P. 1


|Views: 140|Likes:
Published by The Hill Newspaper
“Institutions such as religion, democracy and government anti-diversion policies all significantly enhance a country's long-run economic performance,” Brat wrote in 2004.
“Institutions such as religion, democracy and government anti-diversion policies all significantly enhance a country's long-run economic performance,” Brat wrote in 2004.

More info:

Published by: The Hill Newspaper on Jun 11, 2014
Copyright:Traditional Copyright: All rights reserved


Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See more
See less






(But How can a Fed Guy Forget the Institutions)

Professor David A. Brat
Department of Economics & Business
Randolph-Macon College
Ashland, VA 23005


Is long-run economic growth exogenous? To address this question, Bernanke and Gurkaynak (B&G) show
that the empirical framework of Mankiw, Romer, and Weil (1992) can be extended to test any growth
model that admits a balanced growth path. Their broad conclusion is that long-run growth is significantly
correlated with behavioral variables such as the savings rate. Hence, future empirical studies should focus
on models that exhibit endogenous growth. Fair Enough. But have B&G read the growth literature lately?
Douglass North won a Nobel Prize for his work on Institutions. Hall and Jones have codified these
institutions and shown that they should be considered the “ultimate” source and cause of growth. The
World Bank has institutionalized much of this project. And yet no mention of institutions exists in the
Bernanke paper. Are we all describing the same world? Even if growth is endogenous, why do we have to
go back to variables which economists used to favor? If savings causes growth and growth causes savings,
we are still left with the ultimate question: What is the ultimate cause of savings? Why is Africa poor? Why
don’t they save? Is it because of endogeneity issues? I don’t think so. Institutions are the new story. I will
show how institutions can be placed squarely within the B&G growth framework. As there does not appear
to be a clear way to adjudicate between competing right hand side variables at this time in the literature, I
will follow the herd. I will simply argue that my story is one step more complete than B&G’s 2001 story.

I. Introduction

“This paper takes Robert Solow seriously.” Thus begins one of the most influential and widely
cited pieces in the empirical growth literature, a 1992 article by N. Gregory Mankiw, David Romer, and
David Weil. It also begins the paper by Bernanke and Gurkaynak (B&G) which will be examined here. In
brief, Mankiw, Romer, and Weil (1992), henceforth MRW, performed an empirical evaluation of a
“textbook” Solow (1956) growth model using the Penn World Tables, a multi-country data set constructed
by Summers and Heston (1988) for the years 1960-1985. MRW found support for the Solow model’s
predictions that, in the long-run steady state, the level of real output per worker by country should be
positively correlated with the saving rate and negatively correlated with the rate of labor-force growth.
However, their estimates of the textbook Solow model also implied a capital share of factor income of
about 0.60, high compared to the conventional value (based on U.S. data) of about one-third.

To address this possible inconsistency, MRW considered an “augmented” version of the Solow
model, in which human capital enters as a factor of production in symmetrical fashion with physical capital
and raw labor. They found that the augmented Solow model fits the data relatively better and yields an
estimated capital share more in line with conventional wisdom.

MRW concluded that “an augmented Solow model that includes accumulation of human as well as physical
capital provides an excellent description of the cross-country data.”
As B&G note, the fact that MRW’s
augmented Solow model fits the cross-country data well is an interesting finding (and, as they point out, the
results could have been otherwise). However, as B&G will discuss in some detail below, it is not entirely
clear to what degree the good fit of the MRW specification may be attributed to elements that are common
to many models of economic growth (such as the Cobb-Douglas production structure), and how much of

MRW abstract, pg. 407
the fit is due to elements that are specific to the Solow formulation (such as the exogeneity of steady-state
growth rates). Indeed, as B&G have shown, MRW’s basic estimation framework is broadly consistent with
any growth model that admits a balanced growth path – a category that includes virtually all the growth
models in the literature. Hence, B&G conclude, one might argue that MRW do not actually test the Solow
model, in the sense of distinguishing it from possible alternative models of economic growth.

In their paper, B&G modestly extend the empirical framework introduced by MRW and use it to reevaluate
both the Solow model and some alternatives. In particular, they re-examine the crucial prediction
of the Solow model, that long-run economic growth is determined solely by exogenous technical change
and is independent of variables such as the aggregate saving rate, schooling rates, and the growth rate of
the labor force.

