Market Interrelationships and Applied Demand Analysis: Bridging The Gap Between Theory and Empirics in Commodities Markets Michael K. Wohlgenant

Market Interrelationships and Applied
Demand Analysis: Bridging the Gap

Between Theory and Empirics in
Commodities Markets Michael K.
Wohlgenant
Visit to download the full and correct content document:
https://ebookmass.com/product/market-interrelationships-and-applied-demand-analys
is-bridging-the-gap-between-theory-and-empirics-in-commodities-markets-michael-k-
wohlgenant/
MARKET
INTERRELATIONSHIPS
AND APPLIED
DEMAND ANALYSIS
Bridging the Gap Between Theory and
Empirics in Commodities Markets
Michael K. Wohlgenant
Palgrave Studies in Agricultural Economics and Food
Policy
Palgrave Textbooks in Agricultural Economics

and Food Policy
Series Editor
Christopher Barrett, Cornell University, Ithaca, NY, USA
This book series provides instructors and students with cutting-edge textbooks in
agricultural economics and food policy.
More information about this subseries at http://www.palgrave.com/gp/series/16444

Market Interrelationships
and Applied Demand
Analysis
Bridging the Gap Between Theory
and Empirics in Commodities Markets
North Carolina State University
Raleigh, NC, USA
ISSN 2662-3889 ISSN 2662-3897 (electronic)

Palgrave Studies in Agricultural Economics and Food Policy
ISSN 2662-5474 ISSN 2662-5482 (electronic)
Palgrave Textbooks in Agricultural Economics and Food Policy
ISBN 978-3-030-73143-4 ISBN 978-3-030-73144-1 (eBook)
https://doi.org/10.1007/978-3-030-73144-1
© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature
Switzerland AG 2021
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.
Cover credit: © SimonStock/Alamy Stock Photo
This Palgrave Macmillan imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
To Peggy
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Organization of Chapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 Consumer Demand—Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Consumer’s Choice Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2.1 First-Order Conditions: Application
of the Kuhn-Tucker Theorem . . . . . . . . . . . . . . . . . . . . . . 6
2.2.2 Demand Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Comparative Statics of Income and Price Changes . . . . . . . . . . . . 7
2.3.1 Fundamental Matrix Equation of Consumer
Demand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.3.2 Slutsky Substitution Matrix . . . . . . . . . . . . . . . . . . . . . . . 9
2.3.3 General Restrictions of Consumer Behavior . . . . . . . . . 10
2.4 Comparative Statics of Exogenous Preference Shift
Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.5 Comparative Statics with Strictly Concave Utility Function . . . . . 12
2.6 Duality Theory: Relationships Between Marshallian,
Hicksian, and Frischian Demands . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.7 Inverse Demand Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.7.1 Wold-Hotelling Identity . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.7.2 Comparative Statics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3 Consumer Demand—Separability and Commodity
Aggregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.2 Restrictions on Price Movements . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.3 Separability Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.4 Implications of Weak Separability . . . . . . . . . . . . . . . . . . . . . . . . . . 23
vii
viii Contents
3.4.1 Theorem 1: Conditional Demand Functions . . . . . . . . . 23

3.4.2 Total Differential of First-Stage Allocation
Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.5 Two-Stage Budgeting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.5.1 Theorem 2 and Consistency Requirement . . . . . . . . . . . 26
3.5.2 Form of Price Indices . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.6 Implications of Strong Separability . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.6.1 General Form of Composite Demand Function . . . . . . . 29
3.6.2 Unconditional Demand Elasticities . . . . . . . . . . . . . . . . . 30
3.7 Weak Separability Alone Insufficient to Estimate
Unconditional Demand Elasticities . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.8 Empirical Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4 Consumer Demand—Empirical Analysis I . . . . . . . . . . . . . . . . . . . . . . . 37
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.2 Brief History of Empirical Consumer Demand Models . . . . . . . . . 37
4.2.1 Single-Equation Approach . . . . . . . . . . . . . . . . . . . . . . . . 37
4.2.2 Systems of Demand Functions Derived
from Directly Specified Utility Function . . . . . . . . . . . . 38
4.2.3 Formulation of Directly Specified Demand
Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.2.4 Derivation of Demand Functions Using Locally
Flexible Functional Forms . . . . . . . . . . . . . . . . . . . . . . . . 39
4.2.5 Demand Functions Derived from Globally
Flexible Functional Forms . . . . . . . . . . . . . . . . . . . . . . . . 40
4.3 System of Demand Equations Estimation . . . . . . . . . . . . . . . . . . . . 40
4.4 The Rotterdam Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.5 The Almost Ideal Demand System . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.6 The Linear Approximate AIDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.7 Comparison of RM with AIDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.8 Alternative Differential Demand Systems . . . . . . . . . . . . . . . . . . . . 49
4.9 Generalization of AIDS Demand Systems . . . . . . . . . . . . . . . . . . . 49
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
5 Consumer Demand—Empirical Analysis II . . . . . . . . . . . . . . . . . . . . . . 55
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.2 Application to U.S. Meat Demand . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.2.1 Development of Data Set Based on USDA
Disappearance Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.2.2 Application to Absolute Price Version
of Rotterdam Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.2.3 Application to LA/AIDS and AIDS Models . . . . . . . . . 60
Contents ix
5.2.4 Unconditional Demand Elasticities

from Rotterdam Model . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.3 Imposing Negative Semi-Definiteness . . . . . . . . . . . . . . . . . . . . . . . 72
5.4 Other Data Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
6 Quality, Heterogeneous Goods, and Cross-Section Demand . . . . . . . 79
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.2 Unit Values, Prices, and Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.2.1 Problem of Average Prices Varying Across
Households . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.2.2 Income and Price Elasticities of Quality . . . . . . . . . . . . . 80
6.3 Cox and Wohlgenant Approach to Modeling Unit Values . . . . . . . 81
6.4 Deaton Approach to Unit Values in Estimating Demand
Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
6.5 Theil-Clement Approach to Quality Measurement . . . . . . . . . . . . . 86
6.6 Modeling Systems of Demand Functions for Heterogeneous
Goods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
6.6.1 Multi-Stage Budgeting Approaches to Quality . . . . . . . 87
6.6.2 Nested Logit Framework Approach to Modeling
Heterogeneous Goods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
6.6.3 Hedonic Metric Approach to Estimating
Differentiated Demand Functions . . . . . . . . . . . . . . . . . . 90
6.7 Specifying and Estimating Systems of Demand Functions
with Panel Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
6.7.1 Structure of Panel Data Models . . . . . . . . . . . . . . . . . . . . 91
6.7.2 Demographic Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
6.7.3 Modeling Higher Order Engel Curves . . . . . . . . . . . . . . 93
6.8 Systems of Demand Functions with Missing Observations . . . . . 93
6.8.1 Missing Values on Exogenous Variables . . . . . . . . . . . . 93
6.8.2 Missing Values on Dependent Variable . . . . . . . . . . . . . . 93
6.8.3 Tobit Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
6.8.4 An Alternative Two-Stage Approach . . . . . . . . . . . . . . . 95
6.8.5 Problem of Missing Prices and Amemyia
Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
6.8.6 Stone–Lewbel Prices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
6.9 An Alternative Approach Based on Pseudo-Panel Data . . . . . . . . 97
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
7 Derived Demand, Marketing Margins, and Relationship
Between Output and Raw Material Prices . . . . . . . . . . . . . . . . . . . . . . . 101
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
7.2 Historical Approaches to Modeling Derived Demand
for Raw Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
x Contents
7.2.1 Price Spread Relationships with Quantity/Price . . . . . . 103

7.2.2 Derived Demand Elasticities as Products
Between Elasticities of Price Transmission
and Retail Demand Elasticities . . . . . . . . . . . . . . . . . . . . 104
7.2.3 Derived Demand Elasticities Constrained to Be
Less Than Retail Demand Elasticities . . . . . . . . . . . . . . . 105
7.3 Gardner Model of Retail and Farm Price Relationships . . . . . . . . 105
7.3.1 Long-Run Competitive Model with Variable
Proportions Production Function . . . . . . . . . . . . . . . . . . . 105
7.3.2 Comparative Statics of Retail-to-Farm Price Ratio . . . . 106
7.3.3 Comparative Statics of Farm Value Share . . . . . . . . . . . 108
7.3.4 Biased Estimates of Markup Price Equation . . . . . . . . . 109
7.3.5 Derived Demand Elasticities not Proportional
to Retail Demand Elasticities . . . . . . . . . . . . . . . . . . . . . . 110
7.4 Empirical Analysis: The Wohlgenant Model
of Retail-to-Farm Price Linkages . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
7.4.1 Importance of Input Substitutability in Food
Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
7.4.2 No Direct Estimates of Food Consumption
by Commodity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
7.4.3 Reduced-Form Model of Retail and Farm Prices . . . . . 112
7.4.4 Comparative Statics of Reduced Form . . . . . . . . . . . . . . 113
7.4.5 Relationship of Reduced Form to Structure . . . . . . . . . . 115
7.4.6 Empirical Specification . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
7.5 Empirical Application to US Retail-to-Farm Price
Linkages of Meats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
7.5.1 Data Set Consistent with Data Set for Consumer
Demand for Meats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
7.5.2 Weights in Retail Demand Index . . . . . . . . . . . . . . . . . . . 120
7.5.3 Testing Zero Homogeneity Condition . . . . . . . . . . . . . . . 121
7.5.4 Model Estimation with Symmetry Restriction . . . . . . . . 121
7.5.5 Model Estimation with CRTS Restriction . . . . . . . . . . . 124
7.5.6 Sensitivity of Results to Specification of Retail
Demand Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
7.5.7 Results for Retail-to-Farm Price Linkage
Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
Appendix: Input Substitution and Output Demand Uncertainty . . . . . . . . 131
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
8 Retail-to-Farm Demand Linkages, Imperfect Competition,
and Short-Run Price Determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
8.2 Generalized Derived Demand Relationships . . . . . . . . . . . . . . . . . . 137
8.2.1 Properties of System of Derived Demand
Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
Contents xi
8.2.2 Total Demand Elasticities for Retail-to-Farm

Demand Linkages for Meats . . . . . . . . . . . . . . . . . . . . . . 141
8.3 Imperfect Competition and Market Intermediaries . . . . . . . . . . . . . 143
8.3.1 Review of Different Models . . . . . . . . . . . . . . . . . . . . . . . 143
8.3.2 Implications for Reduced-Form Retail and Farm
Price Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
8.4 Short-Run Lags in Price Determination . . . . . . . . . . . . . . . . . . . . . . 148
8.4.1 Causes of Lags Between Retail and Farm Prices . . . . . . 148
8.4.2 Model of Short-Run Price Determination
with Inventory Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . 149
8.4.3 Empirical Application to the US Beef Short-Run
Farm-to-Retail Price Spreads . . . . . . . . . . . . . . . . . . . . . . 152
8.4.4 Asymmetry in Price Changes . . . . . . . . . . . . . . . . . . . . . . 159
Appendix: Derivation of Inventory Cost Function . . . . . . . . . . . . . . . . . . . 160
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
9 Dynamic Consumer Demand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
9.2 History of Empirical Models of Dynamic Consumer
Demand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
9.2.1 Geometric Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
9.2.2 Partial Adjustment Model . . . . . . . . . . . . . . . . . . . . . . . . . 167
9.2.3 State Adjustment Model . . . . . . . . . . . . . . . . . . . . . . . . . . 167
9.3 Dynamic Models with Theoretical Foundations: Myopic
Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
9.3.1 Dynamic Additive Quadratic Demand Model
and Dynamic Linear Expenditure System . . . . . . . . . . . 170
9.3.2 Dynamic Models for Estimating Disaggregated
Commodities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
9.4 Dynamic Models with Theoretical Foundations:
Intertemporal Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
9.4.1 Rational Addiction Model . . . . . . . . . . . . . . . . . . . . . . . . 172
9.4.2 Multivariate Rational Addiction (MRA) Model . . . . . . 173
9.4.2.1 First-Order Conditions . . . . . . . . . . . . . . . . . . . 174
9.4.2.2 Euler Equations . . . . . . . . . . . . . . . . . . . . . . . . . 174
9.4.2.3 Theorem: Closed-Form Solution
to Matrix Difference Equation . . . . . . . . . . . . . 175
9.4.2.4 Formulas for Elasticities . . . . . . . . . . . . . . . . . . 175
9.5 Simple Non-Additive Preferences and Other Modeling
Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
9.5.1 SNAP Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
9.5.2 Spinnewyn Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
9.5.3 Meghir and Weber Short Memory Model . . . . . . . . . . . . 179
9.6 Empirical Application to Dynamic Demand for US Alcohol . . . . 180
9.6.1 Application of MRA Model to Alcohol Demand . . . . . 180
xii Contents
9.6.1.1 GMM Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

Appendix: Proof of Theorem 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
10 Dynamic Models of the Firm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
10.2 Investment Models of the Firm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
10.2.1 Current-Value Hamiltonian and Necessary
and Sufficient Conditions for Value Maximization . . . . 192
10.2.2 Solution for Optimal Capital Accumulation . . . . . . . . . . 193
10.2.3 Complete Model with Short-Run Output
and Variable Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
10.3 Morrison-Paul Imperfect Competitive Model . . . . . . . . . . . . . . . . . 195
10.4 Multivariate Models of Investment . . . . . . . . . . . . . . . . . . . . . . . . . . 196
10.5 Dynamic Inventory Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
10.5.1 Linear-Quadratic Model with Rising
Marginal Costs of Production and Inventories
and Adjustment Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
10.5.2 First-Order Conditions from Maximizing
Expected Present Discounted Value of Net
Returns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
10.5.3 Solution for Optimal Inventory Holding . . . . . . . . . . . . . 199
10.5.4 Comparative Statics of Parameter Changes . . . . . . . . . . 201
10.5.5 Expectations Formation . . . . . . . . . . . . . . . . . . . . . . . . . . 202
10.5.6 Form of the Structural Inventory Behavioral
Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
10.5.7 Relationship of Linear-Quadratic Inventory
Model to Flexible Accelerator . . . . . . . . . . . . . . . . . . . . . 206
10.5.8 Optimal Sales/Price Decisions . . . . . . . . . . . . . . . . . . . . . 206
10.5.9 Input Decisions and Derived Demand . . . . . . . . . . . . . . . 208
10.5.10 Overall Process of Market Equilibrium . . . . . . . . . . . . . 209
10.6 Other Empirically Based Approaches to Modeling
Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
About the Author
Dr. Michael K. Wohlgenant is William Neal Reynolds Distinguished Professor

