184 120 -.155 491
.120 .127 497 -.113 080 .170 .185 305 -.156 335
.121 .128 625 -.113 966 .171 .186 491 -.157 178
.122 .129 754 -.114 852 .172 .187 678 -.158 021
.123 .130 884 -.115 736 .173 .188 866 -.158 862
.124 .132016 -.116 620 .174 .190 056 -.159 703
.125 .133 148 -.117 503 .175 .191 246 -.160 543
.126 .134 282 -.118 385 .176 .192 438 -.161 382
.127 .135 417 -.119 266 .177 .193 631 -.162 220
.128 .136 553 -.120 147 .178 .194 825 -.163 058
.129 .137 690 -.121 026 .179 .196 021 -.163 894
.130 .138 828 -.121 905 .180 .197 217 -.164 730
.131 .139 968 -.122 782 .181 .198 415 -.165 565
.132 .141 108 -.123 659 .182 .199 614 -.166 399
.133 .142 250 -.124 535 .183 .200 814 -.167 232
.134 .143 393 -.125 410 .184 .202 016 -.168 064
.135 .144 537 -.126 284 .185 .203 218 -.168 896
.136 .145 682 -.127 157 .186 .204 422 -.169 726
.137 .146 828 -.128 030 .187 .205 627 -.170 556
.138 .147 976 -.128 901 .188 .206 834 -.171 385
.139 .149 124 -.129 772 .189 .208 041 -.172214
.140 .150 274 -.130 642 .190 .209 250 -.173 041
.141 .151 425 -.131 511 .191 .210 459 -.173 867
.142 .152 577 -.132 379 .192 .211 671 -.174 693
.143 .153 730 -.133 246 .193 .212 883 -.175 518
.144 .154 884 -.134 112 .194 .214 096 -.176 342
.145 .156 040 -.134 978 .195 .215 311 -.177 165
.146 .157 196 -.135 842 .196 .216 527 -.177 988
.147 .158 354 -.136 706 .197 .217 744 -.178 809
.148 .159 513 -.137 569 .198 .218 962 -.179 630
.149 .160 673 -.138 431 .199 .220 182 -.180 450
.150 .161 834 -.139 292 .200 .221 403 -.181 269
TABLES 475
y ev — 1 e~y — 1 y ey — \ e-y-i
.201 .222 625 -.182 088 .251 .285 310 -.221 978
.202 .223 848 -.182 905 .252 .286 596 -.222 755
.203 .225 072 -.183 722 .253 .287 883 -.223 532
.204 .226 298 -.184 538 .254 .289 172 -.224 308
.205 .227 525 -.185 353 .255 .290 462 -.225 084
.206 .228 753 -.186 167 .256 .291 753 -.225 858
.207 .229 983 -.186 980 .257 .293 045 -.226 632
.208 .231 213 -.187 793 .258 .294 339 -.227 405
.209 .232 445 -.188 605 .259 .295 634 -.228 177
.210 .233 678 -.189 416 .260 .296 930 -.228 948
.211 .234 912 -.190 226 .261 .298 228 -.229 719
.212 .236 148 -.191 035 .262 .299 527 -.230 489
.213 .237 385 -.191 844 .263 .300 827 -.231 258
.214 .238 623 -.192 652 .264 .302 128 -.232 026
.215 .239 862 -.193 459 .265 .303 431 -.232 794
.216 .241 102 -.194 265 .266 .304 735 -.233 561
.217 .242 344 -.195 070 .267 .306 040 -.234 327
.218 .243 587 -.195 875 .268 .307 347 -.235 092
.219 .244 831 -.196 678 .269 .308 655 -.235 857
.220 .246 077 -.197 481 .270 .309 964 -.236 621
.221 .247 323 -.198 283 .271 .311 275 -.237 384
.222 .248 571 -.199 085 .272 .312 587 -.238 146
.223 .249 821 -.199 885 .273 .313 900 -.238 907
.224 .251 071 -.200 685 .274 .315 215 -.239 668
.225 .252 323 -.201 484 .275 .316 531 -.240 428
.226 .253 576 -.202 282 .276 .317 848 -.241 187
.227 .254 830 -.203 079 .277 .319 166 -.241 946
.228 .256 085 -.203 876 .278 .320 486 -.242 703
.229 .257 342 -.204 671 .279 .321 807 -.243 460
.230 .258 600 -.205 466 .280 .323 130 -.244 216
.231 .259 859 -.206 261 .281 .324 454 -.244 972
.232 .261 120 -.207 054 .282 .325 779 -.245 726
.233 .262 381 -.207 846 .283 .327 105 -.246 480
.234 .263 644 -.208 638 .284 .328 433 -.247 233
.235 .264 909 -.209 429 .285 .329 762 -.247 986
.236 .266 174 -.210219 .286 .331 092 -.248 737
.237 .267 441 -.211 009 .287 .332 424 -.249 488
.238 .268 709 -.211 797 .288 .333 757 -.250 238
.239 .269 979 -.212 585 .289 .335 092 -.250 988
.240 .271 249 -.213 372 .290 .336 427 -.251 736
.241 .272 521 -.214 158 .291 .337 765 -.252 484
.242 .273 794 -.214 944 .292 .339 103 -.253 231
.243 .275 069 -.215 728 .293 .340 443 -.253 978
.244 .276 344 -.216 512 .294 .341 784 -.254 724
.245 .277 621 -.217 295 .295 .343 126 -.255 468
.246 .278 900 -.218 078 .296 .344 470 -.256 213
.247 .280 179 -.218 859 .297 .345 815 -.256 956
.248 .281 460 -.219 640 .298 .347 162 -.257 699
.249 .282 742 -.220 420 .299 .348 510 -.258 441
.250 .284 025 -.221 199 .300 .349 859 -.259 182
476 TABLES
y ev — 1 e~y — 1 y ey— 1 e-y-i
.301 .351 209 -.259 922 .351 .420 487 -.296 016
.302 .352 561 -.260 662 .352 .421 909 -.296 720
.303 .353 914 -.261 401 .353 .423 331 -.297 423
.304 .355 269 -.262 139 .354 .424 755 -.298 125
.305 .356 625 -.262 877 .355 .426 181 -.298 827
.306 .357 982 -.263 613 .356 427 608 -.299 527
.307 .359 341 -.264 349 .357 .429 036 -.300 228
.308 .360 701 -.265 085 .358 .430 466 -.300 927
.309 .362 062 -.265 819 .359 .431 897 -.301 626
.310 .363 425 -.266 553 .360 .433 329 -.302 324
.311 .364 789 -.267 286 .361 .434 763 -.303 021
.312 .366 155 -.268 018 .362 .436 199 -.303 718
.313 .367 522 -.268 750 .363 .437 636 -.304 414
.314 .368 890 -.269 481 .364 .439 074 -.305 109
.315 .370 259 -.270 211 .365 .440 514 -.305 803
.316 .371 630 -.270 941 .366 .441 955 -.306 497
.317 .373 003 -.271 669 .367 .443 398 -.307 190
.318 .374 376 -.272 397 .368 .444 842 -.307 883
.319 .375 751 -.273 124 .369 .446 288 -.308 575
.320 .377 128 -.273 851 .370 .447 735 -.309 266
.321 .378 506 -.274 577 .371 .449 183 -.309 956
.322 .379 885 -.275 302 .372 .450 633 -.310 646
.323 .381 265 -.276 026 .373 .452 084 -.311 335
.324 .382 647 -.276 750 .374 .453 537 -.312 023
.325 .384 031 -.277 473 .375 .454 991 -.312711
.326 .385 415 -.278 195 .376 .456 447 -.313 398
.327 .386 801 -.278 916 .377 .457 904 -.314 084
.328 .388 189 -.279 637 .378 .459 363 -.314 770
.329 .389 578 -.280 357 .379 .460 823 -.315 454
.330 .390 968 -.281 076 .380 .462 285 -.316 139
.331 .392 360 -.281 795 .381 .463 748 -.316 822
.332 .393 753 -.282 513 .382 .465 212 -.317 505
.333 .395 147 -.283 230 .393 .466 678 -.318 187
.334 .396 543 -.283 946 .384 .468 145 -.318 869
.335 .397 940 -.284 662 .385 .469 614 -.319 549
.336 .399 339 -.285 377 .386 .471 085 -.320 229
.337 .400 739 -.286 091 .387 .472 556 -.320 909
.338 .402 141 -.286 805 .388 .474 030 -.321 588
.339 .403 543 -.287 518 .389 .475 505 -.322 266
.340 .404 948 -.288 230 .390 .476 981 -.322 943
.341 .406 353 -.288 941 .391 .478 459 - .323 620
.342 .407 760 -.289 652 .392 .479 938 -.324 296
.343 .409 169 -.290 362 .393 .481 418 -.324 971
.344 .410 579 -.291 071 .394 .482 901 -.325 646
.345 .411 990 -.291 780 .395 .484 384 -.326 320
.346 .413 403 -.292 488 .396 .485 869 -.326 993
.347 .414 817 -.293 195 .397 .487 356 -.327 666
.348 .416 232 -.293 901 .398 .488 844 -.328 338
.349 .417 649 -.294 607 .399 .490 334 -.329 009
.350 .419 068 -.295 312 .400 .491 825 -.329 680
TABLES All
y ev —\ e~y — \ y eu — l e-v -1
.401 .493 317 -.330 350 .451 .569 881 -.363 009
.402 .494 811 -.331 019 .452 .571 452 -.363 646
.403 .496 307 -.331 688 .453 .573 024 -.364 282
.404 .497 804 -.332 356 .454 .574 598 -.364 917
.405 .499 302 -.333 023 .455 .576 173 -.365 552
.406 .500 803 -.333 690 .456 .577 750 -.366 186
.407 .502 304 -.334 356 .457 .579 329 -.366 820
.408 .503 807 -.335 021 .458 .580 909 -.367 453
.409 .505 312 -.335 686 .459 .582 491 -.368 085
.410 .506 818 -.336 350 .460 .584 074 -.368 716