B&G find strong statistical evidence against the basic Solow prediction. In particular, they find that a
country’s rate of investment in physical capital is strongly correlated with its long-run growth rate of output
per worker, and that rates of human capital accumulation and population growth are also correlated, though
somewhat less strongly, with the rate of economic growth. They, thus, find that growth is endogenous.
Growth is correlated with savings rates.
This is their new story. It is true but it is not complete.

II. Bernanke’s Replication and Extension of the MRW Results

The original MRW article used cross-national data for the period 1960-1985. In their paper, B&G
replicate the MRW results for 1960-1985 and extend them through 1995. They find that the MRW
conclusions about the fit of the textbook Solow model and the augmented Solow model seem slightly
weaker when they use revised and/or extended data, though their main results survive. B&G also propose a
new test of the Solow model based on joint estimation of equations in the form of their equations (2.17) and
We will use this new specification proposed by B&G and test whether institutions belong in their
model. If their model is not specified properly, then of course, their findings are not statistically valid.

Following MRW and B&G, we draw our basic data from the Summers-Heston Penn World Tables (PWT),
which contain information on real output, investment, and population (among many other variables) for a
large number of countries. The data set used in the original MRW study was PWT version 4.0. The PWT
data have been revised twice since publication of the MRW article; as of this writing, PWT version 5.6
(which extends coverage of most variables through 1992) is the latest publicly available version.

MRW measure n as the average growth of the working-age population (ages 15 to 64). They
obtained these data from the World Bank’s World Tables and the 1988 World Development Report. B&G
use the original MRW data on working-age population in conjunction with the PWT 4.0 data set. For
analyses using PWT 5.6 and PWT 6.0, they use analogous data taken from the World Bank’s World
Development Indicators 2000 CD-ROM. The saving rate relevant to physical capital, K s , is measured as
the average share of gross investment in GDP, as in MRW.

MRW’s estimates of the augmented Solow model (with human capital accumulation) include a
variable they call SCHOOL, analogous to our H s , which is the average percentage of a country’s
workingage population in secondary school. More specifically, MRW define SCHOOL as the percentage of
schoolage population (12-17) attending secondary school times the percentage of the working-age
population that is of secondary school age (15-19). The age ranges in the two components of SCHOOL are
incommensurate but we are inclined to agree with MRW that the imperfect match up is not likely to create
major biases and we use the same construct. Data on enrollment rates, and working-age population and its
components are from the sources noted two paragraphs above and from the UN World Population

B&G pgs. 1-2
B&G, pg. 2
B&G, pg. 8
As in MRW, the three country samples we examine are 1) the “non-oil” sample, the set of all countries for
which complete data are available, excluding oil producers (98 countries); 2) the “intermediate” sample,
which is the non-oil sample excluding countries whose data receive a grade of “D” from Summers and
Heston or whose population is less than one million (75 countries); and 3) the OECD sample, OECD
countries with populations greater than one million (22 countries). Note that, because of missing data, the
sample sizes are in some cases slightly smaller when PWT 5.6 and PWT 6.0 are used for the replication.

A more powerful test of the Solow model

According to B&G, based on the their replication of MRW’s regression results, one might follow MRW
and draw broadly positive conclusions aboutthe fit of the Solow model, especially when augmented with
human capital. Notably, a simple regression using only three variates (the saving rate, the schooling rate,
and the population growth rate) seems to explain a remarkable share of cross-country variation in the level
of output per worker. It is true that the estimates of the production function coefficients are not always
reasonable, and we have found that the over-identifying restriction implied by the Cobb-Douglas structure
is often rejected, but problems with estimation of production relationships are not uncommon. Very
possibly, these statistical rejections are not of great economic significance.

However, as the B&G exposition in Section II suggests, “the results shown so far do not constitute the
strongest test of the Solow model within this framework. In their view, the better test of the Solow model
involves testing the restrictions on the analogue of equation (2.18), the equation explaining long-run
growth. In particular, if the hypothesis that the steady state of the Solow model describes the cross sectional
distribution of output per worker is true, then we should not be able to reject the hypothesis that
factors such as the saving rate or the rate of human capital accumulation do not enter into the determination
of the long-run growth rate.”