Emeritus of Agricultural & Resource Economics at North Carolina State Univer-
sity (NCSU). He was Assistant Professor at NCSU from 1978 to 1982; Associate
Professor at Texas A&M University from 1983 to 1986; Associate Professor at
NCSU from 1986 to 1988; Full Professor at NCSU from 1988 until 2018; and
William Neal Reynolds Distinguished Professor from 1996 to 2018 until retire-
ment. He received a Ph.D. in Agricultural Economics from the University of Cali-
fornia at Davis in 1978, an M.S. degree in Agricultural Economics, and a B.S.
degree in Economics from Montana State University in 1973 and 1972, respectively.
His specialty is development of economic models of agricultural marketing, policy,
demand, and price analysis problems. He has developed economic models to under-
stand farm to retail price linkages, consumer demand, market supply/demand, and the
effects of advertising, among other applications. Dr. Wohlgenant has had extensive
commodity experience, including work on applied price and marketing problems for
cotton, dairy, beef, pork, grapes, sugar, tobacco, wine, and horticultural crops. His
work has been published in professional journals, including the American Journal of
Agricultural Economics, Review of Economics and Statistics, Empirical Economics,
Australian Journal of Agricultural and Resource Economics, Journal of Agricultural
and Resource Economics, and the European Review of Agricultural Economics. His
contributions to agricultural economics are numerous and his research is widely cited
and recognized by awards, with Best Article awards in four journals including the
American Journal of Agricultural Economics. He is also a recipient of the Publication
of Enduring Quality Award by the American Agricultural Economics Association.
His H-index from Google Scholar is 34 and i-10 index is 71.
Dr. Wohlgenant is past editor of the American Journal of Agricultural Economics.
He also served as an associate editor of the American Journal of Agricultural
Economics, the European Review of Agricultural Economics, the Journal of Agri-
cultural and Resource Economics, and the Australian Journal of Agricultural and
Resource Economics. He has served as an economic consultant for the Research
Triangle Institute, the Food and Agriculture Organization of the United Nations,
for various law firms in expert witness testimony (including expert witness testi-
mony before the U.S. International Trade Commission). He is an academic affiliate
xiii
xiv About the Author
of Analysis Group. Dr. Wohlgenant has been called upon to be a consultant to the
North Carolina legislature on economic issues related to the N.C. swine industry, the
U.S. Department of Agriculture’s Economic Research Service, and the U.S. Govern-
ment Accountability Office. He also has served as an external reviewer of outside
research programs. Dr. Wohlgenant taught both undergraduate and graduate courses
in economics and agricultural economics at NCSU. He served as chair or co-chair for
thirty PhD students, many of whom hold prestigious positions in academics, govern-
ment, and industry. For his scholarly and research contributions, he was awarded a
William Neal Reynolds Distinguished Professorship in 1996. Dr. Wohlgenant was
elected a Fellow of the Agricultural and Applied Economics Association in 2001.
List of Tables
Table 5.1 Constant Dollar Expenditures Per Capita for Meats,

1970–2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Table 5.2 Price Indices for Meats (1982–1984 = 100), 1970–2010 . . . . . 59
Table 5.3 Expenditures Per Capita for Meats ($ per person),
1970–2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
Table 5.4 Parameter Estimates of Conditional RM for Meats . . . . . . . . . . 62
Table 5.5 Compensated Price and Expenditure Elasticities
for Conditional RM Model of Meats . . . . . . . . . . . . . . . . . . . . . . 63
Table 5.6 Parameter Estimates of Conditional Demand Functions
for LA/AIDS Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
for Conditional LA/AIDS Model of Meats . . . . . . . . . . . . . . . . . 65
Table 5.8 Parameter Estimates of Conditional Demand Functions
for Full AIDS Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
for Conditional Full AIDS Model of Meats . . . . . . . . . . . . . . . . 67
Table 5.10 Expenditure Per Capita and Price Indices for Food,
Nmeats, Nfood, 1970–2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
Table 5.11 Parameter Estimates of First-Stage Allocation Equations
for Relative Price Version of RM, Eqs. (5.5) and (5.6),
and Second-Stage Conditional Demand Eqs. (5.7) . . . . . . . . . . 70
Table 5.12 Compensated Unconditional Price and Expenditure
Elasticities for Meats when Eqs. (5.5), (5.6), and (5.7)
are Jointly Estimated . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Table 5.13 Compensated Conditional Price and Expenditure
Table 5.14 Uncompensated Unconditional Price and Expenditure
Table 7.1 Farm Quantities (Mil. Lb.), Prices & Wage (1982-100),
and Population (000) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
xv
xvi List of Tables
Table 7.2 Cross-Price and Income Elasticity Estimates, and Farm

Value Shares . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Table 7.3 Parameter Estimates of Constrained Reduced Form
and Structural Retail-to-Farm Price Linkage Equations,
Beef and Veal, 1970–2020 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Pork, 1970–2020 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Poultry, 1970–2020 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
and Structural Retail-to-Farm Price Linkage Equations
with CRTS, Beef and Veal, 1970–2010 . . . . . . . . . . . . . . . . . . . . 126
with CRTS, Pork, 1970–2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
with CRTS, Poultry, 1970–2010 . . . . . . . . . . . . . . . . . . . . . . . . . 127
Table 7.9 Estimates of Retail-to-Farm Price Ratios for Beef
and Veal, Pork, and Poultry, 1970–2010 . . . . . . . . . . . . . . . . . . . 131
Table 8.1 Total Impacts of Farm Supplies, Wage Rate,
and Exogenous Retail Demand Shift Variables on Farm
Prices, Retail Prices, and Derived Demands for Beef
and Veal, Pork, and Poultry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
Table 8.2 Farm-to-Retail Price Spread and Wage Rates for US
Beef, January 2009–December 2018 . . . . . . . . . . . . . . . . . . . . . . 154
Table 8.3 Econometric Estimates of the Farm-to-Retail Price
Spread Model for Beef, January 2009–December 2018 . . . . . . 158
Table 9.1 Per Capita Constant Dollar Expenditures, Price Indices
for Alcohol, and Per Capita Real Personal Consumption
Expenditures, 1995Q1–2015Q4 . . . . . . . . . . . . . . . . . . . . . . . . . 181
Table 9.2 Parameter Estimates of Dynamic Demand for Wine,
Beer, and Spirits using GMM . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
Table 9.3 Elasticities of Dynamic Demand for Wine, Beer,
and Spirits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
Chapter 1
Introduction
1.1 Background
This book is based on a set of class notes I used in a graduate course in economics
and agricultural economics at North Carolina State for over thirty years. The course,
entitled “Consumption, Demand, and Market Interdependencies,” addresses the core
issues facing economists concerning price determination in commodity markets,
especially food and agricultural commodities. The focus is on the conceptual basis
of the various relationships, with special emphasis on market interrelationships, both
horizontally and vertically. Many problems exist when moving from the theoret-
ical or conceptual models to the empirical applications, and considerable attempt is
undertaken to bridge this gap.
Bridging the gap between theory and empirical analysis is a hallmark of this
book. Going from theory to empirics requires that we have data—time-series or
cross section—that match the theoretical constructs. Often the data match is not
perfect, either by definition or how the data are computed. For example, the USDA
Economic Research Service publishes per capita consumption by food commodity,
although the per capita consumption values are based on assumptions about how
food is transformed into the final consumer good, which may or may not be true
for industry practice. While these data have useful purposes for policy purposes,
they may not be appropriate for demand estimation, or may require further care
when used for econometric analysis. In addition to problems of matching data with
theoretical constructs, the researcher needs to know how to specify, estimate, and
interpret results within the context of imperfect and often incomplete data. A prime
example here is how to handle missing values for purchases with cross-section data.
It is for these reasons that I use several data sets in this book to illustrate how one
might address the problems and help bridge the gap between theory and empirics.
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 1

M. K. Wohlgenant, Market Interrelationships and Applied Demand Analysis,
Palgrave Studies in Agricultural Economics and Food Policy,
https://doi.org/10.1007/978-3-030-73144-1_1
2 1 Introduction
1.2 Organization of Chapters
Chapter 2 lays out the basic consumer demand model and determinants of the various
demand relationships one encounters in empirical work. The focus is on the compar-
ative static results and quantitative restrictions resulting from the analysis. I use the
primal approach, based on the direct utility function, to develop these results because
it is the most comprehensive, allowing for quantitative analysis of shifts in prefer-
ences. For completeness, the major duality results are presented, and comparative
static results for inverse demand functions are reported.
In the spirit of Brandow (1961) and George and King (1971), attention is directed
toward demand interrelationships at the retail level and derived demand interrela-
tionships for raw materials producing the final retail products. Separability in pref-
erences plays a central role in deriving these relationships. Much confusion has
surrounded the application of separability in two-stage budgeting. In Chapter 3,
considerable effort is made to clarify application of these concepts in estimation of
demand interrelationships.
Chapter 4 is a review of alternative functional forms used in estimating systems of
demand functions. The focus is on empirical tractability. Particular attention is given
to the Rotterdam Model (RM) and the Almost Ideal Demand System (AIDS), because
of their common and extensive use. Discussion focuses on critique of these models
and relative advantages/disadvantages compared to other modeling approaches in the
literature. Generalizations of the RM and AIDS model, including the recent EASI
demand system, are also presented.
A notable feature of this book is original empirical applications to different data
sets constructed from published data. In Chapter 5, a data set on meats is developed
and applied to estimation of consumer demand using two common functional forms:
RM and the AIDS. Both conditional and unconditional demand models are estimated
and major hypotheses of consumer demand are tested.
Consumption, income, and prices are also spatially interrelated. In Chapter 6,
cross-sectional and panel demand models are presented and evaluated. As with
time-series applications, the focus is on models with demand interrelationships.
Special attention is given to quality variation, caused by product heterogeneity, which
leads to estimation complications, particularly for consumption-income relation-
ships. Models of product heterogeneity, including linear utility and hedonic distance
modeling approaches, are presented and critiqued. Specification and estimation of
panel data and pseudo-panel data models are discussed. A major issue is how to
model zero values of consumption, especially in a system of demand equations.
Although empirically tractable approaches exist and are discussed, problems still
exist and knowledge of sources of zero observations is needed to avoid estimation
problems.
Consistent with Gardner (1975) and Wohlgenant (1989), theoretically consistent
retail-to-farm price linkage models are developed and presented in Chapter 7. The
validity and usefulness of the approach is illustrated through empirical application
to meats. The data set in Chapter 5 is extended to include meats at the farm level. A
1.2 Organization of Chapters 3
large part of the chapter is on the empirical analysis, showing how to implement the
theoretical framework in the absence of data for final consumption. Model testing
and sensitivity of the results to errors in data are highlights of this chapter.
In Chapter 8, the theory of derived demand is extended to general systems of
derived demand functions. The results of Theorem 3 are used to evaluate the retail-
to-farm demand linkages in an interrelated framework, extending the earlier work of
Wohlgenant (1989). The framework allows one to consistently estimate the effects
of exogenous changes in retail demand, marketing input prices, and raw material
(farm output) supplies on retail and raw material prices. This chapter also discusses
the modification of the models of market intermediary behavior to the degree of
competitiveness in food industries, with a focus on meat industries.
In the short run, inventories can play a prominent role in price determination.
In Chapter 8, I discuss theoretical foundations for lags between retail and producer
prices using a model developed by Wohlgenant (1985). The model is applied to
monthly data for beef showing how implicit inventory costs can explain a nega-
tive relationship between farm-to-retail price spreads and farm prices, a widely
proclaimed empirical puzzle. Causes of, and testing for, asymmetry in price response
are also discussed.
In Chapter 9, dynamic models of consumer behavior are presented and discussed.
Habit formation is the main cause of dynamic consumer behavior and significant
models in the literature developed for this purpose are discussed. Special attention is
devoted to intertemporal behavior. As in the static analysis, I present empirical results
to illustrate how to go from theory to empirics. I present and evaluate econometric
results for alcoholic beverages using a data set constructed from published data.
In Chapter 10, formulation and estimation of dynamic factor demand relationships
is the subject. Adjustment costs and product demand and supply uncertainty interact
to affect firms’ investment decisions in capital. Single capital adjustment models as
well as multivariate models of factor adjustment costs are presented. Particular focus
of the chapter is on inventory models of the firm to describe food processor behavior.
As in past chapters, different modeling approaches are discussed, with emphasis on
empirical tractability.
Over thirty-plus years of teaching this course, topics covered in the course have
come and gone. An attempt has been made to update the material to current appli-
cations in the literature. Moreover, I have added new results to many of existing
topics and have added significantly to the empirical illustrations. While the topics
addressed do not exhaust all the topics and models presented in the literature, I believe
the major topics of concern to demand analysts are addressed. My approach has been
to be selective, but rigorous and comprehensive, on the topics addressed. The reader
will find an emphasis on price-taking, competitive behavior in the book, although
reference is made to some of the main ways one can modify the models to account for
imperfectly competitive behavior. In addition, some topics are not covered. Specifi-
cally, spatial equilibrium and spatial economic modeling are not covered in this book
because excellent work can already be found elsewhere (e.g., Fackler and Goodwin
2001).
4 1 Introduction
References
Brandow, G.E. Interrelations Among Demands for Farm Products and Implications for Control of
Market Supply. Pennsylvania Agric. Exp. Sta. Bul. 680, University Park, PA, 1961.
Fackler, P.F., and B.K. Goodwin. “Spatial Price Analysis.” In B.L. Gardner and G.C. Rausser (eds.)
Handbook of Agricultural Economics, Vol. 1b, , pp. 971–1026. Amsterdam: Elsevier Science
B.V., 2001, Chapter 17.
Gardner, B. “The Farm-Retail Price Spread in a Competitive Food Industry.” American Journal of
Agricultural Economics 57(1975): 399–409.
George, P.S., and G.A. King. Consumer Demand for Food Commodities in the United State
with Projections for 1980. Giannini Foundation Monograph No. 26, California Agricultural
Experiment Station, Berkeley, CA, 1971.
Wohlgenant, Michael K. “Competitive Storage, Rational Expectations, and Short-Run Food Price
Determination.” American Journal of Agricultural Economics 67(1985): 739–748.
Wohlgenant, Michael K. “Demand for Farm Output in a Complete System of Demand Functions.”
American Journal of Agricultural Economics 71 (1989): 241–252.
Chapter 2
Consumer Demand—Theory
2.1 Introduction
The neoclassical theory of consumer behavior constitutes the conceptual basis for
the demand analysis framework formulated in this book. In this chapter, I develop
the general results using the primal approach in order to account for all factors
affecting consumer behavior. The main result is the Fundamental Demand Matrix
of Consumer Demand (Barten 1964). Additional topics include duality theory and
inverse demand functions. Comprehensive treatments of the topics covered here
can be found elsewhere (e.g., Phlips 1983; Deaton and Muellbauer 1980b; Theil
1975/1976).
2.2 Consumer’s Choice Problem
The theory of consumer behavior is that the individual makes rational decisions that
satisfy a given set of axioms. These axioms are: (1) reflexivity, (2) completeness,
(3) transitivity, and (4) continuity. These four axioms imply that we can represent
the consumer’s preferences with a single-valued utility function. With the addi-
tional axioms of (5) non-satiation and (6) convexity, the utility function will be
quasi-concave with indifference curves exhibiting diminishing marginal rates of
substitution (Varian 1992).
The consumer’s choice problem is as follows:
max u = v(q)
q
subject to:

https://doi.org/10.1007/978-3-030-73144-1_2
6 2 Consumer Demand—Theory
p q ≤ y
q≥0
where q is an nx1 vector of quantities of commodities consumed, p is the corre-

sponding price vector, y is total expenditures or income, and v(q) is the utility
function. The two sets of constraints are the budget constraint and the non-negativity
requirement of quantities purchased and consumed.
2.2.1 First-Order Conditions: Application

of the Kuhn-Tucker Theorem
Consider the Lagrangian function:
n
L = v(q) + λ y− p q + μi qi
i=1
where λ and μi are Lagrange multipliers associated with the budget and non-
negativity constraints, respectively. By the Kuhn-Tucker Theorem
∂ L ∂v(q)
= −λpi + μi = 0, ∀i
∂qi ∂qi
subject to the budget constraint and non-negativity constraints on the quantities and
the Lagrange multipliers. In addition, there are complementary slackness conditions:

λ y− p q = 0
μi qi = 0, ∀i
By the Envelope Theorem we know that λ is the marginal utility of income, and it
is positive according to axioms 5 and 61 . Thus, taking account of the complementary
conditions, the first-order conditions for utility maximization are:
∂v(q)
−λpi ≤ 0, ∀i
∂qi
1 The Envelope Theorem states that the derivative of the optimum value of the function (in this
case the Lagrangian function) with respect to an exogenous parameter equals the partial derivative
of the function with respect to the specific exogenous parameter (Silberberg 1990). In this case,
the optimized value is maximum utility and the partial derivative of the Lagrangian function with
respect to y is λ; hence, λ is the marginal utility of income.
2.2 Consumer’s Choice Problem 7
p q = y
The first set of first-order conditions indicate that corner solutions are possible
if zero quantities are purchased. This means there are additional factors to consider
when estimating models with data, such as cross-sectional data, where zero purchases
are frequent. On the other hand, if interior solutions exist, qi >0 and μi = 0 according
to the complementary slackness conditions. Therefore, the first-order conditions can
be expressed as follows:
∂v(q)
−λpi = 0, ∀i
∂qi
p q = y
2.2.2 Demand Functions
Because the utility function is quasi-concave, these first-order conditions are both
necessary and sufficient for utility maximization. Moreover, if the utility function is
strictly quasi-concave, unique solutions of q and λ exist for given prices and income:
q = g(y, p)
λ = λ(y, p)
2.3 Comparative Statics of Income and Price Changes
In anticipation of the ensuing empirical specification, a vector of exogenous shift

variables s is introduced into the utility function. Both the first-order conditions and
solutions to the first-order conditions depend on this r x1 vector of parameters:
∂v(q, s)
−λpi = 0, ∀i (2.1)
∂qi
p q = y (2.2)
q = g(y, p, s) (2.3)
λ = λ(y, p, s) (2.4)
2.3.1 Fundamental Matrix Equation of Consumer Demand
Comparative statics of the effects of changes in prices, income, and taste change
parameters on demand can be quantified through totally differentiating (2.1) and
2 2
∂qi ∂q1 ∂q2 ∂qn
(2.2). Define U = ∂q∂i ∂q v
, V = ∂ v
∂qi ∂s j
, Q p = ∂pj
, q y = ∂y
, ∂y
, . . . , ∂y
,

j

λ p = ∂∂λ , ∂λ , . . . , ∂∂λ
p1 ∂ p2 pn
, λs = ∂s∂λ ∂λ
1
∂λ
, ∂s2 , . . . , ∂sr
, dq = (dq1 , dq2 , . . . , dqn ) ,

∂λ
d p = (dp1 , dp2 , . . . , dpn ) , d s = (ds1 , ds2 , . . . , dsr ) , and λ y = ∂y
. Then the
comparative static results can be expressed as follows:
⎡ ⎤
dy
U p dq o λI −V ⎣
= d p⎦ (2.5)
p o −dλ 1 −q 0
ds
The total differentials of Eqs. (2.3) and (2.4) are as follows:
⎡ dy ⎤
dq q y Q p Qs ⎣
= d p⎦ (2.6)
−dλ −λ y −λ p −λs
ds
Therefore, at the initial equilibrium point, equating (2.5) with (2.6) yields:

U p q y Q p Qs o λI −V
= (2.7)
p o −λ y −λ p −λs 1 −q 0
Equation (2.7) is the Fundamental Matrix Equation of Consumer Demand (Barten

1964; Phlips 1983)2 . This matrix provides a concise summary of all the comparative
static effects of the static theory of consumer behavior. We can derive the specific
comparative static results through solving for the (n + 1)x(n + r + 1) second matrix
on the left-hand side of (2.7), provided the first matrix is non-singular. But we know
that it is non-singular because of the assumption that the utility function is strictly
quasi-concave in q, which implies that this matrix is non-singular and in fact is
strictly negative semi-definite. This also implies that the matrix
2 Note that this is really the augmented Fundamental Matrix Equation of consumer demand because it
includes the comparative static results for preference parameter shift variables, whereas the original
result did not include the preference shift variables.
2.3 Comparative Statics of Income and Price Changes 9
−1
U p Z z
=
p o z z
exists and is strictly negative semi-definite. Using this result in (2.7) yields the specific
comparative static results:

q y Q p Qs Z z o λI −V
= (2.8)
−λ y −λ p −λs z z 1 −q 0
Writing out the solutions for the partial derivatives of the demand functions and
marginal utility of income gives:
qy = z (2.9)
Q p = λZ − zq (2.10)
Q s = −ZV (2.11)
−λ y = z (2.12)

−λ p = −λz − zq (2.13)

−λs = −z V (2.14)
2.3.2 Slutsky Substitution Matrix
Substituting (2.9) into (2.10) and defining λZ = K we obtain the matrix of

generalized Slutsky equations:

Q p = K − qyq (2.15)

The matrix K is the substitution matrix and −q y q is the matrix of income effects.
To show that K is indeed the substitution matrix, note that utility held constant

means du = vq dq = 0. From the first-order conditions (2.1), du = vq dq =

λ p dq = 0. Substituting into the differential form of the budget constraint (2.5),
holding s constant, gives: dy|d s=0 = q d p. Substitution into the differential of the
demand functions (2.6) (holding s constant) gives dq|d s=0 = Q p d p + q y dy =

Q p d p + q y q d p = Q p + q y q d p = K d p. Therefore, the matrix K is the matrix
of income-compensated price changes to maintain the original utility level.
The Slutsky substitution matrix has the following properties: (1) negative semi-
definite, (2) non-positive own-price effects, (3) symmetric price effects, and (4) rank
≤ n − 1. The matrix K is negative semi-definite because Z = λ−1 K is the principal
−1
U p
n − 1 submatrix of which we already know is strictly negative semi-
p o
definite. Because any principal submatrix of any negative semi-definite matrix must
also be negative semi-definite, K is negative semi-definite because λ−1 > 0. This
immediately implies all diagonal elements of K are non-positive. This matrix is
−1
U p
also symmetric because it is proportional to the principal submatrix of ,
p o
which is symmetric. Therefore, ki j = k ji ∀i = j, or all cross-compensated price
effects are equal. Finally, the matrix K has less than full rank because the columns
of K are linearly dependent. To see this, post-multiplying K by the price vector, we

have K p = Q p + q y q p = Q p p + q y q p = Q p p + q y y, upon substituting
for the budget constraint. This vector is a zero vector for the simple reason that the
uncompensated (Marshallian) demand functions (2.3) are homogeneous of degree
zero in all prices and income. This is easy to see because the proportional price and
income changes only affect the budget constraint since preferences are independent
of prices and income. The expression Q p p + q y y is then a zero vector because it is
the set of Euler equations for functions homogeneous of degree zero in prices and
income. Therefore, K p = 0 or the Slutsky substitution matrix has at most rank equal
to n − 1.
In addition to the properties of the Slutsky equation, we also know that the
demand functions, as just shown, are homogeneous of degree zero in prices and
income, known as the Homogeneity Condition. In addition, the demand functions
must satisfy the adding-up property, which states that the sum of the demand func-
tions, weighted by their respective prices, must exactly equal income. There are two
important derivative properties of the adding-up condition.
n
The first is the Engel Aggregation Condition: pi ∂g
∂y
i
= 1. The second derivative
i=1

n
∂g
property is the Cournot Aggregation Condition: −qi = p j ∂ pij ∀i.
j=1
2.3.3 General Restrictions of Consumer Behavior
In general, there are two types of demand functions: (1) Marshallian or nominal-
income constant demand functions (2.3), and (2) Hicksian or real income constant
demand functions:
q = h(u, p, s) (2.16)
2.3 Comparative Statics of Income and Price Changes 11
The Homogeneity requirement for this set of demand functions is that they be
homogeneous of degree zero in p, implying K p = 0. There are no adding-up
requirements on this set of demand functions and the conditions on the substitution
matrix are as indicated above.
The general restrictions of consumer behavior are summarized below for both the
Marshallian and Hicksian demand functions in elasticity form. Define ei j = ∂∂gpij qij ,
p
ei = ∂g , ei∗j = ∂∂hp ij qij , and wi =

i y p pi qi
∂ y qi y
. Then the conditions are as follows:
Engel aggregation:

n
wi ei = 1 (2.17)
i=1
Cournot aggregation:

n
−wi = w j e ji ∀i (2.18)
j=1

n
Homogeneity condition: ei + ei j = 0;
j=1

n
ei∗j = 0 (2.19)
j=1
Symmetry condition: ei j + w j ei = e ji + wi e j ;
wi ei∗j = w j e∗ji (2.20)

Negative semi-definiteness: wi ei∗j is negative semi-definite with
rank ≤ n − 1 (2.21)
Negativity: eii + wi ei ≤ 0, eii∗ ≤ 0∀i (2.22)
2.4 Comparative Statics of Exogenous Preference Shift

Variables
The effects of preference shift parameters from (2.11) are

Q s = −ZV = −λ−1 K V (2.23)
These effects are proportional to the Slutsky substitution effects. This means
there is an intrinsic relationship to price effects in the Marshallian demand functions.
Consider the set of total differentials of the Marshallian demand functions (2.6).
Substituting (2.15) and (2.23) into (2.6) and collecting terms gives:

dq = K d p − λ−1 V d s + q y dy − q d p (2.24)
This specification shows, as before, that demand can be decomposed into substi-
tution and income effects. This decomposition also shows that effects of exoge-
nous preference changes can be quantified as adjustments to price changes. If, for
example, s represents quality characteristics, then the relevant price changes are
quality-adjusted prices. One can envision following a two-step process, in such an
instance, where one first estimates hedonic price equations and then uses these resid-
uals to obtain the quality-adjusted prices in the demand model. Such an approach
has been taken by Cowling and Raynor (1970) in the context of tractor prices, and
Cox and Wohlgenant (1986) in the context of demand for foods. To the author’s
knowledge, this is the first time a theoretical justification has been provided for this
specification. In general, the decomposition of price changes shows that specific
restrictions are imposed on the demand structure from introducing exogenous shift
parameters into the utility function, whatever they may be.
2.5 Comparative Statics with Strictly Concave Utility

Function
We can relate changes in marginal utilities to changes in demand by assuming further

that the utility function is strictly concave in q. This assumption, although more
restrictive than quasi-concavity, is allowable because strict quasi-concavity is an
increasing, monotonic transformation of strict concavity; therefore, the preference
ordering is preserved and one obtains the same solutions for q and λ. When the utility
function is strictly concave, we know that U is non-singular so that U −1 exists. By
the rule for partitioned inverse (e.g., Hadley 1961):
⎡ −1 −1 ⎤
U p
−1
U −1 − U −1 p p U −1 p p U −1 U −1 p p U −1 p
=⎣ −1 −1 ⎦
p o p U −1 p p U −1 − p U −1 p
Z z
Replacing on the right-hand side of (2.8) with the above partitioned
z z
solution, we obtain the comparative static results:
2.5 Comparative Statics with Strictly Concave Utility Function 13

K = λZ = λU −1 − λ λ y q y q y

(2.25)

−λ p = −λq y − λ y q (2.26)
−1
−1
q y = U −1 p p U p (2.27)
−1
−1
−λ y = − p U p (2.28)
Equation (2.25) shows that thesubstitution matrix can be decomposed into two

effects: the matrix λU −1 and − λ λ y q y q y . With U i j referring to the (i, j) th

element of the inverse matrix U −1 , the (i, j) th element of K is:

∂q ∂q
ki j = λU i j − λ λ y
i j
(2.29)
∂y ∂y
Houthakker (1960) refers to λU i j as the specific substitution effect of the total

substitution effect because it gives the income-compensated price change to hold the
marginal utility of income, λ y , constant. It follows that − λ ∂qi ∂q j
λ y ∂ y ∂ y is the general
substitution effect, because it shows the overall effect on demand from an income-
compensated price change. As will be shown later, this decomposition will prove
very important when it comes to imposing restrictions on preferences to aggregate
price effects. As an example, suppose that all goods are independent in the sense that
the marginal utility ofgoodi only depends on consumption of good i. Then it is easy

to show that ki j = − λ λ y ∂q i ∂q j
∂y ∂y
∀i = j. In this case, all cross-compensated price
effects are proportional to income effects. As we shall see, weaker restrictions than
this can produce desired aggregation effects.
It is useful to express (2.29) in elasticity form:
ei∗j = ei j − φei w j e j ∀i, j

f
(2.30)
The decomposition of the Hicksian price elasticities is into Frisch-compensated

f p
elasticities, ei j = λU i j qij , which is named after the renowned econometrician,
Ragnar Frisch (Frisch 1959). It is the income-compensated price elasticity to keep the
marginal utility of income constant, thereby reflecting specific substitution between
goods. The scalar, φ, is the so-called “money flexibility,” or the reciprocal of the
elasticity of marginal utility of income with respect to income.
2.6 Duality Theory: Relationships Between Marshallian,

Hicksian, and Frischian Demands
In addition to Marshallian demand functions, (2.2), and Hicksian demand functions,