.411 .508 325 -.337 013 .461 .585 659 -.369 347
.412 .509 834 -.337 676 .462 .587 245 -.369 978
.413 .511 345 -.338 338 .463 .588 833 -.370 607
.414 .512 857 -.338 999 .464 .590 423 -.371 236
.415 .514 371 -.339 660 .465 .592 014 -.371 865
.416 .515 886 -.340 320 .466 .593 607 -.372 493
.417 .517 403 -.340 979 .467 .595 201 -.373 120
.418 .518 921 -.341 638 .468 .596 797 -.373 746
.419 .520 440 -.342 296 .469 .598 395 -.374 372
.420 .521 962 -.342 953 .470 .599 994 -.374 998
.421 .523 484 -.343 610 .471 .601 595 -.375 622
.422 .525 009 - 344 266 .472 .603 197 -.376 246
.423 .526 534 -.344 921 .473 .604 801 -.376 870
.424 .528 062 -.345 576 .474 .606 407 -.377 493
.425 .529 590 -.346 230 .475 .608 014 -.378 115
.426 .531 121 -.346 884 .476 .609 623 -.378 737
.427 .532 653 -.347 536 .477 .611 233 -.379 357
.428 .534 186 -.348 189 .478 .612 845 -.379 978
.429 .535 721 -.348 840 .479 .614 459 -.380 598
.430 .537 258 -.349 491 .480 .616 074 -.381 217
.431 .538 796 -.350 141 .481 .617 691 -.381 835
.432 .540 335 -.350 791 .482 .619 310 -.382 453
.433 .541 876 -.351 440 .483 .620 930 -.383 070
.434 .543 419 -.352 088 .484 .622 552 -.383 687
.435 .544 963 -.352 735 .485 .624 175 -.384 303
.436 .546 509 -.353 382 .486 .625 800 -.384 918
.437 .548 056 -.354 029 .487 .627 427 -.385 533
.438 .549 605 -.354 674 .488 .629 055 -.386 147
.439 .551 155 -.355 319 .489 .630 685 -.386 761
.440 .552 707 -.355 964 .490 .632 316 -.387 374
.441 .554 261 -.356 607 .491 .633 949 -.387 986
.442 .555 816 -.357 250 .492 .635 584 -.388 598
.443 .557 372 -.357 893 .493 .637 221 -.389 209
.444 .558 930 -.358 535 .494 .638 859 -.389 819
.445 .560 490 -.359 176 .495 .640 498 -.390 429
.446 .562 051 -.359 816 .496 .642 140 -.391 038
.447 .563 614 -.360 456 .497 .643 783 -.391 647
.448 .565 179 -.361 095 .498 .645 427 -.392 255
.449 .566 745 -.361 734 .499 .647 073 -.392 863
.450 .568 312 -.362 372 .500 .648 721 -.393 469
478 TABLES
y ev-\ e~v — 1 y ey — l e~y-i
.501 .650 371 -.394 076 .551 .734 987 -.423 627
.502 .652 022 -.394 681 .552 .736 723 -.424 203
.503 .653 675 -.395 286 .553 .738 461 -.424 778
.504 .655 329 -.395 891 .554 .740 200 -.425 353
.505 .656 986 -.396 494 .555 .741 941 -.425 928
.506 .658 643 -.397 098 .556 .743 684 -.426 502
.507 .660 303 -.397 700 .557 .745 428 -.427 075
.508 .661 964 -.398 302 .558 .747 175 -.427 647
.509 .663 627 -.398 904 .559 .748 923 -.428 219
.510 .665 291 -.399 504 .560 .750 672 -.428 791
.511 .666 957 -.400 105 .561 .752 424 -.429 362
.512 .668 625 -.400 704 .562 .754 177 -.429 932
.513 .670 295 -.401 303 .563 .755 932 -.430 502
.514 .671 966 -.401 902 .564 .757 689 -.431 071
.515 .673 638 -.402 499 .565 .759 448 -.431 640
.516 .675 313 -.403 097 .566 .761 208 -.432 208
.517 .676 989 -.403 693 .567 .762 970 -.432 775
.518 .678 667 -.404 289 .568 .764 734 -.433 342
.519 .680 346 -.404 885 .569 .766 500 -.433 909
.520 .682 028 -.405 479 .570 .768 267 -.434 475
.521 .683 711 -.406 074 .571 .770 036 -.435 040
.522 .685 395 -.406 667 .572 .771 807 -.435 604
.523 .687 081 -.407 260 .573 .773 580 -.436 169
.524 .688 769 -.407 853 .574 .775 354 -.436 732
525 .690 459 -.408 445 .575 .777 131 -.437 295
.526 .692 150 -.409 036 .576 .778 909 -.437 858
.527 .693 843 -.409 627 .577 .780 688 -.438 419
.528 .695 538 -.410217 .578 .782 470 -.438 981
.529 .697 234 -.410 806 .579 .784 253 -.439 541
.530 .698 932 -.411 395 .580 .786 038 -.440 102
.531 .700 632 -.411 983 .581 .787 825 -.440 661
.532 .702 334 -.412 571 .582 .789 614 -.441 220
.533 .704 037 -.413 158 .583 .791 405 -.441 779
.534 .705 742 -.413 745 .584 .793 197 -.442 337
.535 .707 448 -.414 331 .585 .794 991 -.442 894
.536 .709 157 -.414916 .586 .796 787 -.443 451
.537 .710 867 -.415 501 .587 .798 585 - .444 007
.538 .712 578 -.416 085 .588 .800 384 -.444 563
.539 .714 292 — .416 669 .589 .802 185 -.445 118
.540 .716 007 -.417 252 .590 .803 988 -.445 673
.541 .717 724 -.417 834 .591 .805 793 -.446 227
.542 .719 442 -.418 416 .592 .807 600 -.446 780
.543 .721 163 -.418 997 .593 .809 409 -.447 333
.544 .722 885 -.419 578 .594 .811 219 -.447 886
.545 .724 608 -.420 158 .595 .813 031 -.448 437
.546 .726 334 -.420 738 .596 .814 845 -.448 989
.547 .728 061 -.421 317 .597 .816 661 -.449 539
.548 .729 790 -.421 895 .598 .818 478 -.450 090
.549 .731 521 -.422 473 .599 .820 298 -.450 639
.550 .733 253 -.423 050 .600 .822 119 -.451 188
TABLES 479
y ev — 1 e~y — 1 y ey — 1 e-y — i
.601 .823 942 -.451 737 .651 .917 457 -.478 476
.602 .825 767 -.452 285 .652 .919 376 -.478 997
.603 .827 593 -.452 832 .653 .921 296 -.479 518
.604 .829 422 -.453 379 .654 .923 218 -.480 038
.605 .831 252 -.453 926 .655 .925 143 -.480 558
.606 .833 084 -.454 471 .656 .927 069 -.481 077
.607 .834 918 -.455 017 .657 .928 997 -.481 596
.608 .836 754 -.455 561 .658 .930 927 -.482 114
.609 .838 592 -.456 106 .659 .932 859 -.482 632
.610 .840 431 -.456 649 .660 .934 792 -.483 149
.611 .842 273 -.457 192 .661 .936 728 -.483 665
.612 .844 116 -.457 735 .662 .938 666 -.484 181
.613 .845 961 -.458 277 .663 .940 605 -.484 697
.614 .847 808 -.458 818 .664 .942 547 -.485 212
.615 .849 657 -.459 359 .665 .944 491 -.485 726
.616 .851 507 -.459 899 .666 .946 436 -.486 240
.617 .853 360 -.460 439 .667 .948 383 -.486 754
.618 .855 214 -.460 979 .668 .950 333 -.487 267
.619 .857 070 -.461 517 .669 .952 284 -.487 779
.620 .858 928 -.462 056 .670 .954 237 -.488 291
.621 .860 788 -.462 593 .671 .956 193 -.488 803
.622 .862 650 -.463 130 .672 .958 150 -.489 314
.623 .864 513 -.463 667 .673 .960 109 -.489 824
.624 .866 379 -.464 203 .674 .962 070 -.490 334
.625 .868 246 -.464 739 .675 .964 033 -.490 844
.626 .870 115 -.465 274 .676 .965 998 -.491 352
.627 .871 986 -.465 808 .677 .967 965 -.491 861
.628 .873 859 -.466 342 .678 .969 934 -.492 369
.629 .875 734 -.466 875 .679 .971 905 -.492 876
.630 .877 611 -.467 408 .680 .973 878 -.493 383
.631 .879 489 -.467 941 .681 .975 853 -.493 889
.632 .881 370 -.468 472 .682 .977 829 -.494 395
.633 .883 252 -.469 004 .683 .979 808 -.494 901
.634 .885 136 -.469 534 .684 .981 789 -.495 405
.635 .887 022 -.470 065 .685 .983 772 -.495 910
.636 .888 910 -.470 594 .686 .985 757 -.496 414
.637 .890 800 -.471 123 .687 .987 743 -.496 917
.638 .892 692 -.471 652 .688 .989 732 -.497 420
.639 .894 585 -.472 180 .689 .991 723 -.497 922
.640 .896 481 -.472 708 .690 .993 716 -.498 424
.641 .898 378 -.473 235 .691 .995 710 -.498 925
.642 .900 278 -.473 761 .692 .997 707 -.499 426
.643 .902 179 -.474 287 .693 .999 706 -.499 926
.644 .904 082 -.474 813 .694 1.001 71 -.500 426
.645 .905 987 -.475 337 .695 1.003 71 -.500 926
.646 .907 894 -.475 862 .696 1.005 71 -.501 424
.647 .909 803 -.476 386 .697 1.007 72 -.501 923
.648 .911 714 -.476 909 .698 1.009 73 -.502 421
.649 .913 626 -.477 432 .699 1.011 74 -.502 918
.650 .915 541 -.477 954 .700 1.013 75 -.503 415
480 TABLES
y 1 e~v — \ y e*' — 1 e-y-1
.701 1.015 77 . -.503 911 .751 1.119 12 -.528 106