According to B&G, “formally, equations (2.17) and (2.18), together with the assumptions that all
countries share the same production function parameters and long-run growth rate, imply that where the
growth rate g is constant across countries. A straightforward statistical implication of the model, easily
tested in this framework, is that the coefficients on variables such as the saving rate, the schooling rate, and
the growth rate of the workforce rate should be zero, when they are entered on the right side of eq. (3.2).
(More precisely, we divide both sides of eq. (3.2) by the number of periods t, so that the annual growth rate
is on the right-hand side).”

Table I reports the results of the primary B&G test. (It is table 5 in the B&G paper.) Equations (3.1) and
(3.2) are estimated jointly by seemingly unrelated regression (SUR), with equation (3.2) being augmented
by the variables I/GDP,SCHOOL, and the labor-force growth rate n . The prediction of the Solow model
(under the auxiliary assumption of steady states) is that the estimated coefficients of the last three variables
should all be zero.

Table I shows the parameter estimates and standard errors for the augmented equation (3.2). In brief, B&G
show that the Solow model’s implication that growth is exogenous is strongly rejected for the nonoil
and intermediate samples. The z-statistics for I/GDP are bolded to show its statistical significance. The
implication is that growth is endogenous.

There are at least two possible reasons for the statistical rejections found in Table I: First, growth
may not be truly exogenous, in the sense of the Solow model. Second, the maintained hypothesis that the
countries in the sample are in the steady state may be wrong, i.e., we may be picking up transition

B&G, pg.12
B&G, pg.12
B&G, pg.12
B&G, pg.13
B&G show that “one simple test of the second possibility is to consider only the 22 countries in our sample
that are located in the Western hemisphere. Arguably, the assumption of steady states makes more sense for
Western hemisphere countries than for the rest of the world, as the Americas have not been the scene of
major wartime destruction, post-colonial transitions, or (except for Cuba, which is not in our sample)
sustained non-market experiments during the past century. Interestingly, as Table V shows, the
restrictions of the Solow model cannot be rejected for the countries of the Western hemisphere as a group.
Thus, it remains possible that the results of this section arise because of transition dynamics, not because
the Solow model is fundamentally wrong about long-run growth. In the latter part of their paper, B&G
address this issue directly by considering the determinants of TFP growth rather than output.

III. Bernanke’s Omission of Institutions: The Ultimate Cause of Growth!

The analysis thus far has reviewed the work of Bernanke and Gurkaynak. They present a solid case that
growth is not exogenous and their modeling exercise is helpful in understanding some of the shortcomings
of the Solow model and of the Mankiw paper. However, Bernanke’s critique applies to a 1992 paper. In the
meantime, much has taken place in the literature of Growth Economics. The highlight may be the 1993
Nobel Prize in Economics awarded to Douglass North for his pioneering work on institutions. Since that
time, prominent authors such as Hall and Jones and Barro, and indeed even Brat, have investigated the
effects of institutions on economic growth in just the cross-country framework which Bernanke uses.

The purpose of the present paper is to show that institutional variables belong in the framework of
Bernanke as presented above. Standard econometric evidence will clearly prove this case. It will also be
shown that institutions of all sorts are compatible with Bernanke’s story. Thus it is not simply one variable
which is being omitted. Much work needs to be done here. Previous work by Brat has shown that a
Protestant institutional variable explains why some countries are rich and why some are poor within the
Mankiw framework. Interestingly, the Protestant institutional variable is statistically significant in
explaining differences in income levels but not in explaining current growth rates. The same Protestant
variable produced similar results when included in the Barro framework and within the growth framework
of Hall and Jones. In all of these cases, arguments have been given for why the Protestant variable is
actually the best “ex ante” exogenous variable in the literature. Bradford Delong deserves the credit for
putting this argument forward in the American Economic Review back in 1988.

Doug North as Motivation

The motivation for this study is to provide an alternative perspective to the new macroeconomic growth
theory. While Mankiw, Romer and Weil, (MRW) have claimed that "the (human capital) augmented Solow
model provides an almost complete explanation of why some countries are rich and others are poor"
Douglass North has taken another tack.

In his 1993 Nobel lecture, North stated that "Neoclassical theory is simply an inappropriate tool to analyze
and prescribe policies that will induce development. It is concerned with the operation of markets, not with
how markets develop. In fact, most societies throughout history got “stuck” in an institutional matrix that
did not evolve into the impersonal exchange essential to capturing the productivity gains that came from
the specialization and division of labor that have produced the Wealth of Nations.”