(2.16), there are also Frischian demand functions:
q = f (y, r, p, s) (2.31)
where r = λ−1 is the price of utility. These sets of demand functions are related to,
and can be derived from, the indirect utility function, the expenditure function, and
the (consumer) profit function, respectively3 .
First, with respect to the Marshallian demand functions we have:

q = g(y, p, s) = arg max v(q, s)| p q = y (2.32)
q

ϕ(y, p, s) = max v(q, s)| p q = y (2.33)
q
The indirect utility function, ϕ(y, p, s), is increasing in income, decreasing in

prices, convex in prices, and homogenous of degree zero in income and prices.
Roy’s identity can be applied to the indirect utility function to obtain Marshallian
demand functions4 :
− ∂ϕ(y, p,s)
∂ pi
qi = − ∂ϕ(y, p,s)
(2.34)
∂y
The relationships associated with the Hicksian demand functions are as follows:

q = h(u, p, s) = arg min p q = y u = v(q, s)| (2.35)
q

C(u, p, s) = min p q = y u = v(q, s)| (2.36)
q
The expenditure or cost function,C(u, p, s), is increasing in prices and utility,

concave in prices, and homogeneous of degree one in prices. Shephard’s lemma can
be applied to the cost function to obtain Hicksian demand functions5 :
3 See, e.g., Sproule (2013) for a concise discussion on the relationship between these various
relationships. See Cornes (1992) for a comprehensive approach to duality.
4 Roy’s identity can be derived from the Envelope Theorem through differentiation of the maximized
value of the Lagrangian function associated with the constrained maximization problem (2.33) with
respect to both y and pi .
5 Shephard’s lemma can be derived from the Envelope Theorem through differentiation of the
maximized value of the Lagrangian function associated with the constrained minimization problem
(2.36) with respect to pi .
2.6 Duality Theory … 15
∂C(u, p, s)
qi = (2.37)
∂ pi
Finally, we have the relationships corresponding to the Frischian demand

functions:

q = f (r, p, s) = arg max r v(q, s) − p q (2.38)
q

π (r, p, s) = max r v(q, s)− p q (2.39)
q
The profit function, π(r, p, s), is increasing in the price of utility, decreasing in
prices, homogenous of degree one in r and p, and convex in r and p. Hotelling’s
theorem applied directly to the profit function yields the Frischian demand functions6 :
∂π (r, p, s)
qi = − (2.40)
∂ pi
At the same initial equilibrium point, it can be shown that all three demand
functions are equal, i.e.,:
q = g(y, p, s) = h(u, p, s) = f (r, p, s) (2.41)

To see this, note first that because r = λ−1 , q = f (r, p, s) = f ∂C(u, p,s)
, p, s =
−1 ∂u
h(u, p, s). Second, q = f (r, p, s) = f ∂ϕ(y, ∂y
p,s)
, p, s = g(y, p, s).
Equating the two leads to (2.41).
2.7 Inverse Demand Functions
It is sometimes the case that we want demand functions, expressed in terms of

prices, as functions of quantities and income. For example, if supplies of goods are
predetermined, such as quantities of different varieties of fresh fish, market prices
will adjust to clear the market in response to changes in fixed supplies of goods
offered for sale. The following discussion parallels Barten and Betandorf (1989).
A distance function approach to specification of inverse demand functions can be
found in Anderson (1980).
6 By the Envelope Theorem, partial differentiation of the maximized value of the profit function,
(2.39), with respect to pi equals the negative of the Frischian demand function.
2.7.1 Wold-Hotelling Identity
Recall the system of Marshallian demand functions, q = g(y, p), where we have
suppressed preference shift variables, s, for convenience. These demand functions,
which are homogeneous
of degree zero in income and prices can be expressed as
q = g(π ), π = y p. We can solve for normalized prices to obtain
1
π = g −1 (q) = b(q) (2.42)
which is the vector of uncompensated inverse demand functions. The first-order

conditions for utility maximization are
v q = μπ (2.43)
1 = π q (2.44)
where μ = yλ. Multiplying (2.43) by q and using (2.44) yields the so-called Wold-
Hotelling identity:

1
π= v q = b(q) (2.45)
q vq
The Wold-Hotelling identity delivers directly the uncompensated inverse demand

functions, which we see express normalized prices (i.e., prices deflated by total
expenditures) as functions of all n commodity quantities.
2.7.2 Comparative Statics
In order to determine the properties of uncompensated inverse demand functions,

totally differentiate (2.45) to obtain:

1
dπ = −π v q dq + I − π v q dv q
q vq

= −π π dq + I −π v q U ∗ dq
where U ∗ = q 1v U, with U = ∂q∂i ∂q
v
2
j
as defined previously. We can rearrange
q
this last term to obtain:
dπ = gπ dq + Gdq (2.46)
2.7 Inverse Demand Functions 17
where

g = − π − I − qπ U ∗ q (2.47)

G = I − qπ U ∗ I − qπ (2.48)
The form of (2.46) shows that the total change in inverse demand can be decom-
posed into two effects given by (2.47) and (2.48). The first is a scale effect (Anderson,
1980). Define a proportionate increase in the components of q as dq = κq, κ a posi-
tive scalar. It follows that π dq = κπ q = κ. Note using (2.48) that Gq = 0. There-
fore, the scale effect only works through the first term in (2.46). The change in scale

is monotonously related to change in utility. Note that du = v q dq = μπ dq = μκ.
This means Gdq is the utility or real income-compensated effect of quantity changes.
The matrix G, called the Antonelli substitution matrix, is the counterpart to the Slutsky
substitution matrix.
The scale measure, π dq, can be representedas the Divisia volume index,
d log Q. This follows from the fact that π dq = π j dq j = π j q j d log q j =
j j
w j d log q j = d log Q. Therefore, we can rewrite (2.46) as
j
dπ = gd log Q + Gdq (2.49)
Totally differentiating the budget constraint gives π dq + q dπ = 0, which

implies q dπ = −π dq = −d log Q. Premultiplying both sides of (2.49) by q
gives q dπ = −d log Q, provided q g = −1 and
q G = 0 (2.50)
which are adding-up properties. We also have the result
Gq = 0 (2.51)
or the homogeneity restriction. Finally, we have the result that G is a negative semi-
definite matrix with rank ≤ n − 1. This follows from Eq. (2.48), showing G is a
quadratic function of U ∗ which is negative semi-definite, and from (2.51), showing
the n columns of G are linearly dependent.
The compensated inverse demand functions, π = a(q, u), are negative semi-
definite in quantities, implying all own-quantity changes are non-positive. Commodi-
ties whose cross-quantity effects are negative are called q-substitutes and cross-
quantity effects that are positive are called q-complements. The counterpart to the
Slutsky equation can be derived from the relationship between compensated and
uncompensated inverse demand functions:

π = a(q, u) = a q, v(q) = b(q) (2.52)

∂πi ∂ai ∂ai ∂v ∂bi
For the ith good, ∂q j
= ∂q j u
+ ∂u ∂q j
= ∂q j
. In light of (2.49), this expression
becomes
∂πi ∂bi
= gi j + gi π j = (2.53)
∂q j ∂q j
The total effect of a change in quantity of good j on normalized price i is the sum
of the Antonelli substitution effect and the scale effect. These functions are subject to
the adding-up, homogeneity, symmetry, and negativity restrictions discussed above.
Problems
2.1 Using the Kuhn-Tucker Theorem, derive the Marshallian demand functions
for the linear utility function, u = v(q1 , q2 ) = α1 q1 + α2 q2 .
2.2 Verify the elasticity form of the general restrictions, Eqs. (2.17)–(2.18).
2.3 Show in the two good case that diminishing MRS does not imply diminishing
MU, and vice versa. (Hint: note that the indifference curve in this case is
2
u 0 ≡ v[q1 , φ(q1 )], q2 = φ(q1 ). The MRS is − dq dq1
= vv21 . Diminishing
u=u 0
MRS means d M RS
dq1
< 0.)
2.4 Prove that the matrix λU i j is the specific substitution matrix obtained by
income-compensating price changes holding the marginal utility of income
constant.
2.5 Using the Envelope Theorem, derive: (a) Roy’s Identity, Eq. (2.34); (b)
Shephard’s lemma, Eq. (2.37); and (c) Hotelling’s Theorem, Eq. (2.40).
n
β
2.6 Consider the Cobb-Douglas utility function, u = v(q) = qj j.
j=1
(a) Derive Marshallian, Hicksian, and Frischian demand functions using

the duality relationships in Eqs. (2.31)–(2.40). Specifically, verify that
condition (2.41) holds.
(b) Show that the same results would be obtained if the utility func-
tion had either the form (i) u = v(q) = nj = 1 β j log q j or (ii)

β / n β
u = v(q) = nj = 1 q j j k = 1 j . Why is it the case that the resulting
demand functions are the same with these two utility functions?
2.7 Show that the elasticity form of the Antonelli equation is f i j = f i∗j + f i w j ,
where f i j is the elasticity of the ith normalized price with respect to the jth
quantity (flexibility of ith price with respect to jth quantity), f i∗j is the compen-
sated price flexibility, f i is the scale flexibility (i.e.„ relative change in price
to an equal proportional change in all quantities), and w j is the budget share
Problems 19
as defined before. Compare and contrast this form with the Slutsky equation,
ei j = ei∗j − ei w j .
References
Anderson, R.W. “Some Theory of Inverse Demand for Applied Demand Analysis.” European
Economic Review 14(1980): 281–290.
Barten, A.P. “Consumer Demand Functions under Conditions of Almost Additive Preferences.”
Econometrica 32(1964): 1–38.
Barten, A.P., and L.J. Bettendorf. “Price Formation of Fish: An Application of an Inverse Demand
System.” European Economic Review 33(1989): 1509–1515.
Cornes, R. Duality and Modern Economics. New York, NY: Cambridge University Press, 1992.
Cowling, K., and A.J. Raynor. “Price, Quality, and Market Share.” Journal of Political Economy
78(1970): 1292–1309.
Cox, T., and M.K. Wohlgenant. “Prices and Quality Effects in Cross-Sectional Demand Analysis.”
American Journal of Agricultural Economics 68(1986): 908–919.
Deaton, A.S., and J. Muellbauer. Economics and Consumer Behavior. Cambridge: Cambridge
University Press, 1980b.
Frisch, R.A. “A Complete Scheme for Computing All Direct and Cross Demand Elasticities.”
Econometrica 27(1959): 177–196.
Hadley, G. Linear Algebra. Reading, MA: Addison-Wesley Publishing Company, Inc., 1961.
Houthakker, H.S. “Additive Preferences.” Econometrica 28(1960): 244-257.
Phlips, L. Applied Consumption Analysis. Amsterdam: North-Holland Publishing Co., 1983.
Silberberg, E. The Structure of Economics: A Mathematical Analysis. New York: McGraw Hill
Publishing Co., 1990.
Sproule, R. “A Systematic Analysis of the Links Amongst the Marshallian, Hicksian, and Frischian
Demand Functions: A Note.” Economics Letters 121(2013): 555–557.
Theil, H. Theory and Measurement of Consumer Demand, Vol. I. Chicago: University of Chicago
Press, 1975/1976.
Varian, H.R. Microeconomic Analysis. New York: W.W. Norton & Co., 1992.
Chapter 3
Consumer Demand—Separability
and Commodity Aggregation
3.1 Introduction
A significant problem to address prior to empirical implementation of the theory of

consumer demand is the problem of degrees of freedom. In particular, because of the
large number of goods purchased by the consumer, there are literally thousands of
individual goods and, therefore prices of related goods, one would need to include
in the analysis to make it complete. To see why this is a problem, there are n + n 2
income and price elasticities to estimate. The general restrictions
provide a total
of 1 + n + 21 n(n−1) restrictions, leaving a net total of 21 n 2 + n−2 elasticities to
estimate. Even for a modest number of commodities, there are still a relatively large
number of elasticities to estimate requiring a relatively large data sample1 .
3.2 Restrictions on Price Movements
The traditional and most common approach to commodity aggregation has been
Hicks Composite Commodity Theorem, which says that for a group of commodities
whose prices move together we can aggregate them into a single composite group with
a single price index. In the case of proportionalprice changes within the group, the
composite good for quantity of, say, group G is i∈G pi* qi (Deaton and Muellbauer,
1980b, Chapt. 5), where pi* denotes price for some base period. The group price
index would be obtained by dividing group expenditures,
y G = i∈G pi qi , by the
pi qi
composite quantity to obtain the Paasche price index, p * q . A less known form
i∈G
i∈G i i
of the Composite Commodity Theorem is when price differences of goods within
the composite move together. The composite quantity in this case is the simple
1 As an example, if the number of goods is 10, then we still have 54 income and price elasticities to
estimate. This is clearly too large for most time-series data sets as well as many panel data sets.

https://doi.org/10.1007/978-3-030-73144-1_3
22 3 Consumer Demand …

sum aggregate of goods within the group, i∈G qi (Weissand Sharir 1978). The
pq
appropriate price index would then be the unit value index, i∈G qi i i .
i∈G
Although both forms of the Composite Commodity Theorem (CCT) are used
(particularly for very disaggregated goods), it’s use is quite limited due to the fact that
it requires exact price proportionality (respectively, price difference proportionality).
Lewbel (1996) has relaxed the assumption of the price proportionality version of
the CCT by allowing for deviations in price proportionality so that on average price
proportionality occurs. This method is called the Generalized Composite Commodity
Theorem (GCCT). While this approach shows some promise, it is still somewhat
restrictive in its restrictions on price movements. Moreover, empirical applications
of the GCCT to date are limited2 .
This chapter focuses on the widely applicable approach based on imposing restric-
tions on preferences rather than movements of relative prices within commodity
aggregates. Major separability approaches are reviewed and the implications of
weak separability and two-stage budgeting are discussed. An attempt is made to
make a difficult and complex subject clearer with an emphasis placed on empirical
tractability.
3.3 Separability Concepts
There are two main concepts of separability: weak and strong. Weak separability
states that preferences may be partitioned into N <n partitions such that

u = F v1 q 1 , v2 q 2 , . . . , v N q N (3.1)
N
where q G = q1 , q2 , . . . , qn G for G = 1 n G = n. Weak separability implies the
following:
∂v

∂qi
∂ ∂v
∂qk
= 0 i, k ∈ G; j ∈ H ; H = G (3.2)
∂q j
And
∂qi ∂q j
ki j = μG H i ∈ G; j ∈ H ; H = G (3.3)
∂y ∂y
The last restriction on compensated cross-price effects is both a necessary and

sufficient condition for weak separability (Goldman and Uzawa 1964).
2 Notable applications to date, aside from Lewbel (1996), include Davis et al. (2000) and Reed et al.
(2005).
3.3 Separability Concepts 23
Strong separability states that preferences may be partitioned into N <n partitions
such that
N
u=F vj q j (3.4)
j =1

where q G = q1 , q2 , . . . , qn G for GN = 1 n G = n. Strong separability implies the
following:
∂v