.702 1.017 78 -.504 407 .752 1.121 24 -.528 577
.703 1.019 80 -.504 902 .753 1.123 36 -.529 048
.704 1.021 82 -.505 397 .754 1.125 48 -.529 519
.705 1.023 85 -.505 891 .755 1.127 61 -.529 989
.706 1.025 87 -.506 385 .756 1.129 74 -.530 459
.707 1.027 90 -.506 879 .757 1.131 87 -.530 928
.708 1.029 93 -.507 372 .758 1.134 00 -.531 397
.709 1.031 96 -.507 864 .759 1.136 14 -.531 866
.710 1.033 99 -.508 356 .760 1.138 28 -.532 334
.711 1.036 03 -.508 847 .761 1.140 42 -.532 801

.712 1.038 06 -.509 338 .762 1.142 56 -.533 268
.713 1.040 10 -.509 829 .763 1.144 70 -.533 734
.714 1.042 14 -.510318 .764 1.146 85 -.534 201
.715 1.044 19 -.510 808 .765 1.148 99 -.534 666
.716 1.046 23 -.511 297 .766 1.151 14 -.535 131
.717 1.048 28 -.511 785 .767 1.153 30 -.535 596
.718 1.050 33 -.512 273 .768 1.155 45 -.536 060
.719 1.052 38 -.512 761 .769 1.157 61 -.536 524
.720 1.054 43 -.513 248 .770 1.159 77 -.536 987
.721 1.056 49 -.513 734 .771 1.161 93 -.537 450