Such a theory sounds plausible, but how does one test such a theory empirically? One paper in particular
set the stage for such an empirical analysis. Bradford DeLong discovered that "there is one striking ex ante
association between economic growth in the period 1870-1979 and a predetermined variable: this variable
is a nation's dominant religious establishment."

B&G, pg. 13
MRW, pg. 408
North, pgs. 359, 364
DeLong, pg. 1146
DeLong further notes that while difficult to interpret, his regression results "do serve as an example of how
culture may be associated with substantial divergence in growth performance.....a country's religious
establishment has been a surprisingly good proxy for the social capability to assimilate modern

Is there a systematic relationship between a country's dominant religious establishment and the standard of
living across countries? How do we examine this question? The heart of the argument in this paper is that
while MRW claim “an almost complete explanation” of economic development based on their three right
hand side variables, they have not really “explained” development at all. It is true that rich countries have
relatively more capital, both physical and human. It is also true that these variables are correlated with high
incomes. However, the more interesting development question is why these countries acquired these types
of capital inputs in the first place.

To conclude that rational actors choose to acquire less capital and remain in the midst of high levels of
absolute poverty commits the rational actor theory to a strange position indeed. Douglass North concludes
that “it is necessary to dismantle the rationality assumption underlying economic theory in order to
approach constructively the nature of human learning. History demonstrates that ideas, ideologies, myths,
dogmas and prejudices matter.”
DeLong’s regressions presented above reveals that “religion” does
matter in history.

The Protestant Institution

The positive coefficient on the Protestant institutional variable indicates that countries with a Protestant
religious establishment have “on average” higher income levels. As DeLong noted above, it is this variable
and not capital or education or democracy or OECD status which has been associated with economic
growth over the long run. It is therefore logical that these countries should now have higher incomes on
average. They do.

Equally significant is that this variable is the “only ex ante” variable of interest. It is the variable which
comes first historically and explains future economic behavior. A dummy variable representing OECD
countries would be trivial as it is a collection of those countries which have become rich and therefore begs
the whole question of what caused this richness. It is quite another thing to claim that a country’s religious
establishment in 1600 would predict future income levels.

Why might Protestant countries have grown more quickly over the long run as suggested by DeLong?
Should we still expect a positive connection between Protestantism and economic performance? Douglass
North has given the broad theoretical answer to this question. To the extent that Protestantism provides an
efficient set of property rights and encourages a modern set of economic incentives, one might anticipate
positive economic performance. If one looks to the Reformation period itself, it is clear that at a minimum,
the Protestant countries provided a relatively more decentralized economic environment for economic
actors. Swanson gives an extensive analysis of this claim.

Swanson shows that a remarkable pattern emerges in Western Europe that remains constant for over three
centuries. Countries with Catholic regimes tend to be more centralized. Countries with Lutheran or
Anglican (Protestant) regimes tend to be intermediate and Calvinist (Protestant) regimes tend to be the
most decentralized. No normative claims are being made here. But a clear linkage is being established
between religion, regime and decentralization in order to add content to North’s claim that institutions
matter in history. Could this decentralization encourage the institutional matrix of “ impersonal exchange
essential” to modern economic life?
I will put this forward as my best tentative hypothesis thus far. I will
also show that many other tentative hypotheses are not as strong.

DeLong, pg.1148
North, pg. 362
See Swanson.

IV. The Institutional Results – Protestant, Democracy and Government Corruption Variables

With our hypothesis in mind, let’s take a look at the regression results. Table 1 is the exact replication of
the B&G paper. Table I shows the parameter estimates and standard errors for B&G’s augmented equation
(3.2). In brief, B&G show that the Solow model’s implication that growth is exogenous is strongly rejected
for the nonoil and intermediate samples. The z-statistics are bolded to show statistical significance. The
implication is that growth is endogenous. But this story is not complete as Tables 2 – 4 will make clear.

Looking primarily at the intermediate sample of 72 countris, the B&G regression shows that all of the
Mankiw right hand side variables are statistically significant in explaining the level of income per capita in
1995, but only investment is significant in explaining growth rates from 1960 – 1995.