∂qi
∂ ∂v
∂qk
= 0 i ∈ H ; k ∈ I ; j ∈ G; H, I = G (3.5)
∂q j
And
∂qi ∂q j
ki j = μ i ∈ G; j ∈ H ; H = G (3.6)
∂y ∂y
The last restriction on compensated cross-price effects is both a necessary and

sufficient condition for strong separability (Goldman and Uzawa 1964). Also note
that with F a monotonic, increasing function,
N an
equivalent
representation of
preferences for strong separability is u = j =1 v j q j .
Additional concepts of separability are developed and discussed in Blackorby,
Primont, and Russell (1978). For use in this book, we will also make use of
homogeneous separability. Homogeneous separability is weak separability plus the
requirement that each group utility function be homogeneous of degree one.
3.4 Implications of Weak Separability
3.4.1 Theorem 1: Conditional Demand Functions
One implication of weak separability is that we may formulate conditional demand

functions. That is, weak separability implies:
Theorem 1 (Pollack 1970)

qi = giG yG , pG ∀i ∈ G (3.7)

where p G = p1 , p2 , . . . , pn G is the price vector of commodities only in group
G. This system of demand functions can be treated in all respects as a system of
Marshallian Demand functions with all the properties inherent in the theory.
Proof:
Under weak separability, the first-order conditions for utility maximization are:
∂u ∂ F ∂vG
= −λpi = 0 ∀i ∈ G
∂qi ∂vG ∂qi
y= pi qi + pjqj
i∈G j∈H
H =G
Eliminating λ we obtain
∂F ∂vG ∂vG
∂vG ∂qi ∂qi pi
∂F ∂vG
= ∂vG
= ∀i, k ∈ G
∂vG ∂qk ∂qk
pk
y= pi qi + pjqj
i∈G j∈H
H =G
These conditions give us demand functions:
qi = m i y− p j q j , pG ∀i ∈ G
j∈H
H =G
If q j = g j (y,p) ∀ j ∈ H, H = G,

then qi = m i y− j∈H p j q j (y, p), pG = giG yG , pG . QED
H =G
3.4.2 Total Differential of First-Stage Allocation Equations
Another implication of weak separability concerns the total differential of group

expenditure:
N
p G dq G = bG (dy−qd p) + ϕG H dπ H (3.8)
H =1
where p G dq G is the change in real expenditure on group G, bG = p G q yG ,

∂q1G ∂q2G ∂qnGG
q yG = ∂ yG , ∂ yG , . . . , ∂ yG , ϕG H = bG b H μG H , ϕGG = − H =G ϕG H and
∂q jH
dπ H = j∈H ∂ yG dp j is an index of prices associated with group H (Barten
1977). It would appear that this result can be used directly with the total differential
of the conditional demand functions (3.7) to obtain comparative statics of the total
effect of changes in prices on demand. The total differential of (3.7) can be written
as follows:
∂giG ∂giG
dqi = (dyG −qdpG ) + kiGj dpk = ( pG dqG ) + kiGj dp j
∂ yG j∈G ∂ yG j∈G
3.4 Implications of Weak Separability 25
where kiGj is the compensated price effect associated with the Slutsky submatrix for
group G.
Substituting from (3.8) then gives:
∂giG N
dqi = bG (dy−qd p) + ϕG H dπ H + kiGj dp j (3.9)
∂ yG H =1 j∈G
Therefore, we have as comparative effects for total changes in prices and income:
∂giG ∂g j
G
ki j = ϕGG + kiGj ∀i, j ∈ G (3.10)
∂ yG ∂ yG
∂giG ∂g Hj
ki j = ϕG H ∀i ∈ G, j ∈ H, H = G (3.11)
∂ yG ∂ y H
∂qi ∂giG
= bG ∀i ∈ G (3.12)
∂y ∂ yG
Uncompensated price effects can be calculated directly using the Slutsky equation:
∂qi
∂pj
= ki j −q j ∂q
∂y
i
.
Consider the log differential of Eq. (3.8). After dividing both sides by y, it can be
written as (see, e.g., Deaton and Muellbauer 1980b, Chapt. 5):
N
wG dlog ȳG = bG (dlogy - wdlogp) + ϕG* H dlogπ H (3.13)
H =1

where wG = i∈G wi , dlog P = wdlog p = nj = 1 w j dlogp j is the relative change
in real expenditure of group G, w = (w1 , w2 , . . . , wn ), ϕG* H = ϕG H /y, and the two
price indexes are:
n
dlog P = wdlog p = w j d log p j (3.14)
j =1
dlogπ H = b Hj dlogp j (3.15)

j∈H
∂g H
where b Hj = p j ∂ yiH is the marginal expenditure value of the jth good from changing
∂g j
expenditure on group H and bG = j∈G p j ∂ y is the marginal expenditure value of
group G. In general, we do not have numerical estimates of these price indices because
they depend on the unknown functions b Hj One might use conditional expenditure
shares, wiG = pyiGqi , in place of the conditional marginal expenditure values in (3.15).
Barten and Turnovsky (1966) argue that such an approximation need only have minor
effects on the estimated parameters.
3.5 Two-Stage Budgeting
3.5.1 Theorem 2 and Consistency Requirement
What one would like to do, of course, is use price indices such as the widely published
Laspeyres or Paasche type indices. An important question then is: under what condi-
tion will two-stage budgeting yield the same solution as single-stage budgeting? That
is, when will the following conditions occur?

qi = giG yG , pG = giG yG y, P 1 , P 2 , . . . , P N , pG = gi (y, p) (3.16)
where P G , G = 1, 2, . . . , N are price indices. This condition is called the

Consistency
Requirement for aggregation. For local price indices of the form
d P H = i∈H aiH (y, p)dpi to exist, it must be true that
∂qi ∂giG ∂ yG ∂ P H ∂PH

∂ ps ∂ yG ∂ P H ∂ ps ∂ ps
∂qi
= =
∂giG ∂ yG ∂ P H ∂PH
∂ pt ∂ yG ∂ P H ∂ pt ∂ pt
is independent of G (Strotz 1957, Gorman 1959). The following theorem gives the
necessary and sufficient conditions for aggregation (Gorman, 1959 Green 1964):
Theorem 2
The Consistency Requirement (3.16) holds if and only if the utility function is
weakly separable and:
(i) there are only two groups; or
(ii) each group utility function is homogeneous of degree one; or
(iii) the quantity indices for all groups but one (say the first) are homoge-
neous and the utility function may be written as a function of the one non-
homogeneous group and a function of the remaining homogeneous functions,
H (vd + 1 , vd + 2 , . . . , v N ); or
(iv) the quantity indices beyond the dth are homogeneous and the utility function
d
may be written as u = F j = 1 v j q j , H (vd + 1 , vd + 2 , . . . , v N ) .
Proof:
The proof is long and difficult to reproduce all the details. I will outline the
significant aspects of it. See Strotz (1957), Gorman (1959), and Green (1964, Chapt.
3) for additional details.
Consider the first-stage problem which for simplicity, but still generality, can be
written as follows:

max u = v QG , Q H , Q I
{Q G ,Q H ,Q I }
subject to:
3.5 Two-Stage Budgeting 27
y = yG + y H + y I
This maximization problem is equivalent to the problem:
max u = V (ϕG , ϕ H , ϕ I )
{yG ,y H ,y I }
subject to:
y = yG + y H + y I
where ϕG = maxqG { vG (qG )| pG qG = yG }. The first-order conditions for utility

maximization are:
VG ϕG −VI ϕ I = 0
VH ϕ H −VI ϕ I = 0
y = yG + y H + y I
where VG = ∂ϕ ∂V
G
, ϕG = ∂ϕ G
∂ yG
, etc. Solutions to these first-order conditions are:
yG = yG ( ps , pt ), y H = y H ( ps , pt ), y I = y I ( ps , pt ) ∀s, t ∈ I ; I = G, I = H
where I have suppressed notation for all other prices and income. These solu-
tions are substituted into the above first-order conditions and then differentiated with
respect to both ps and pt . The algebra is tedious, but one can show that:

ϕ I s ϕ I t
− (VG I ϕG −VH I ϕ H ) = 0
ϕI s ϕI t

So either (a) ϕ Is
ϕI s
− ϕ It
ϕI t
= 0 or
(b) (VG I ϕG −VH I ϕ H ) = 0.
∂ϕ I
∂ ∂ yI
For case (a), ϕ I s = ∂ ps
,

∂ϕ I ∂ϕ I
∂ ∂ yI ∂ ∂ yI
∂ϕ I ∂ϕ I
ϕI s = ∂ ps
, etc., which implies ∂ ps
∂ϕ I = ∂ pt
∂ϕ I . Let λ I = ∂ yI
so that condition (a)
∂ ps ∂ pt
∂λ I ∂λ I
∂ϕ I
can be written as ∂ ps
∂ϕ I = ∂ pt
∂ϕ I . By Roy’s identity (2.34), ∂ ps
= −qs λ I and likewise for
∂ ps ∂ pt
∂λ I
∂ϕ I
∂ pt
. Substituting in the above expression and rearranging terms gives: qqst = ∂ ps
∂λ I . From
∂ pt
Eq. (2.26), we infer that ∂λ I

∂ ps
= λ I ∂∂ pyIs + ∂λ I
∂ yI
qs ; similarly for ∂λ I
∂ pt
. Upon substituting for
∂qs
qs ∂ yI
the numerator and denominator above and rearranging terms: qt
= ∂qt . Therefore,
∂ yI
we find that condition (a) is that all income- consumption curves must be straight
lines through the origin. This means preferences for each group must be homothetic,
which in turn, implies homogeneous preferences.
Condition (b) can be shown to mean preferences are strongly separable.
∂u
∂vG
∂ ∂u
∂v H
Strong separability means that: ∂v I = 0. Carrying out the differentiation we
find: VG I VHV−V
2
H I VG
= 0, which implies VG I VH −VH I VG = 0. From the first-order
H
conditions,
VG ϕG −VH ϕ H = 0, or VVGH = ϕ H
ϕG
. Substituting into the equation,
VG I VH −VH I VG = 0, we obtain: (VG I ϕG −VH I ϕ H ) = 0, which is precisely
condition (b). QED Homogeneous separability is the least attractive of the two
requirements. In particular, because homogeneous separability implies all condi-
tional demand functions have unitary income elasticities, it is usually rejected
empirically as too restrictive3 . Typically, the least restrictive condition is strong
separability, which has formed the basis of much empirical work.
3.5.2 Form of Price Indices
Theorem 2gives the consistent criteria for local price indices. For global price indices,
P H = P H p H , the conditions are more stringent. As proved by Gorman (1959),
global price indices exist if there is strong separability plus the requirement that each
group have the Gorman Polar Form:

C G u G , p G = aG p G + u G bG p G (3.17)
This functional form implies all Engel curves are linear, but does not require the
curves to go through the origin as in homogeneous preferences.
Indeed, homogeneous
preferences are a special case of (3.17) when aG p G = 0. However, the quasi-linear
nature of (3.17) leads to quite stringent conditions on conditional demand functions.
Sometimes one would like price and quantity indices such that the product of the
two equals total expenditure on the group in question. If one requires that the price
indices depend only on prices within the group and the quantity indices depend only
on quantities within the group, then group preferences must be homogeneous (Green
1964, Chapt. 4). Fortunately, for two-stage budgeting only existence of local price
indices is required, leading us back to Theorem 2.
∂qs
3 Note from part (a) of Theorem 2 that the condition qs ∂ yI
qt = ∂qt is equivalent to assuming each second
∂ yI
stage income elasticity equals one.
3.6 Implications of Strong Separability 29
3.6 Implications of Strong Separability
3.6.1 General Form of Composite Demand Function
As indicated above, strong separability is the least restrictive of the requirements to

attain the Consistency Requirement for two-stage budgeting. Theil (1980, Chapt. 2)
shows that the composite demand function for the first-stage utility maximization
problem with strong separability can be written as
wG dlog Q G = bG (dlogy−wdlog p) + φbG (dlogπG −dlogπ) (3.18)

where Q G = i∈G wwGi dlogqi , dlogπG as defined in (3.15), φ is the “money flex-
n
ibility” as defined in Chapter
2, and dlogπ = H b H dlogπ H = j = 1 b j dlogp j .
The functions bG = j∈G b j are group marginal expenditure values, with
∂g
b j = p j ∂ yj . The empirical counterpart of this specification is called the “relative price
version of the Rotterdam Model” because the price index for group G is deflated by
the general price index, π , known as the “Frisch price index.”
In Chapter 2, we showed that substitution effects could be decomposed into
specific and general effects. The total differential of demand in the case of strong
separability can be written as
n
wi dlogqi = bi (dlogy−wdlog p) + vi j −φbi b j dlogp j
j =1
where, using (2.30), we have that ci j = wi ei∗j = vi j −φbi b j , where vi j = wi ei j , and

f

wi ei = bi so that the last term becomes φwi ei w j e j = φbi b j . Because nj = 1 b j = 1
n
from adding-up, j = 1 vi j = φbi . Therefore, the above log differential can be
expressed as
n n n
wi dlogqi = bi (dlogy−wdlog p) + vi j dlogp j − vi j bk dlogpk
j =1 j =1 k=1
n
= bi (dlogy−wdlog p) + k = 1 vik dlogpk − nj = 1 b j dlogp j or
n
wi dlogqi = bi (dlogy−wdlog p) + vik (dlogpk −dlogπ ) (3.19)
k=1
n
Noting that for block additivity, i = 1 vi j = i∈G vi j and vik = vki , upon
summing over both sides we get
n
wi dlogqi = bG (dlogy−wdlog p) + vki (dlogpk −dlogπ )
i∈G i∈G k=1