.722 1.058 55 -.514 220 .772 1.164 09 -.537 912
.723 1.060 61 -.514 706 .773 1.166 26 -.538 374
.724 1.062 67 -.515 191 .774 1.168 42 -.538 835
.725 1.064 73 -.515 675 .775 1.170 59 -.539 296
.726 1.066 80 -.516 160 .776 1.172 76 -.539 757
.727 1.068 86 -.516 643 .777 1.174 94 -.540 217
.728 1.070 93 -.517 126 .778 1.177 11 -.540 676
.729 1.073 01 -.517 609 .779 1.179 29 -.541 135
.730 1.075 08 -.518 091 .780 1.181 47 -.541 594
.731 1.077 16 -.518 573 .781 1.183 65 -.542 052

.732 1.079 23 -.519 054 .782 1.185 84 -.542 510
.733 1.081 32 -.519 535 .783 1.188 03 -.542 967
.734 1.083 40 -.520 015 .784 1.190 22 -.543 424
.735 1.085 48 -.520 495 .785 1.192 41 -.543 880
.736 1.087 57 -.520 974 .786 1.194 60 -.544 336
.737 1.089 66 -.521 453 .787 1.196 80 -.544 792
.738 1.091 75 -.521 931 .788 1.198 99 -.545 247
.739 1.093 84 -.522 409 .789 1.201 19 -.545 701
.740 1.095 94 -.522 886 .790 1.203 40 -.546 155
.741 1.098 03 -.523 363 .791 1.205 60 -.546 609

.742 1.100 13 -.523 839 .792 1.207 81 -.547 062
.743 1.102 23 -.524 315 .793 1.21002 -.547 515
.744 1.104 34 -.524 791 .794 1.212 23 -.547 967
.745 1.106 44 -.525 266 .795 1.21444 -.548 419
.746 1.108 55 -.525 740 .796 1.216 66 -.548 870
.747 1.11066 -.526 214 .797 1.218 87 -.549 321
.748 1.11277 -.526 688 .798 1.221 09 -.549 771
.749 1.11488 -.527 161 .799 1.223 32 -.550 221
.750 1.11700 -.527 633 .800 1.225 54 -.550 671
TABLES 481
y ev -\ e-y-1 y ey -\ e~y — 1
.801 1.227 77 -.551 120 .851 1.341 99 -.573 012

.802 1.230 00 -.551 569 .852 1.344 33 -.573 439
.803 1.232 23 -.552 017 .853 1.346 68 -.573 865
.804 1.234 46 -.552 465 .854 1.349 02 -.574 291
.805 1.236 70 -.552 912 .855 1.351 37 -.574 717
.806 1.238 93 -.553 359 .856 1.353 73 -.575 142
.807 1.241 17 -.553 805 .857 1.356 08 -.575 567
.808 1.243 42 -.554 251 .858 1.358 44 -.575 991
.809 1.245 66 -.554 697 .859 1.360 80 -.576 415
.810 1.247 91 -.555 142 .860 1.363 16 -.576 838
.811 1.250 16 -.555 587 .861 1.365 53 -.577 261

.812 1.252 41 -.556 031 .862 1.367 89 -.577 683
.813 1.254 66 -.556 475 .863 1.370 26 -.578 106
.814 1.256 92 -.556 918 .864 1.372 63 -.578 527
.815 1.259 18 -.557 361 .865 1.375 01 -.578 948
.816 1.261 44 -.557 803 .866 1.377 38 -.579 369
.817 1.263 70 -.558 245 .867 1.379 76 -.579 790
.818 1.265 96 -.558 687 .868 1.382 14 -.580 210
.819 1.268 23 -.559 128 .869 1.384 53 -.580 629
.820 1.270 50 -.559 568 .870 1.386 91 -.581 048
.821 1.272 77 -.560 009 .871 1.389 30 -.581 467

.822 1.275 05 -.560 448 .872 1.391 69 -.581 886
.823 1.277 32 -.560 888 .873 1.394 08 -.582 303
.824 1.279 60 -.561 327 .874 1.396 48 -.582 721
.825 1.281 88 -.561 765 .875 1.398 88 -.583 138
.826 1.284 16 -.562 203 .876 1.401 28 -.583 555
.827 1.286 45 -.562 641 .877 1.403 68 -.583 971
.828 1.288 74 -.563 078 .878 1.406 08 -.584 387
.829 1.291 03 -.563 514 .879 1.408 49 -.584 802
.830 1.293 32 -.563 951 .880 1.410 90 -.585 217
.831 1.295 61 -.564 387 .881 1.413 31 -.585 632