In Table 2, the DeLong Protestant variable is added to the B&G framework. Again, looking primarily at
the intermediate sample of 70 countries, the Protestant variable is statistically significant in explaining both
income level and growth rates. However, the sign is different and this result appears to be robust across the
other samples as well. How can Protestant be positive on income but negative on growth? If you are
familiar with the growth literature, it becomes clear that the Protestant variable in the growth equation is
probably just stealing the explanatory power from another variable which is usually present, specifically,
the initial income of the country at 1960. This is the case. When initial income is added to the right hand
side of the growth equation, Protestant is no longer negatively signed and it is not statistically significant in
any equation. So the net result is that Protestant explains the level of income in the cross-country
regressions but does not explain current growth. Is this true for other institutional variables as well?

In Table 3, Barro’s Democracy variable is added to the B&G equation instead of the Protestant variable.
The results are similar, but the democracy variable is not quite as robust. It also suffers from endogeneity
problems which will be discussed below. It is clear that there is two-way correlation between income and
democracy over time. In the largest non-oil sample, democracy is statistically significant in explaining the
level of income. In the intermediate sample, it is not. In the OECD sample, it is statistically significant.

In Table 4, I have included perhaps the strongest institutional variable in the literature. Hall and Jones have
constructed a variable called GADP. GADP is their measure of Government Anti-Diversion Policy. It is
their primary “institutional variable.” They follow Knack and Keefer (1995) using an equally weighted
average of 5 categories for the years 1986 – 1995. Two of the five categories relate to the government’s
(positive) role in protecting against private diversion: (i) law and order, and (ii) bureaucratic quality. Three
categories relate to the government’s possible (negative) role as diverter: (iii) corruption, (iv) risk of
expropriation, and (v) government repudiation of contracts.

As Table 4 makes clear, this variable is highly significant in all of the level regressions. This variable
therefore demands serious attention. However, a potential problem with their specification is that their most
important variable for explaining long run growth is constructed using very recent data for a variable which
they acknowledge to be endogenous. They address this problem in their 1999 paper. Even with this
correction, however, they are still using 1990’s institutional data to explain economic growth from 1800-
2000. This is a problem.

A thorough discussion of several problems with their model including identification, endogeneity,
measurement error and simultaneity is given in Hall and Jones (1999). One of these problems, however,
feeds directly into the discussion which follows.

Hall and Jones claim that “the other important characteristic of an instrument is lack of correlation with the
disturbance term. To satisfy this criterion, they ask whether European influence was somehow more
intensively targeted toward regions of the world that are more likely to have high output per worker today.
They note that in fact, this does not seem to be the case. Europeans did seek to conquer and exploit areas
of the world that were rich in natural resources such as gold and silver or that could provide valuable trade
in commodities. There is no tendency today for these areas to have high output per worker.”
They need
this assumption for their results to hold. By focusing on “these areas”, Hall and Jones, commit themselves
backward in economic theory and look only at the basic economic variables (resources) for explanatory
power. However, their current thesis is about institutions. When one looks at “these areas” with
“institutional” glasses on, one might reach a very different conclusion.

Using their own language, we again ask “whether European influence was somehow more intensively
targeted toward regions of the world that are more likely to have high output per worker today?” The
answer is yes and the channel by which it traveled may be the institution called Protestant religious
establishment. This is precisely the result which DeLong reports. Our data include a list of the 13 Protestant
countries used in the Bradford DeLong paper. One sees not only a striking European influence in several
now rich countries but also a striking correlation between Protestant and about every explanatory variable
used by Hall and Jones. Remember, all variables are scaled [0,1] in the Hall paper. Thus, when one sees
mean values for the Protestant country institutional variables in the 0.7 (0-1) range and mean values for the
non-Protestant countries in the 0.4 (0-1) range, it seems that Hall and Jones have some explaining to do.
They are aware of the DeLong variable and so the omission of its impact is a mystery.

V. Conclusion
B&G have re-visited Mankiw, Romer, and Weil’s classic empirical study of the Solow model of
economic growth. They showed that the MRW framework applies broadly to almost any economic growth
model that admits a balanced growth path, and that the restrictions specifically imposed by the Solow
model tend to be rejected. In particular, they find that variables such as the saving rate seem to be strongly
correlated with long-run growth rates. The correlation of variables like the saving rate with long-run output
growth rates is inconsistent with the joint hypothesis that the Solow model is true and the economies being
studied are in their respective steady states. The finding that the saving rate and the growth rate of the labor
force are correlated with estimated TFP growth is inconsistent with the standard Solow model, even if we
do not assume steady states.
This paper agrees with all of B&G thus far. However, as the research above
makes clear, B&G left out institutions. And institutions may well be the “ultimate” cause of long-run
economic growth. This is therefore no small omission. The regressions above show that institutions such as
religion, democracy and government anti-diversion policies all significantly enhance a country’s long-run
economic performance. The theory above suggests that the religion variable may be the strongest ex ante,
exogenous institutional variable in the literature. If all of this is true, then the B&G model is not specified
properly and the results which flow from it are not the final word on economic growth. Institutions may not
be the final word either, but this research indicates that their inclusion in the economic growth literature is
one step in the right direction.