= bG (dlogy−wdlog p) + i∈G φbi (dlogpk −dlogπ) or
wi dlogqi = bG (dlogy−wdlog p) + φbG (dlogpk −dlogπ)

i∈G

Multiplying and dividing i∈G wi dlogqi by wG then gives (3.18).
3.6.2 Unconditional Demand Elasticities
The first-stage composite demand function (3.18) and the conditional demand func-
tions can be combined to produce unconditional demand responses as in the case
of weak separability shown above. The total differential of the conditional demand
functions can be written as

wi dlogqi = biG wG dlog Q G + G
cik dlogpk (3.20)
i∈G
where ciGj = wi kiGj . The unconditional demand responses under strong separability
are obtained upon substituting (3.18) into (3.20):

wi dlogqi = biG bG (dlogy−wdlog p) + φbG (dlogπG −dlogπ ) + G
cik dlogpk
j∈G
(3.21)
The unconditional demand elasticities under strong separability are therefore
φbiG b Gj bG (1−bG ) + ciGj

ei∗j = ∀i, j ∈ G (3.22)
wi
−φbiG b Hj bG b H
ei∗j = ∀i ∈ G, j ∈ H, H = G (3.23)
wi
biG bG
ei = (3.24)
wi
These comparative statics, in contrast to those in (3.10), (3.11) and (3.12), indicate
additional information on the money flexibility, φ, and group expenditure values, bG ,
is required, in addition to the conditional demand parameters, to estimate uncondi-
tional elasticities. To see this more clearly, we can express Eqs. (3.10), (3.11) and
(3.12) in elasticities as follows:
biG b Gj ϕGG
*
+ ciGj
ei∗j = ∀i, j ∈ G (3.25)
wi
biG b Hj ϕG* H
ei∗j = ∀i ∈ G, j ∈ H, H = G (3.26)
wi
Another random document with
no related content on Scribd:
spoke anything rather than brotherly love. Robin kept a respectful distance,
for he seemed to have for the Saxon youth no stronger predilection than his
master.
Meantime the mind of the rebel chief was little disturbed by the
disastrous intelligence which he had received. The leader, upon whom his
party had placed the greatest reliance, was taken; and the easy defeat of the
Flemings had taught him a lesson which every one that embarks in a great
undertaking should learn betimes—that it is a perilous thing to put trust in
foreign auxiliaries. Yet so accustomed was he to this irregular mode of
warfare, and so inured to all the vicissitudes to which the fickle temper of
Dame Fortune might subject him, that his mind was at this moment
perfectly calm, and hardly rested a thought upon the perilous situation in
which he found himself placed. He seated himself at the rude banquet,
which his followers were now preparing, with perfect indifference, although
the possibility of his enjoying another tranquil meal was at least a matter of
doubt. After some time spent in noisy revelling—for when their assistance
was required in an affair of so much danger, the chief thought it no scorn to
join in the merriment and court the goodwill of his vassals—Richard began
to reflect upon his interview with his two captives; and, with a
contemptuous smile, he asked who was the Saxon divinity to whom they
must attribute the loss of so able a coadjutor in the person of Reginald
d’Arennes.
A dozen sturdy voices were lifted up at once, in commendation of the
Lady Elfrida. Her tall and commanding stature—her long flaxen hair—her
dignified countenance—her cheeks, whose bright complexion invited the
flattery which they blushed to hear—and her light blue eye, whose glance
beamed so mildly on the meek, and met so proudly the gazes of the proud—
were alternately the themes of admiration. At last the chieftain, impatient of
these rapturous effusions, which he began to think were endless, poured out
his last cup “To the health of the Rose of Kennet Hold,” and deserted the
board. He busied himself for a time in giving the necessary orders for their
departure early in the ensuing morning; and then, calling Eustace aside,
exclaimed: “We will ourselves look upon this Saxon beauty: by our Lady, if
she deserve but one half of the praises of these boors, she may haply be the
companion of our onward march.” And with these words, attended by his
esquire, De Mallory strode from the enclosure.
While this scene was going on, Reginald and his companion had made
considerable progress on their journey, and were within a few miles of its
termination; yet not a word had been exchanged between them. They
looked from time to time towards each other, apparently with a mutual
feeling of dislike, if not of apprehension. At last Lothaire led the way to
conversation, in a tone which betrayed a strong disposition to offer an
insult, although the disposition appeared to be checked or subdued for a
time by the counter-agency of some equally powerful motive.
“Sir Reginald,” said he, “knowest thou the qualities which are required
in him who would sue for the hand of my sister Elfrida?”
“I have doubts touching this matter,” replied Reginald.
“Methinks,” rejoined his companion, “it were worth the while to instruct
thyself further, ere thou settest foot on my father’s threshold; for, of a truth,
Elfrida hath a right Saxon spirit and a right Saxon speech: she hath proud
eyes, that smile on whom they list and frown on whom they will; and proud
thoughts, that respect not so much the glittering of the corselet as the valour
of the knight that wears it.”
This was somewhat like a thunderclap to poor Reginald. He had
anticipated no difficulties of this nature: the timidity of his nature would
have shrunk back with horror from the mention of a protracted courtship. In
short, he had expected a path strewn with roses, and he found it beset with
briars; he came to wed an obedient and passive bride, and he began to
suspect she was little better than an intractable virago. After having spent
some moments in reflections of this nature, he gave utterance to his secret
musings in a brief soliloquy:
“I am doubting whether or no I shall proceed.”
He was answered by a loud laugh from his intended brother-in-law, who
proceeded forthwith to dispel the apprehensions which he had himself
excited.
“Cheer thee, noble Knight; be not afraid for a woman. Thou hast,
princely Reginald, many valorous and knightly qualities, the least of which
might win a richer bride than the daughter of Leofwyn and the sister of
Lothaire. Surely thou dost obtain honour at those splendid jousts, from
which thou knowest our Saxon habits do utterly revolt; and, doubtless, thou
hast skill in foreign music, which thou knowest our Saxon ears do utterly
detest; and thou art also skilled in that foreign language which thou knowest
a Saxon doth so loath, that he would have his tongue torn from his throat
rather than give utterance to its accents.”
“Brother,” said Reginald, who began to perceive the necessity of
conciliating Lothaire, “I have meddled but little with courts, and, in my
ignorance of these accomplishments, I am a perfect Saxon. But I prithee tell
me, in love and fellowship, by what means or endeavours it is possible for
me to win the goodwill of thy sister.”
“I will show thee,” said Lothaire. “First, thou must learn to speak, not
tardily through thy teeth, as is thy present method, but boldly, openly, and
fearlessly, as one man should do to another.”
“Whether this be possible, I doubt,” observed Reginald.
“Secondly,” said his instructor, “at my father’s board thou must not be
too ready to relinquish the goblet.”
“I will do thee reason—I will do thee reason, Sir Lothaire,” returned
Reginald. “Marry, I shall need but little instruction upon that head.” And he
strained his eye as he spoke in the direction of Kennet Hold, as if he would
measure the space which lay between his lip and the flagon.
“Thirdly,” resumed Lothaire, “thou must hate a Norman as thou
wouldest hate the foul fiend.”
“I do,” cried Reginald; “I do hate a Norman: the Norman we parted from
e’en now, Richard de Mallory. A blight upon him! He hath bound me,
scoffed at me, worried my body and my mind, until I can scarcely keep my
saddle on my journey or recollect whither the journey tendeth. A murrain
on the proud knight! Doth he fancy that I care aught whether the father or
the son hath the better? whether the Henry I serve be called the second or
the third?”
“If I may risk prophecy,” muttered the Saxon, “thou wilt never see the
third Henry wearing his father’s crown. We have worn the yoke of your
tyrants long enough; and it is time that the throne of Alfred should be again
filled by one of his descendants. Despised and oppressed as we are, there
are still true Saxons enow to drive ye headlong from the land ye have
spoiled.”
The two young men continued to ride as far apart as courtesy and their
roads would permit, and the line of conversation into which they had fallen
did not seem likely to promote kinder feelings between them. Reginald’s
national prejudices began to rise high within him, and to overpower the
want of energy which was his failing. “Sir Lothaire,” he replied doggedly,
“methinks thou hast forgotten Hastings.”
“Sir Knight,” said his companion, in a melancholy voice, “it is not
possible for thee or for me to forget Hastings. Thine ancestor did obtain
there power, and title, and riches; mine did win nothing but honour and his
grave. The chance may be ours in another field. If valour and desert in arms
had had their meed the bastard of Normandy had never set foot upon the
corpse of Harold.”
“Thou errest, thou errest, good brother,” said Reginald unthinkingly;
“the single arm of King William was sufficient to beat down Harold and his
brothers to boot. Thine ancestor himself, Sir Lothaire, was light in the
balance when weighed with the least of our Norman chivalry!”
“Norman liar!” exclaimed Lothaire, and immediately giving his horse
the spur, and causing him to make a demivolte, which brought him close to
his companion’s side, he raised his ponderous arm, and dealt with his
mailed hand so terrible a blow between the corselet and headpiece of his
future brother-in-law, that Launcelot reeled upon his haunches, and his rider
fell to the ground without sense or motion. Lothaire gazed for a moment
upon the fallen Knight; and then, after beckoning to Robin to come up, put
his horse into a hand gallop, and continued his route.
Robin, when the formidable Saxon was out of sight, ventured to
approach the scene of the fracas. Piteous was the sight which presented
itself. Launcelot was standing beneath a neighbouring tree, still trembling
with the shock he had received. Reginald lay motionless in the dust: his
bright armour was soiled with earth and blood, which gushed out plentifully
from his mouth and nostrils. Robin took off his helmet, and endeavoured,
by throwing water over his features, to restore animation. After having
spent a long time in the vain endeavour, he looked upon his fallen patron
with an expression of utter despair, and muttered to himself, “My master is
certainly dead; and there will be no wedding, nor revel, nor wassailing.” He
continued for some minutes in deep contemplation, and then exclaimed,
“An my project hold good, I will be revenged on the Saxon churl.” And
with these words he began to disarm his master.
While these incidents were taking place among those personages to
whom our attention has been hitherto confined, the state of the inhabitants
of Kennet Hold was such as calls for our notice. The MS., indeed, from
which I draw this narrative, goes through all the minute particulars of
Reginald’s journey, until it sets him down at the gate of his father-in-law;
but, to avoid greater prolixity than is necessary, I will reserve this
explanation for my dénouement, and for the present leaving my hero on his
bed of earth, I will introduce my reader, without further delay, to the hall at
Kennet Hold.
Everything seemed to be in a state of unusual confusion at the residence
of the Saxon. This was, no doubt, partly to be attributed to the extraordinary
preparations made by the cooks, and to the wish of the domestics to appear
in the sprucest attire before the eyes of the Norman guest. But there was
something more than this in the bustle which pervaded Kennet Hold. There
seemed to be in every countenance, from the swineherd to the thane, the
consciousness of some concealment, some unspeakable secret lingering on
the lips, and awaiting a fit opportunity for disclosure. Many of the menials
were staring at each other in silence, although they had abundant
occupation before them; and many were looking inordinately busy,
although it was their chance to have nothing to do. The expression of their
faces was various. In some you could perceive little more than a repressed
desire to laugh; but on the features of the higher sort of vassals you might
read pride, contempt, resentment, together with a visible exultation, which
plainly told that all these vindictive feelings were on the eve of
gratification.
Leofwyn himself was seated on the chair of his hall, beneath a scarlet
canopy, in all the rude state which his Saxon prejudices permitted. He was
of short stature, with a round good-humoured face, which spoke, as plainly
as face could speak, that its owner was willing to be upon friendly terms
with the rest of the world, if the rest of the world would give him leave. In
fact, Leofwyn was of a disposition to prefer the beginning of a banquet to
the conclusion of a broil; and if he had been at liberty to consult his own
inclination, there would have been much wine, and but very little blood,
poured out annually by the retainers of Kennet Hold. Many causes,
however, conspired to make these pacific qualities of no effect. In the first
place, the chief had an hereditary feud to support against the invaders of the
land; and, although he himself saw nothing in these foreigners which should
deserve his malediction, he deemed it his duty to hate them most
religiously, because his father had done so before him; secondly, his son
Lothaire was of a terribly violent temper, and was always seeking an
opportunity for embroiling his father with some Norman landholder; and
thirdly, this opportunity was frequently afforded by the predatory attacks of
the surrounding nobles.
In the retaliation which Leofwyn exercised for these outrages he
frequently put in practice some cunning and jocose device, which accorded
ill with the professions of hate and enmity which he was perpetually
making. For instance, it appears that when the vassals of Sir Robert de
Vallice had made considerable depredations upon the Saxon’s swine, he
carried off the only son of the offender, and, after confining him in
company with the porkers for a night and a day, sent him back to Sir
Robert, with a message that “he had sent him his swineherd also.” Such
freaks as these had among his dependents secured to him the reputation of
having a right sharp wit: among his powerful neighbours he was considered
little better than a madman, in consequence of which, amidst the
oppressions to which his race was daily subject, he had been allowed to
pass his days in despised security.
Upon the present occasion it seemed that he had some unusually clever
design in view. He was perpetually giving some instructions to the
domestics, in a tone of voice mysteriously low, and again relapsing into
deep and silent meditation. In short, in the anxiety which he evinced for the
approaching nuptials, he showed all the assiduity and precaution of a
modern match maker. Reginald did not come at the appointed time; the old
man began to grow impatient; he asked for his son. “Lothaire,” replied one
of the attendants, “bore forth thy message in the morning, being desirous of
looking on the Norman guest. He hath not yet returned.” “It is the better,”
said Leofwyn to himself. “His hastiness might defeat what my prudence
hath devised. Nevertheless, I cannot but marvel at his stay. Is the bride
apparelled?” “She is: the maidens have been busied about her head-gear
since noonday. Marry they have no light task; for the hair they decorate
hath been but little used to the operation.” “Peace!” said Leofwyn.
Hours passed away in rapid succession, evening came gradually on; and
still there were no traces or tidings of Reginald d’Arennes. The Saxon’s
choler began to rise in earnest. “Surely,” he muttered inwardly, “surely, that
hot-brained fellow Lothaire hath not overturned the structure my counsel
hath been so long a-building; mischief light upon him if he hath dared to
make or to meddle! The forward boy is ever at bullying and drawing of
swords. Boys’ play, boys’ play; but it were a brave thing to put this slight
upon the Norman. Marry, hang him if he hath despoiled my daughter of her
husband.”
Suddenly his soliloquy was interrupted by the blast of horn announcing
the arrival of strangers. Leofwyn leaped from his seat in an ecstasy; but
immediately resumed it, with a studied look of gravity, that restrained the
inclination of merriment which was predominant among his dependents.
Every one, therefore, was silent, as the folding-doors were slowly unclosed,
and the major-domo introduced to the presence of his lord—Sir Reginald
d’Arennes.
He was greeted by his future father-in-law with cold and distant
courtesy, which he returned in a manner of still greater reserve. “Sir
Knight,” said Leofwyn, “it is my will that thy nuptials be solemnized ere
thou sittest down to the banquet. My son Lothaire is choleric (his guest
gave an involuntary motion of assent); and if he should return before the
wedding, I know not whether thy head might not lie in the castle-moat
sooner than on the bridal pillow.” The bridegroom shuddered.
“Is the Lady Elfrida attired?” continued Leofwyn, in a tone of mock
gravity, which was exceedingly ludicrous. The attendants caught the
infection, and many unrepressed jests circulated among them, as they
departed to bear their lord’s summons to his daughter.
Presently Elfrida made her appearance. The bridegroom started as she
entered the hall: perhaps the exterior qualifications of the Saxon beauty
might not altogether correspond with the exaggerated reports which his ears
had greedily drunk. Her figure might be called elegant, but was certainly
too short to deserve the appellation of dignified; her face might be deemed
pretty, but the pertness which was its prevalent characteristic disqualified it
for the epithet of beautiful. Instead of the soft yellow hair which her adorer
had expected, he beheld a profusion of dark brown ringlets; and in lieu of
the languishing blue eye, which he deemed would have dissolved him into
rapture, he met the glance of a sparkling black one, in which there lurked a
very strong inclination to laugh in his face. To his disappointment, however,
if he felt any, Reginald gave no vent; he seemed to have a great reluctance
to unclose, in the presence of Saxons, either his visor or his lips. Both
parties betrayed a wish to have the ceremony performed as speedily as
possible; and the nuptials of Sir Reginald d’Arennes with Elfrida, the
daughter of Leofwyn of Kennet Hold, were accordingly celebrated in the
chapel which was attached to the residence of the Saxon. The Lady Elfrida
was splendidly attired; but, in other respects, the nuptial rites were graced
with little pomp save the attendance of a large body of Leofwyn’s retainers,
who, bearing in their hands each a flaming torch, cast an air of rude
magnificence over the scene.
A sumptuous banquet awaited them upon their return to the hall. The
merriment of the vassals was loud and unremitting. The bridegroom,
however, did not seem to enjoy the situation in which he found himself
placed. He fidgeted upon his seat, and turned his eyes alternately to the
ceiling and to the wall, as if he suspected that more than half the joviality of
the party was at his expense. His embarrassment was increased by the
malicious endeavours of his bride, who rallied him upon his gravity and
look of despondency, in a style to which he had evidently no spirits to reply.
It must be confessed that the young man’s suspicions were not altogether
without foundation. The occupants of the lower part of the board, who, of
course, were the most obstreperous in their mirth, were, from time to time,
indulging themselves in very acute criticisms upon the figure and features
of their master’s son-in-law. These did not altogether answer their
expectations. Much as they contemned the Normans, they had pictured to
themselves, in the person of Reginald d’Arennes, a countenance noble even
to sternness, and a bearing at once courteous and martial. They knew he
was a Norman, but they also knew he was a handsome and a friendly
Norman; in consequence of which they had made up their minds to hate
him, and, at the same time, to find nothing in him worthy of hate. They
were much surprised, therefore, when they found the young Knight so
perfectly different from the image report had drawn. His face seemed
perfectly void of all expression of majesty or valour. At present its
predominant expression was embarrassment, mingled with a strong tincture
of fear; but there was a slight curve upon the lips, and a sly twinkle under
the eye, which betrayed a strong disposition to cunning and risibility. His
figure appeared slender and diminutive, and a gorgeous steel harness hung
dangling about it, as if the bark of the forest oak had been stripped off to
give an appearance of strength to the willow. This was all very strange: the
attendants looked, and laughed, and wondered; and Leofwyn showed no
disposition to check their humour. Indeed he seemed to participate cordially
in their malicious propensities.
“Sir Knight,” said he, “methinks there is in thy demeanour a greater
degree of bashfulness than thy noble presence and thy lofty lineage do
warrant.”
“It is a feeling,” replied the guest, “which I have inherited from my
mother Bridget—I mean, from my mother the Lady Marie,” he added,
turning very pale.
“Ha, ha!” exclaimed his entertainer; “now, by my verity, I dreamed not
thy father had been so gay in his young days. What! play the Lady Marie
false! Come, come, it was ill done, ill done; she was a lady of most
excellent carriage; it was ill done. But be not cast down. The sin was not
thine. Pledge me, noble Reginald. Thou standest in need of refreshment; for,
in truth, thy look is weariness itself, and thou art as silent as the oaken
board on which thou leanest. Come, come, the pigment is worth the
tasting.”
Reginald blushed, and seemed doubting whether it were not well to
make a precipitate retreat. The Lady Elfrida turned away her head, and let
down her veil, with a gesture of affected horror at the indelicate sallies of
her father. Nothing daunted, the old man continued his pertinacious system
of annoyance, while the domestics applauded, by ill-repressed
acclamations, the surprising jocularity of their lord.
“Thou art sparing of thy food, Sir Knight; but doubtless thou art used to
other diet than this: the board of a Saxon thane hath but little to tempt the
palate of the son of a Norman noble.”
“Thou wrongest thine own hospitality, noble thane,” replied the other,
collecting his spirits, and making an effort to be polite. “Womanly indeed
should I be, if I were not used to harder fare than this! My father, the
forester—that is, I mean, my father, the baron.” And again Reginald looked
confused, and paused, and was silent.
“Cheer thee, noble Reginald,” said his host; “thou art wearied with thy
journey, and thy wits wander.” “Perchance,” said the fair Elfrida, “Sir
Reginald hath lost them on the way!” The menials echoed applause, and
Reginald looked yet more foolish than before. “Thou dost belie thy
character strangely,” continued the old man; “fame hath told us that in the
whole shire there is not a jollier boon companion nor a truer lover of the
cup.” “It is true that Sir Reginald d’Arennes hath had that reputation,”
replied the Norman, “and his best friends have judged that he would do well
to put it away.” “By the holy Confessor,” cried Leofwyn, “not upon his
wedding-day! Out upon the idea! What, ho! Osric, fill up for Sir Reginald.