.832 1.297 91 -.564 822 .882 1.415 73 -.586 046
.833 1.300 21 -.565 257 .883 1.418 14 -.586 460
.834 1.302 51 -.565 691 .884 1.420 56 -.586 873
.835 1.304 81 -.566 126 .885 1.422 98 -.587 286
.836 1.307 12 -.566 559 .886 1.425 41 -.587 698
.837 1.309 43 -.566 992 .887 1.427 84 -.588 110
.838 1.311 74 -.567 425 .888 1.430 26 -.588 522
.839 1.31405 -.567 858 .889 1.432 70 -.588 933
.840 1.316 37 -.568 289 .890 1.435 13 -.589 344
.841 1.318 68 -.568 721 .891 1.437 57 -.589 755

.842 1.321 00 -.569 152 .892 1.440 00 -.590 165
.843 1.323 33 -.569 583 .893 1.442 45 -.590 574
.844 1.325 65 -.570 013 .894 1.444 89 -.590 984
.845 1.327 98 -.570 443 .895 1.447 34 -.591 392
.846 1.330 31 -.570 872 .896 1.449 78 -.591 801
.847 1.332 64 -.571 301 .897 1.452 24 -.592 209
.848 1.334 97 -.571 729 .898 1.454 69 -.592 616
.849 1.337 31 -.572 157 .899 1.457 14 -.593 024
.850 1.339 65 -.572 585 .900 1.459 60 -.593 430
482 TABLES
TABLE D2 (concluded)
y ev — l e~y — 1 y ey-l e~y — 1
.901 1.462 06 -.593 837 .951 1.588 30 -.613 646

.902 1.464 53 -.594 243 .952 1.590 89 -.614 032
.903 1.466 99 -.594 648 .953 1.593 48 -.614417
.904 1.469 46 -.595 053 .954 1.596 07 -.614 803
.905 1.471 93 -.595 458 .955 1.598 67 -.615 188
.906 1.474 41 -.595 862 .956 1.601 27 -.615 572
.907 1.476 88 -.596 266 .957 1.603 87 -.615 957
.908 1.479 36 -.596 670 .958 1.606 48 -.616 341
.909 1.481 84 -.597 073 .959 1.609 09 -.616 724
.910 1.484 32 -.597 476 .960 1.611 70 -.617 107
.911 1.486 81 -.597 878 .961 1.614 31 -.617 490

.912 1.489 30 -.598 280 .962 1.61693 -.617 872
.913 1.491 79 -.598 682 .963 1.619 54 -.618 254
.914 1.494 28 -.599 083 .964 1.622 16 -.618 636
.915 1.496 78 -.599 483 .965 1.624 79 -.619 017
.916 1.499 27 -.599 884 .966 1.627 41 -.619 398
.917 1.501 77 -.600 284 .967 1.630 04 -.619 778
.918 1.504 28 -.600 683 .968 1.632 67 -.620 158
.919 1.506 78 -.601 082 .969 1.635 31 -.620 538
.920 1.509 29 -.601 481 .970 1.637 94 -.620 917
.921 1.511 80 -.601 879 .971 1.640 58 -.621 296

.922 1.51431 -.602 277 .972 1.643 23 -.621 674
.923 1.516 83 -.602 675 .973 1.645 87 -.622 053
.924 1.519 35 -.603 072 .974 1.648 52 -.622 430
.925 1.521 87 -.603 469 .975 1.651 17 -.622 808
.926 1.524 39 -.603 865 .976 1.653 82 -.623 185
.927 1.526 92 -.604 261 .977 1.656 47 -.623 561
.928 1.529 45 -.604 656 .978 1.659 13 -.623 938
.929 1.531 98 -.605 052 .979 1.661 79 -.624 313
.930 1.534 51 -.605 446 .980 1.664 46 -.624 689
.931 1.537 04 -.605 841 .981 1.667 12 -.625 064

.932 1.539 58 -.606 235 .982 1.669 79 -.625 439
.933 1.542 12 -.606 628 .983 1.672 46 -.625 813
.934 1.544 67 -.607 021 .984 1.675 14 -.626 187
.935 1.547 21 -.607 414 .985 1.677 81 -.626 561
.936 1.549 76 -.607 807 .986 1.680 49 -.626 934
.937 1.552 31 -.608 199 .987 1.683 17 -.627 307
.938 1.554 87 -.608 590 .988 1.685 86 -.627 679
.939 1.557 42 -.608 981 .989 1.688 54 -.628 052
.940 1.559 98 -.609 372 .990 1.691 23 -.628 423
.941 1.562 54 -.609 763 .991 1.693 93 -.628 795

.942 1.565 11 -.610 153 .992 1.696 62 -.629 166
.943 1.567 67 -.610 542 .993 1.699 32 -.629 536
.944 1.570 24 -.610 932 .994 1.702 02 -.629 907
.945 1.572 81 -.611 320 .995 1.704 72 -.630 277
.946 1.575 39 -.611 709 .996 1.707 43 — .630 646
.947 1.577 96 -.612 097 .997 1.710 14 -.631 015
.948 1.580 54 -.612 485 .998 1.712 85 -.631 384
.949 1.583 13 -.612 872 .999 1.715 56 -.631 753
.950 1.585 71 -.613 259 1.000 1.718 28 -.632 121
INDEX
Additive preferences, see Preference C.E.S. production function 312-313

independence and block-independence Chenery, H. B. 312m
Aggregation theory of information Chi-square as an approximation to the
concepts, see Information concepts, information inaccuracy 38-39
aggregation of Climatological chance 8-9, 12
Agriculture, forestry and fishing 39-42 Clothing industry (German and Danish),
in input-output aggregation 349 see Textile and clothing industry
Allen, R. G. D. 308m Coal miner families, household data on
Allocation discrepancy 143-146, 150, 154, 138-150, 158, 169-172, 268-275
165-166, 172, 180-181 Coleman, E. J. 104m
in relation to demand equations Common Market, description of the
202-203, 227 development of the - in informational
in relation to the true index of real terms 377-387
income 226 Concentration (industrial)
Anderson, O. 14, 52 definition in informational terms 290-293
Appraisal of inventories, see Production application to passenger cars in the U.S.
plan revisions 291-302
Arrow, K. J. 312m aggregation over makes 291-295
Attneave, F. 20 aggregation over regions 295-302
See also Herfindahl’s concentration index
Barten, A. P. 152, 172, 190m, 198m, 233m, Concentration (international trade)
238m, 239m, 314 for individual countries and commodity
Bauer, R. K. 14, 52n groups 367-372
Berkson, J. 72m aggregative measures of 372-377
Bit 5 Conditional distributions 49-51
Bivariate information theory, see See also Entropy, average conditional -
Information theory, bivariate Constant returns to scale, see Returns,
Block-independent preferences, see constant - to scale
Preference independence and Construction Programme of the
block-independence Netherlands 36-38
Boas, J. 76/z Coombs, C. H. 220n
Brillouin, L. 11 9m Correlation coefficients compared with
Brown, A. 363, 388m average conditional entropies 58-59
Budget constraint 182 Cost of living price index, see Price index
number, cost of living
Cell effect (input-output aggregation) 340 Covariance of logarithmic price and
Central Bureau of Statistics (The Hague), quantity differences, see Variance and
see Statistics, Central Bureau of covariance of logarithmic price and
Central Planning Bureau (The Hague), see quantity differences
Planning, Central-Bureau Cramer, J. S. 73m
483
484 INDEX
Davis, R. L. 220n consumer demand theory 189-190

De Jongh, B. H., see Jongh, B. H. de factor demand theory 305-306;
De Leeuw, C. G., see Feeuw, C. G. de solution of 310
Debreu, G. 220/7
Decomposition forecasts 35-46 Garner, W. R. 165/7
“Deflated” per capita income 102, 110, 121 Gini’s concentration ratio 96, 121-123
Demand equations (consumer goods) 182, Goldman, S. 391«, 392n
185-186 Graham Jr., R. E. 104/?, 105/7
in infinitesimal changes 197
in finite changes 200 Hartley, R. V. F. 19, 20
error terms in 228-233 Herfindahl, O. C. 316/7
aggregation and disaggregation 233-237 Herfindahl’s concentration index 128, 293/7,
See also Preference independence and 316-318
block-independence Hessian matrix of the production function
Demand equations (production factors) 305
303-304 assumed to be negative semi-definite
Deming, W. E. 360/7 306-307
Determinative message, expected Hessian matrix of the utility function 189
information of, see Information, assumed to be negative definite 194
expected - in a determinative message as an indicator of prices to be introduced
Divisia, F. 138/7 into the demand equations 221
Dosser, D. 107 diagonalization of 276-280
See also preference independence and
Ellis Island, see Maxwell’s demon block-independence
Emanuel, H. 69n Hicks, J. R. 185/7
Entropy 19, 26, 34-35, 47-48 Hirschman, A. O. 316n
average conditional - 50-51, 57-59 Hooper, J. W. 104/7
as a measure of (income) equality 91-95, Houthakker, H. S. 187/7, 192/7, 196/7, 198/7
127
as an inverse concentration measure Ifo-Institut filr Wirtschaftsforschung 11
290-293, 372-374 Inaccuracy, information -, see Information
of a continuous distribution 390-392 inaccuracy
Equivalent income changes 263-268 Income changes, equivalent -, see
Equivalent income changes
Factor reversal test 143-146, 154 Income effect of price changes 191-192
Finney, D. J. 72/7 Income elasticity of consumption good
Fisher, I. 138/7 205-206
Flexibility, see Income flexibility and variance of 260-263
Substitution flexibility Income flexibility 195-196, 205-207
Food industry (input-output aggregation) statistical testing of 256-258
345-350 Income inequality, see Inequality (of
Forecasts, see (1) Decomposition forecasts, income)
(2) Planning, Central - Bureau, (3) Income, marginal utility of, see Marginal
Survey forecasts, (4) Weather forecasts utility of income
Frisch, R. 135/7, 196/7, 198/7, 213/7 Indirect message, expected information of
Fuhrer, H. 52/7 an -, see Information, expected - of an
Fundamental matrix equation indirect message
INDEX 485
Indirect utility function 208-209 Information (difference) component of two

income solution of 209-210 sets of value shares, see Allocation
Inequality (of income) 91-134 discrepancy
definition in informational terms 91 Information, expected -
alternative measures 121-128 of a determinative message 24-26, 47
aggregation and disaggregation 93-95 of an indirect message 27-29, 47-48
lognormal distribution 97, 127 Information gain
Pareto distribution 97-98, 99/7, 100-101, of forecasts (Central Planning Bureau)
127 and estimates (Central Bureau of
White and Nonwhite families in the U.S. Statistics) 406-409, 417-420
98-101, 104-105, 128-131, 134 of survey forecasts 12-18, 30-34, 43, 47
states of the U.S. 102-106, 112-114 of weather forecasts 9-10, 47
countries of the world 107-110, 112 Information improvement of forecast
when population shares remain constant revisions 43, 48
110-114 Information inaccuracy 37-46, 48
effect of migration 114-120 applied to demand predictions 241-251
limits to the inequality level 99, 128, Information, mutual - 34-35, 50-51
132-134 applied to the development of
Inequality coefficient (weighted) 411-412, international trade 363-365, 378-387
420 See also two-stage information forecasts
Inferior good 186-187, 193-194, 197, 199, Information received with a message 10, 47
200 Information theory, bivariate 33-35,49-51;
Information concepts, aggregation of 43-46 multivariate 55-59
See also (1) Cell effect (input-output Input heterogeneity (input-output
aggregation), (2) Concentration aggregation) 338-339
(industrial), aggregation over makes and reduction in the case of disaggregation
over regions, (3) Concentration 347-348
(international trade), aggregative in relation to input-output prediction
measures of, (4) Inequality (of income), 350-354
aggregation and disaggregation, (5) Input-output prediction
Information decomposition equation of standard procedure 320-322
input-output aggregation, (6) Input aggregation bias of intermediate
heterogeneity (input-output demand predictions 322-328
aggregation), (7) Output heterogeneity aggregation bias of primary demand
(input-output aggregation), (8) Value predictions 328-331
share, decomposition of the dispersion of homogeneous input structure 327-330,
- changes 334-335, 338-339
Information content See also (1) Information content of an
of a definite message 3-5, 46 input-output table, (2) Information
of an input-output table 333 decomposition equation of
of a message on stochastically input-output aggregation, (3) Input
independent events 4, 6-7 heterogeneity (input-output
of survey forecasts 29-33, 35, 48 aggregation)
See also Information received with a Input structure prediction 39-46
message Interaction of determining factors 61-65,
Information decomposition equation of 69-70, 399
input-output aggregation 337-340 Intermediate price vector, see Real income,
486 INDEX
true index of, evaluated at the as a function of income and prices