Pledge me, gallant Knight. The health of thy bride—of Elfrida!”
“I will do thee reason,” said Reginald, raising the cup to his lips; but, at
the mention of the name of Elfrida, some of the vassals burst into such a
clamorous fit of laughter that he set it down in astonishment.
Leofwyn remarked his surprise, and endeavoured to dispel it. “Thou
seest, good son, that there is a kind of pageant toward, at which these boors
are marvellously pleased; but be not the less inclined to join in our banquet.
We wait but for the arrival of my son Lothaire, and all disguise shall be
stripped off.” “Disguise!” cried the guest, dropping the cup, and starting
from his seat, “a murrain on the tell-tale! How didst thou learn——” “Nay,
my son,” said the Saxon, as if endeavouring to retract an unguarded
expression, “we are all somewhat disguised—in liquor.”
Reginald resumed his seat, and, in a short time, began to drink most
valorously, as if striving to drown in the rich pigment some unpleasant
suspicions. By degrees, his head, which was evidently weaker than the one
fame had attributed to Reginald d’Arennes, began to be overpowered by the
frequent potations which were forced upon him by his host; and while
Leofwyn and his retainers, and even the modest Elfrida, were immensely
amused by his awkward situation, the hapless bridegroom showed the
effects of Saxon hospitality in rhapsodical and unintelligable exclamations.
“Of a truth, good thane, thy drink is marvellous good! marvellous good
is thy drink! Better have not I tasted since we rifled old Ambrose, the
hermit of Torney Low! Very rich was the old rogue: he had store of gold
and of silver, and an admirable cellar withal. Right merry we were and
jovial; and, for the hoary man, we made him sit by the board, and chaunt a
merry stave. That did I; for truly my fellow thief had some quirks of
conscience. Health to the old man! May his bags and his cellar be
replenished before next Whitsuntide! What care I for abbot and friar, mitre
and cowl! I roam through glade and greenwood, over hill, and rock, and
stream, free as the hawk, free as the passing wind. Marry, I had forgot how I
have linked myself to a wife! Kiss me, fair Elfrida! I love thee very much,
Elfrida; but thou knowest, when war calleth us away, we soldiers leave ye
like a whistle. How dost thou, old father-in-law, how dost thou? Of a verity,
thy face is as black as a November cloud, and that spear by thy side is
wondrous sharp: it is well I have a Milan corselet. Mark ye my Milan
corselet, father and bride? The zecchins that were paid for it! It hath not
borne blow yet. Certainly I like not blows; but the lace of my helmet is
snapped in twain. Thy son, most noble Leofwyn, could explain unto thee
the manner of it. Surely it was a mighty blow, and a perilous, given with a
strong arm and a right goodwill. Launcelot shook like an aspen leaf.
Howbeit, noble Saxon, thy drink is marvellous good; it maketh a man
valorous, and doth as it were put to flight the whimsies, and the visions, and
the phantasies of the brain. Fill up, valiant Leofwyn! Plague on them that
flinch! Mine harness is much soiled for a wedding garment, but I shall wear
a new doublet to-morrow. A blight upon the brambles in the coppice! How
now, good father-in-law, why dost thou not speak? Thy face is as round as
the bowl, and as silent as the roasted crab that is floating within it. Fill up!
Off with care! Shall I not be merry, when steel, and nobility, and a wife are
put upon my shoulders?”
“My lord groweth complimentary,” said Elfrida, hardly able to speak for
laughter. “I do feel afraid that the air of Kennet Hold, and the drink it
affords, have somewhat unsettled his brains!”
“Beautiful Elfrida,” said the bridegroom, “true it is that the brains of Sir
Reginald had a terrible knock this day, and thy brother knows whence it
came; but we will forget these quarrelsome topics, and give up the evening
to merriment. My brains are as firm as thine own. Marry, the wine is
marvellous good!” He was sinking gradually into intoxication.
“I marvel wherefore Lothaire delayeth his coming,” said Leofwyn.
“Truly,” replied Elfrida, “it were well to conclude the farce without him.
I am weary of this mummery.”
“Mar-vel-lous good!” repeated the Norman, and closed his eyes.
“Girl,” said Leofwyn, “thou speakest foolishly; until my son’s return we
will keep up the disguise.”
“Disguise!” cried Reginald, recovering some little sense of what was
going forward. “Who talked of disguise? Was it thou, most rustic Leofwyn,
or thou, most black-browed Elfrida? Who talked of disguise? I care not. If I
am not——”
A loud and piercing shriek interrupted the speaker. You might have
thought all the maidens of the shire had conspired to deafen the ears of the
Saxon proprietor. A door was suddenly flung open, and a warder, with
terror and consternation pictured on every limb and feature, rushed up to the
daïs, and bending his head as if to receive the chastisement which his
negligence would call down, exclaimed, “The Lady Elfrida hath been taken
away from the castle!”
It were difficult to describe minutely the astonishment which pervaded
the hall. Vassals and menials of every degree snatched their arms and fled
from the apartment.
Nothing was heard but inquiries, and weeping, and imprecations.
Nothing was known but that the lady had been within the last few minutes
carried off by a strange knight mounted on a swift bay horse, and attended
by one follower. It was supposed that he must have entered and departed by
swimming the moat, which, as it was now midnight, was an attempt by no
means impracticable. He had been seen by a peasant who was returning
from an adjacent forest; his lovely prize was thrown across a led palfrey,
and appeared to be in a swoon.
All was confusion. The retainers of Leofwyn ran to and fro in all
directions but the right one. Armour resounded with a dismal clang, as it
was hastily thrown over the shoulders of the domestics; torches were
flinging their red glare in every direction; the voices of the pursuers were
repeated by frequent echoes, as they shouted and called to one another
through the darkness. In the meantime the chief personages in the hall were
in a situation partaking strongly of the ludicrous. The black-eyed damsel,
who had figured throughout the banquet as the daughter of Leofwyn, had
cried out, as the warder had delivered his news, “My dear mistress, my poor
mistress!” and fainted upon her throne. The bridegroom had been in some
measure roused from his intoxication, but was still unable to collect his
ideas, so as to form any idea of the origin or meaning of the tumult.
Leofwyn appeared to be in a state of mental stupefaction. In spite of the
foibles of the old man’s character, he was doatingly fond of his daughter;
and the news of her loss, coming in the midst of revelry, seemed to have
withered him like a thunder-bolt. He sat still, looking on the confusion with
a vacant gaze, and inquiring from time to time, “Is my daughter well? How
fares it with the Lady Elfrida? Does she not come to her old father?” These
three personages, therefore, remained quietly upon their seats, while every
one around them was in commotion; like the bronzed images in modern
halls, that hold their candelabras so calmly, while the guests are all in the
bustle of departure.
Things remained in this disagreeable position for some minutes, when
the blowing of a horn, and a loud talking and shouting without, announced
that something had taken place. Presently, accompanied by a crowd of
peasants half accoutred for the pursuit, Lothaire entered the hall. Leofwyn
raised his head, and being in some measure recalled to his recollection by
the sight of his son, repeated his inquiry, “Is my daughter well?”
“She is well!” said Lothaire, “and I am well! No thanks to my new
friend, the doughty Sir Richard de Mallory, from whom, to say truth, mine
headpiece hath received a most mischievous contusion. Thanks to thee,
good steel,” he continued, taking off his helmet, and surveying the deep
indenture which appeared on its summit; “had not thy temper been true, thy
master’s head had lain on the couch from which no man lifteth himself up.”
He was interrupted by a thousand interrogatories, a great proportion of
which proceeded from Leofwyn, who had by this time recovered from the
effects of his sudden shock, and began to feel great curiosity to know the
particulars of the story.
“I know but little of the matter,” said Lothaire; “ye see I have been
overthrown in no light fashion”—(they perceived for the first time that his
apparel bore marks of a recent fall)—“and in truth had it not been for the
intervention of my good friend in the ragged doublet, I had hardly lived to
tell ye the tale.”
“Of whom dost thou speak?” said Leofwyn.
“That is more than I can tell,” replied the young Saxon. “Not many paces
hence did I encounter the valorous Sir Richard, who is now, peace be with
him, no longer a man of this world. I had a heavy stroke, as ye may witness;
nevertheless, it was my horse’s fault, or I had not been so foiled. I believe
another minute would have caught the last breath of Lothaire, but for the
help of the aforesaid knight of the ragged doublet. By the sword of Harold!
he overthrew that proud Norman as if he were wrestling with a child. I saw
not his features, but by his apparel he seemed to be the esquire of thine
hopeful son-in-law, Reginald d’Arennes. But ye will see him presently.”
Lothaire was supported from the hall, and put under the care of the
leech; for his wound, although he made so light of it in his story, wore a
dangerous appearance.
As he retired, another loud acclamation announced the arrival of
Elfrida’s deliverer. A tall, well-made figure advanced towards the daïs, clad,
as Lothaire had intimated, in a short ragged doublet, with a small cap which
was quite insufficient to confine the long dark tresses that floated
luxuriantly down his neck. His arm supported the real Elfrida, whose
personal charms amply deserved the encomium which had been lavished
upon them in the forest. Animation seemed hardly restored to that beautiful
form. Her eyes were half closed and her cheek very pale.
“Providence be thanked,” cried Leofwyn, “that my child is restored to
me!”
Now it has been already hinted that Elfrida was possessed of a
disposition somewhat untractable; in fact, loth as I am to speak aught ill of
the fair sex, I must confess that the Lady Elfrida partook, in no trifling
degree, both of the fantastic whims of her father Leofwyn and the violent
obstinacy of her brother Lothaire. The reader, therefore, will not be
surprised when he hears that the Saxon beauty, bowing respectfully to her
father, thus addressed him:
“Not to thee, my father, not to thee is thy daughter restored; in good and
in evil, in life and in death, she shall abide with her preserver—with him
who hath delivered her from the grasp of the spoiler.”
“Thou art mad, my child!” said the old man in astonishment; “the knight
that sued for thee thou didst contemn and reject, and wilt thou now wed
with his serving-man?”
Elfrida appeared to recollect the circumstances which had preceded her
capture, the suitor who had solicited her hand, and the deceit which she had
conspired to put upon him: she looked up to the daïs, and beheld Bertha, her
waiting-woman, seated by the side of the Norman guest; she glanced round
and met the eye of her preserver turned upon her with an expression of the
deepest adoration. She looked no further, but immediately, addressing her
father, said:
“Why should it not be so, my father? To-day thou hast married thine
handmaid to the Knight; to-morrow thou shalt marry thy daughter to the
Knave.”
Her unknown deliverer at these words began to stare about him; he
gazed upon his dress, upon his attendants, upon Elfrida; and then, with all
the embarrassment of a performer who comes forward to play in a pageant
without the smallest acquaintance with his part, observed: “This morning
was I a knight, mounted on a goodly steed, and clad in goodly apparel; but
whether I am now Norman or Saxon, knight or knave, by my grandfather’s
sword—I doubt.”
Leofwyn stared; his large eyes were dilated into a truly comic expression
of astonishment. “Who art thou?” he cried at last to the bridegroom; “art
thou Reginald d’Arennes, or must we hang thee for a rogue?”
“Peace, good father-in-law,” said the sham Reginald, shaking off his
drunkenness, and leering around him with an arch look of self-satisfaction;
“I am not Reginald d’Arennes, but yet as good a man! I am Robin, the son
of Egwulph; truly a cunning knave, and a wily.”
“I do begin to perceive,” said the waiting-woman, Bertha, looking on the
sham Reginald with a disappointed air, “that our plot hath altogether
failed.”
“Mine hath fared no better!” said the knave, returning a glance of equal
disappointment upon the mock Elfrida. “In this I have been but a silly
knave, and a witless!”
Dost thou comprehend, gentle reader, the circumstances which led to
these mistakes? or is it necessary for me to inform thee that the knave
Robin proceeded to Kennet Hold in Reginald’s apparel, with the purpose of
revenging, by his wedding with the heiress, the death of his master, which
he fancied had been occasioned by the heir; that at Kennet Hold the said
knave met with the counterplot which had been prepared by the jocose
Saxon, and became the husband of the maid instead of the mistress; that
Reginald, recovering from his swoon after the departure of his attendant,
advanced towards Kennet Hold, and encountered, in his way, his new
acquaintance, Richard de Mallory, from whom he had the good fortune to
rescue the life of Lothaire and the honour of Elfrida?
There is yet one point unexplained. The reader must be aware that a
considerable interval took place between the memorable blow given by
Lothaire and his rencontre with De Mallory. Upon this point the MS. makes
mention of Winifred, a certain arch-damsel, who—— but Decorum puts her
forefinger on her mouth—I have done.
Rather than desert a long established custom, I proceed to state that the
personages of my tale lived and loved to a green old age. Robin died before
it was thoroughly decided whether he was more properly termed “the Wily”
or “the Witless.” Reginald, it appears, never got rid of his old trick of
hesitation, for it is upon record that, when he told the story of his
adventures to Cœur de Lion, at the siege of Acre, and was asked by the
humorous monarch whether the knight or the knave were the more fortunate
bridegroom, he scratched his chin for a few minutes, played with his sword
for a few more, and replied slowly, “I have doubts as touching this matter.”
MAD—QUITE MAD!
“Great wits are sure to madness near allied.”—Dryden.
It has frequently been observed that Genius and Madness are nearly allied;
that very great talents are seldom found unaccompanied by a touch of
insanity, and that there are few bedlamites who will not, upon a close
examination, display symptoms of a powerful, though ruined, intellect.
According to this hypothesis, the flowers of Parnassus must be blended
with the drugs of Anticyra; and the man who feels himself to be in
possession of very brilliant wits may conclude that he is within an ace of
running out of them. Whether this be true or false, we are not at present
disposed to contradict the assertion. What we wish to notice is the pains
which many young men take to qualify themselves for Bedlam, by hiding a
good, sober, gentleman-like understanding beneath an assumption of
thoughtlessness and whim. It is the received opinion among many that a
man’s talents and abilities are to be rated by the quantity of nonsense he
utters per diem, and the number of follies he runs into per annum. Against
this idea we must enter our protest; if we concede that every real genius is
more or less a madman, we must not be supposed to allow that every sham
madman is more or less a genius.
In the days of our ancestors, the hot-blooded youth who threw away his
fortune at twenty-one, his character at twenty-two, and his life at twenty-
three, was termed “a good fellow,” “an honest fellow,” “nobody’s enemy
but his own.” In our time the name is altered; and the fashionable who
squanders his father’s estate, or murders his best friend—who breaks his
wife’s heart at the gaming-table, and his own neck at a steeple-chase—
escapes the sentence which morality would pass upon him, by the plea of
lunacy. “He was a rascal,” says Common Sense. “True,” says the World,
“but he was mad, you know—quite mad.”
We were lately in company with a knot of young men who were
discussing the character and fortunes of one of their own body, who was, it
seems, distinguished for his proficiency in the art of madness. “Harry,” said
a young sprig of nobility, “have you heard that Charles is in the King’s
Bench?” “I heard it this morning,” drawled the Exquisite; “how distressing!
I have not been so hurt since poor Angelica (his bay mare) broke down.
Poor Charles has been too flighty.” “His wings will be clipped for the
furture!” observed young Caustic. “He has been very imprudent,” said
young Candour.
I inquired of whom they were speaking. “Don’t you know Charles
Gally?” said the Exquisite, endeavouring to turn in his collar. “Not know
Charles Gally?” he repeated, with an expression of pity. “He is the best
fellow breathing; only lives to laugh and make others laugh; drinks his two
bottles with any man, and rides the finest mare I ever saw—next to my
Angelica. Not know Charles Gally? Why everybody knows him! He is so
amusing! Ha! ha! And tells such admirable stories! Ha! ha! Often have they
kept me awake”—a yawn—“when nothing else could.” “Poor fellow!” said
his lordship; “I understand he’s done for ten thousand!” “I never believe
more than half what the world says,” observed Candour. “He that has not a
farthing,” said Caustic, “cares little whether he owes ten thousand or five.”
“Thank Heaven!” said Candour, “that will never be the case with Charles:
he has a fine estate in Leicestershire.” “Mortgaged for half its value,” said
his lordship. “A large personal property!” “All gone in annuity bills,” said
the Exquisite. “A rich uncle upwards of fourscore!” “He’ll cut him off with
a shilling,” said Caustic.
“Let us hope he may reform,” sighed the Hypocrite; “and sell the pack,”
added the Nobleman; “and marry,” continued the Dandy. “Pshaw!” cried the
Satirist, “he will never get rid of his habits, his hounds, or his horns.” “But
he has an excellent heart,” said Candour. “Excellent,” repeated his lordship,
unthinkingly. “Excellent,” lisped the Fop, effeminately. “Excellent,”
exclaimed the Wit, ironically. We took this opportunity to ask by what
means so excellent a heart and so bright a genius had contrived to plunge
him into these disasters. “He was my friend,” replied his lordship, “and a
man of large property; but he was mad—quite mad. I remember his leaping
a lame pony over a stone wall, simply because Sir Marmaduke bet him a
dozen that he broke his neck in the attempt; and sending a bullet through a
poor pedlar’s pack because Bob Darrell said the piece wouldn’t carry so
far.” “Upon another occasion,” began the Exquisite in his turn, “he jumped
into a horse-pond after dinner in order to prove it was not six feet deep; and
overturned a bottle of eau-de-cologne in Lady Emilia’s face, to convince me
that she was not painted. Poor fellow! The first experiment cost him a dress,
and the second an heiress.” “I have heard,” resumed the Nobleman, “that he
lost his election for—— by lampooning the mayor; and was dismissed from
his place in the Treasury for challenging Lord C——.” “The last accounts I
heard of him,” said Caustic, “told me that Lady Tarrel had forbid him her
house for driving a sucking-pig into her drawing-room; and that young
Hawthorn had run him through for boasting of favours from his sister!”
“These gentlemen are really too severe,” remarked young Candour to us.
“Not a jot,” we said to ourselves.
“This will be a terrible blow for his sister,” said a young man who had
been listening in silence. “A fine girl—a very fine girl,” said the Exquisite.
“And a fine fortune,” said the Nobleman; “the mines of Peru are nothing to
her.” “Nothing at all,” observed the Sneerer; “she has no property there. But
I would not have you caught, Harry; her income was good, but is dipped,
horribly dipped. Guineas melt very fast when the cards are put by them.” “I
was not aware Maria was a gambler,” said the young man, much alarmed.
“Her brother is, Sir,” replied his informant. The querist looked sorry, but yet
relieved. We could see that he was not quite disinterested in his inquiries.
“However,” resumed the young Cynic, “his profusion has at least obtained
him many noble and wealthy friends.” He glanced at his hearers, and went
on: “No one that knew him will hear of his distresses without being forward
to relieve them. He will find interest for his money in the hearts of his
friends.” Nobility took snuff; Foppery played with his watch-chain;
Hypocrisy looked grave. There was long silence. We ventured to regret the
misuse of natural talents, which, if properly directed, might have rendered
their possessor useful to the interests of society and celebrated in the
records of his country. Every one stared, as if we were talking Hebrew.
“Very true,” said his lordship, “he enjoys great talents. No man is a nicer
judge of horseflesh. He beats me at billiards, and Harry at picquet; he’s a
dead shot at a button, and can drive his curricle-wheels over a brace of
sovereigns. “Radicalism,” says Caustic, looking round for a laugh. “He is a
great amateur of pictures,” observed the Exquisite, “and is allowed to be
quite a connoisseur in beauty; but there,” simpering, “every one must claim
the privilege of judging for themselves.” “Upon my word,” said Candour,
“you allow poor Charles too little. I have no doubt he has great courage—
though, to be sure, there was a whisper that young Hawthorn found him
rather shy; and I am convinced he is very generous, though I must confess
that I have it from good authority that his younger brother was refused the
loan of a hundred when Charles had pigeoned that fool of a nabob but the
evening before. I would stake my existence that he is a man of unshaken
honour—though, when he eased Lieutenant Hardy of his pay, there
certainly was an awkward story about the transaction, which was never
properly cleared up. I hope that when matters are properly investigated he
will be liberated from all his embarrassments; though I am sorry to be
compelled to believe that he has been spending double the amount of his
income annually. But I trust that all will be adjusted. I have no doubt upon
the subject.” “Nor I,” said Caustic. “We shall miss him prodigiously at the
Club,” said the Dandy, with a slight shake of the head. “What a bore!”
replied the Nobleman, with a long yawn. We could hardly venture to
express compassion for a character so despicable. Our auditors, however,
entertained very different opinions of right and wrong! “Poor fellow! he
was much to be pitied: had done some very foolish things—to say the truth
was a sad scoundrel—but then he was always so mad.” And having come
unanimously to this decision, the conclave dispersed.
Charles gave an additional proof of his madness within a week after this
discussion by swallowing laudanum. The verdict of the coroner’s inquest
confirmed the judgment of his four friends. For our own parts we must
pause before we give in to so dangerous a doctrine. Here is a man who has
outraged the laws of honour, the ties of relationship, and the duties of
religion; he appears before us in the triple character of a libertine, a
swindler, and a suicide. Yet his follies, his vices, his crimes, are all palliated
or even applauded by this specious façon de parler—“He was mad—quite
mad!”