intermediate price vector 206- 208, 251-253
Intermediate utility level, see Price index as the reciprocal of a price deflator
number, true cost of living -, evaluated 207- 208
at the intermediate utility level See also Income flexibility
Inventories 13, 31 Marginal utility shock model, see Demand
Inventory appraisal, see Production plan equations (consumer goods), error terms
revisions in
Investment in machinery 13, 31 Marginal value share 195-197, 205-206
Markowitz, H. 221k
Jongh, B. H. de 19, 20
Matrix equation, fundamental ~, see
Fundamental matrix equation
Kendall, M. G. 75/z
Maxwell’s demon 119-120
Khinchin, A. I. 6n
McFadden, D. 203k
Kimball, G. E. 6n
McGill, W. J. 165k
Klaassen, L. H. 69n
Michaeli, M. 368k
Kloek, T. 138k, 146k, 172k, 208k, 217k,
Migration, see Inequality (of income),
223k, 268k
effect of migration
Klijn, N. 8k
Minhas, B. S. 312k
Koopman, B. O. 6k
Mnookin, R. H. 241k
Konjunkturinstitutet 11
Munksgaard, H. 14
Kullback, S. 20
Mutual information, see Information,
Laspeyres bias 146 mutual -
Leeuw, C. G. de 357k, 389
Lemma on local quadratic approximation, Necessity, see Luxury good
see Quadratic, lemma on local - Neudecker, H. 228n
approximation Nit 5
Leser, C. E. V. 198k
Lisman, J. H. C. 20 Odds 19, 62-63
Logit 63-90 One-dimensional price and quantity scales
Logit relations in qualitative determining 146-150, 270, 272-275
factors 65-71, 88-90 Orders received 13, 31
Logit regressions 71-87 surprises on -, see Production plan
Lognormal distribution, see Inequality (of revisions
income), lognormal distribution Output heterogeneity (input-output
Lorenz curve 121-123 aggregation), 339-340
Luxury good 205-206
Luxury-necessity index 258-263 Pareto distribution, see Inequality (of
income), Pareto distribution
MacDonald, D. K. C. 5 Peacock, A. T. 107
Malinvaud, E. 322n Pearce, I. F. 198k
Marginal cost 304 Petersen, J. P. 52k
Marginal price index, see Price index Pikler,A. 20
number, marginal - Planning, Central - Bureau (The Hague)
Marginal utility of income 184, 187, 209 392-395, 407, 421
in relation to the general substitution Posterior probabilities, see Prior and
effect of price changes 192-193 posterior probabilities
INDEX 487
Praag, B. M. S. van 221 n intermediate price vector

Prediction-realization table (surveys) 11-18 Principal components of consumer
Preference independence and preferences 276-280
block-independence 198-200 Prior and posterior probabilities 27
in relation to demand disturbances Probit analysis 72-73
231-233 Production 13-18, 21-23, 31
in relation to the aggregation of demand Production function 303
equations 234-237 See also C.E.S. production function
See also Principal components of Production plan revisions 52-56, 59-63,
consumer preferences 67-68
Price index number, cost of living -137-138 Profit 13, 31
for coal miner families (cross-section) Purchase prices 12-13, 30-31
141-143, 268-275
Dutch consumption (time series) 150-154 Quadratic, lemma on local -
as income deflator 200-201 approximation 201-202, 219, 222-223,

225-226, 269-270
as a determinant of the marginal utility
Quantity index number 138
of income 207
See also Price index number, true cost of for coal miner families (cross-section)
living - 141-143
Dutch consumption (time series) 150-154
Price index numbers in factor demand
theory 313« relation with left-hand variable of
demand equation 201-202
Price index number, marginal - 201,
See also Real income
251-253
Quantity index number in factor demand
as a deflator of prices in demand
theory 313
equations 201
Quantity index number, partial - 158-162
as a determinant of the marginal utility
of income 207 Rajaoja, V. 208n
See also Price index number, true RAS method 363, 388-389
marginal - Rate of profit, see Profit
Price index number, partial cost of living - Raw materials bought 13, 31
158-162; partial marginal 234-236 Real income 200
Price index number, true cost of living - true index of 223-226, 272«; price
210-212 elasticity of 224-225; evaluated at
price elasticity of 215 intermediate price vector 225-226
dependence on utility level 216-217 Redundancy 92/7
evaluated at intermediate utility level Returns, constant - to scale 306
217-219 Revisions of forecasts, see Information
application to household data of coal improvement of forecast revisions
miner families 268-275 Revisions of production plans, see
Price index number, true marginal - 213-214 Production plan revisions
price elasticity of 215-216 Rosenbluth, G. 316/7
application to household data of coal Rosenstein-Rodan, P. N. 107
miner families 274—275 Roy, R. 209
Prices (in surveys), see Purchase prices and
Sales prices Sales, 13, 31
Price vector, intermediate -, see Real Sales prices 13, 17-18, 21-23, 31
income, true index of, evaluated at Sampling variability of the information
488 INDEX
value 18-19 Uncertainty and information (duality) 25