Market Interrelationships and Applied Demand Analysis: Bridging The Gap Between Theory and Empirics in Commodities Markets Michael K. Wohlgenant

Uploaded by

Document Information

Original Title

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

Market Interrelationships and Applied Demand Analysis: Bridging The Gap Between Theory and Empirics in Commodities Markets Michael K. Wohlgenant

Uploaded by

Copyright:

Available Formats

Market Interrelationships and Applied

Demand Analysis: Bridging the Gap

Palgrave Textbooks in Agricultural Economics

More information about this subseries at http://www.palgrave.com/gp/series/16444

ISSN 2662-3889 ISSN 2662-3897 (electronic)

Cover credit: © SimonStock/Alamy Stock Photo

3.4.1 Theorem 1: Conditional Demand Functions . . . . . . . . . 23

5.2.4 Unconditional Demand Elasticities

7.2.1 Price Spread Relationships with Quantity/Price . . . . . . 103

8.2.2 Total Demand Elasticities for Retail-to-Farm

9.6.1.1 GMM Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

Dr. Michael K. Wohlgenant is William Neal Reynolds Distinguished Professor

Table 5.1 Constant Dollar Expenditures Per Capita for Meats,

Table 7.2 Cross-Price and Income Elasticity Estimates, and Farm

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 1

1.2 Organization of Chapters

2.2 Consumer’s Choice Problem

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 5

where q is an nx1 vector of quantities of commodities consumed, p is the corre-

2.2.1 First-Order Conditions: Application

Consider the Lagrangian function:

2.2.2 Demand Functions

2.3 Comparative Statics of Income and Price Changes

In anticipation of the ensuing empirical specification, a vector of exogenous shift

2.3.1 Fundamental Matrix Equation of Consumer Demand

The total differentials of Eqs. (2.3) and (2.4) are as follows:

Equation (2.7) is the Fundamental Matrix Equation of Consumer Demand (Barten

2.3.2 Slutsky Substitution Matrix

Substituting (2.9) into (2.10) and defining λZ = K we obtain the matrix of

2.3.3 General Restrictions of Consumer Behavior

ei = ∂g , ei∗j = ∂∂hp ij qij , and wi =

wi ei∗j = w j e∗ji (2.20)

Negativity: eii + wi ei ≤ 0, eii∗ ≤ 0∀i (2.22)

2.4 Comparative Statics of Exogenous Preference Shift

The effects of preference shift parameters from (2.11) are

Q s = −ZV = −λ−1 K V (2.23)

2.5 Comparative Statics with Strictly Concave Utility

We can relate changes in marginal utilities to changes in demand by assuming further

element of the inverse matrix U −1 , the (i, j) th element of K is:

Houthakker (1960) refers to λU i j as the specific substitution effect of the total

ei∗j = ei j − φei w j e j ∀i, j

The decomposition of the Hicksian price elasticities is into Frisch-compensated

2.6 Duality Theory: Relationships Between Marshallian,

In addition to Marshallian demand functions, (2.2), and Hicksian demand functions,

The indirect utility function, ϕ(y, p, s), is increasing in income, decreasing in

The expenditure or cost function,C(u, p, s), is increasing in prices and utility,

Finally, we have the relationships corresponding to the Frischian demand

q = g(y, p, s) = h(u, p, s) = f (r, p, s) (2.41)

2.7 Inverse Demand Functions

It is sometimes the case that we want demand functions, expressed in terms of

2.7.1 Wold-Hotelling Identity

π = g −1 (q) = b(q) (2.42)

which is the vector of uncompensated inverse demand functions. The first-order

The Wold-Hotelling identity delivers directly the uncompensated inverse demand

2.7.2 Comparative Statics

In order to determine the properties of uncompensated inverse demand functions,

dπ = gd log Q + Gdq (2.49)

Totally differentiating the budget constraint gives π dq + q dπ = 0, which

which are adding-up properties. We also have the result

(a) Derive Marshallian, Hicksian, and Frischian demand functions using

A significant problem to address prior to empirical implementation of the theory of

3.2 Restrictions on Price Movements

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 21