Samuelson, P. A. 302 Uribe, P. 290/?, 331/?, 357/?, 389
Scholes, M. 290//, 398« Utility as a probability 221, 232/7
Schwartz, C. F. 104/?, 105/7 Utility function 183
Schwartz, L. S. 92n monotonic transformation of 220-221
Shannon, C. E. 19, 20 See also Indirect utility function and
Ship-airplane preference 69-71 Preference independence and
Skolka, J. 331/7, 333/7 block-independence
Solow, R. M. 312/7 Utility, intermediate - level, see Price
SOMERMEIJER, W. H. 198/7 index number, true cost of living -,
Spinning industry (German) 17-18, 22 evaluated at the intermediate utility level
Statistics, Central Bureau of - (The Hague) Uzawa, H. 312/7
11, 342/;, 392, 395, 407, 421
Stone, R. 198/7, 363, 388n Value share 136
Strotz, R. H. 198/7 decomposition of the dispersion of -
Stuart, A. 75n changes 163-166, 180-181
Substitution effect of price changes 155-158 quantity component of - change 182-183,
general substitution effect 192-194 194, 202, 204-206
specific substitution effect 194 See also Marginal value share
in relation to variances and covariances Van Praag, B. M. S., see Praag, B. M. S.
of demand disturbances 230-233 VAN
Substitution, elasticity of 308 Variance analysis 61/7
See also C.E.S. production function comparison with information analysis
Substitution flexibility 309 163-166, 180-181
Sunshine predictions 8-10 Variance and covariance of logarithmic
Surprises on orders received, see price and quantity differences 154—158,
Production plan revisions 253-258, 265-268; decomposition of
Survey forecasts 10-18, 29-33, 35 162-163
Variance decomposition
Textile and clothing industry (German and of income logarithms 124-125
Danish) 14-18, 21-23 of prediction and estimation errors
Theil, H. 8/7, 69/7, 104/7, 138/7, 144/7, 172/7, 398-399; estimation of the
190/7, 198/7, 228/7, 233/7, 241/?, 290n, 302/7, decomposition model 399-404
322/7, 331/?, 357«, 389, 398n Variation, coefficient of - as a measure of
Thrall, R. M. 220/? income inequality 125
Tilanus, C. B. 302n, 341 Volume index number, see Quantity index
Tobin, J. 73/? number and Real income
Tobit analysis 73/?
Tong Hun Lee 73/7 Waelbroeck, J. 360
Tornqvist, L. 138/7 Weather forecasts 8-10
Trade flow prediction 357-365 Weaver, W. 20
Turnovsky, S. J. 233/7, 239n Weaving industry (German) 17-18, 22
Two-stage information forecast 358-360 Wold, H. 184/?
application to the trade flows of eight Work force 13, 31
regions 360-365
Two-stage random selection procedure Zellner, A. 73/?
138, 159 ZWEZERIJNEN, W. J. 8/7
\
Date Due
HB 74 .M3T46
Theil, Henri. 010101 000
Economics and information theo
163 0226 42 5
TRENT UNIVERSITY
HB74 . M3 Tk6
Theil, Henri
Economics and information
theory
DATE ISSUED TO
-88690
88690
ik x * : ;>p;nx30' ;x*x:xu
• ‘ • 11 ><* > «/* • fp«< SU?I x

rii'.MKiif&hiM IKK v KJ tK»<J*N WT<V
ZSWMxrSIir. [ftpppp ii'.I ’■ r
5Jnj?/,?52K;K:>;sir
;j!K;3irN: ?ir?5f ii iijii-l: * S8jR»y«g8K8^{!

S&kttSeSfiaSKMi
v. SIS £iiiHb9nSfe«
XlOfK
^3i0fJ8oi3xS«s ji'SnS’cfS
!§g!3tei
*» * «* «**>- ><in »¥#IKJ
x vf?f 2Sh. :< nij xjdii *T;
3»Bf
O^kSH; SS> 3€U>f5
.<xw

Economics and Information Theory

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Economics and Information Theory

Uploaded by

Copyright:

Available Formats

ECONOMICS

Studies in Mathematical and Managerial Economics

NORTH-HOLLAND PUBLISHING COMPANY - AMSTERDAM

RAND McNALLY & COMPANY - CHICAGO

NORTH-HOLLAND PUBLISHING COMPANY - AMSTERDAM

RAND McNALLY & COMPANY - CHICAGO

No part of this book may be reproduced

Sole distributors for U.S.A. and Canada:

PRINTED IN THE NETHERLANDS

This is a series of books concerned with the quantitative approach to

The use of probability theory in modern economics is at present substantial.

Summer 1966 Henri Theil

Introduction to the Series. v

Outline of the Book.xix

1. The Information Concept. 1

8. Industrial Concentration and the Allocation Problem of the Firm . . 290

Introduction to the Series. v

Abbreviated Contents. VIII

Outline of the Book.xix

Chapter 1. The Information Concept

1. The Dog and the Chess-board. 1

Chapter 2. Expected Information

1. The Expected Information in a Determinative Message; The Entropy

Chapter 3. Economic Relations Involving Conditional Probabilities

1. Conditional Distributions and Their Entropies.49

Chapter 4. The Measurement of Income Inequality

1. Inequality at the Individual Level; The Aggregation Problem ... 91

Chapter 5. A Statistical Approach to the Problem of Price and

1. Random Selection of Commodities.135

6. Partial Index Numbers and the Decomposition of Price and Quantity

8. Concluding Remarks. 166

Chapter 6. The Consumer’s Allocation Problem

1. The Consumer’s Optimality Conditions and the General Form of

Chapter 7. Empirical Implications of the Allocation Model of the

1. A Reformulation of the Demand Equations; A Marginal Utility Shock

8. Equivalent Income Changes.263

Chapter 8. Industrial Concentration and the Allocation Problem of

1. The Measurement of Industrial Concentration.290

Chapter 9. Input-Output Analysis and Its Aggregation Problems

1. The Input-Output Technique: A Brief Summary.319

Chapter 10. Information Measures in the Analysis of International

2. An Application to the Trade of Eight Regions.360

Chapter 11. Continuous Information Theory and a Multiplicative

Decomposition of Prediction Error Variances

1. The Entropy of a Continuous Distribution; The Case of the Normal

A. Logarithms to the Base 2 and Natural Logarithms of Integers from 1

2.1. The Information Content of Survey Forecasts of Ten Variables . 31

3.1. Production Plan Revisions Determined by Surprises on Orders

4.1. Income Distributions of White Families and of Nonwhite Families in

5.1. Sub-regions and Their Weights within Their Region.139

5.13. Logarithmic Variances of Value Share Changes and Their Information

7.1. Dependent Variables and Disturbances of the Demand Equations for

8.1. The Entropy of Passenger Car Production in the United States, by

9.1. The Information Decomposition of Twelve Input-Output Tables . 343

11.1. Seven Successive Forecasts and Estimates of the Log-changes in Five

A. Logarithms to the Base 2 and Natural Logarithms of Integers from I

1.1. The Information Content h{ ). 5

2.1. The Expected Information H() for n=2.26

3.1. The Logit as a Function of the Probability.64

11.1. Seven Successive Forecasts and Estimates.393

1. The Information Concept

4. The Measurement of Income Inequality

The family household is in the center of the discussions of these chapters.

Chapter 6 is based on classical economic theory. It defines the cost of living

. “Information Theory and Its Application to Taxonomy," f«/ +P *>,

h(x1x2) = h(- logx1x2) = h(- logxx - logx2) = h*(y1 + y2)

(3-7) h(yt + y2) = fc(yl} + h*(y2) yuy2 > 0

(3-8) h(yi)<h(y2) if 0^yi<y2

We proceed to substitute yx=y2 = 0 in (3.7), which gives A(0) = 2/z(0)