Hall & Jones, 1999, pg. 101.
B&G, pg. 28
Barro, Robert J. and Jong-Wha Lee, “International Data on Educational Attainment: Update and
Implications”, NBER working paper 7911, September 2000.

Barro, Robert J., N. Gregory Mankiw, and Xavier Sala-i-Martin, “Capital Mobility in Neoclassical Models
of Growth”, American Economic Review, 85, March 1995, 103-115.

Barro, Robert J. and Xavier Sala-i-Martin, Economic Growth, Cambridge: MIT Press, 1999.

Bernanke, Ben and Gurkaynak,Refet, “Is Growth Exogenous? Taking Mankiw, Romer and Weil
Seriously” NBER Macroeconomics Journal, 2001.

DeLong, Bradford "Productivity Growth, Convergence, and Welfare: Comment" American Economic
Review, 1988 Vol.78 No.5.

Durlauf, Steven N. and Danny T. Quah, “The New Empirics of Economic Growth”, in John B. Taylor and
Michael Woodford, eds., Handbook of Macroeconomics Vol. 1, Amsterdam: Elsevier Science, 1999, 235-

Hall, Robert & Jones, Charles. “The Productivity of Nations”, NBER Working Paper 5812, 1996

Hall, Robert E. and Charles I. Jones, “Why Do Some Countries Produce So Much More Output per Worker
than Others?’, Quarterly Journal of Economics, 114, February 1999, 83-116.

King, Robert G. and Sergio T. Rebelo, “Transitional Dynamics and Economic Growth in the Neoclassical
Model”, American Economic Review, 83, September 1993, 908-31.

Klenow, Peter and Andres Rodriguez-Clare, “The Neoclassical Revival in Growth Economics: Has It Gone
Too Far?”, in Ben Bernanke and Julio Rotemberg, eds., NBER Macroeconomics Annual, Cambridge MA:
MIT Press, 1997, pp. 73-103.

Lucas, Robert E., Jr., “On the Mechanics of Economic Development,” Jouranl of Monetary Economics, 22,
June 1988, 3-43.

Mankiw, N. Gregory, David Romer, and David N. Weil, “A Contribution to the Empirics of Economic
Growth”, Quarterly Journal of Economics, 107, May 1992, 407-37.

North, Douglass “Economic Performance Through Time” American Economic Review, June 1994. Nobel

Ramsey, Frank, “A Mathematical Theory of Saving”, Economic Journal, 88, 1928, 543-559.

Solow, Robert M., “A Contribution to the Theory of Economic Growth”, Quarterly Journal of Economics,
70, February 1956, 65-94.

Summers, Robert and Alan Heston, “A New Set of International Comparisons of Real Product and Price
Levels Estimates for 130 Countries, 1950-1985”, Review of Income and Wealth, 34, March
1988, 1-26.

Young, Alwyn, “The Tyranny of Numbers: Confronting the Statistical Realities of the East Asian Growth
Experience, “Quarterly Journal of Economics, 110, August 1995, 641-80.

Bernanke Table 5 Replicated: Test of exogeneity of growth in the Augmented Solow Model
(I/GDP is statistically significant in Non Oil and Inter. – Solow Mod Rejected - Growth is not Exogenous)

Dependent Variables: Top of Each Table - Log GDP per working-age person, 1995 (Level)
Bottom of Each Table - Change in log GDP per working-age person, 1960-1995 (Growth)

Non-Oil Sample: 90 Obs OECD Sample: 21 Obs
Level Coef. Std.Err. Z Level Coef. Std.Err. Z
I/GDP 0.54 0.11 5.12 I/GDP 0.08 0.28 0.27
PopGro -2.38 0.38 -6.2 PopGro -1.18 0.54 -2.17
School 0.63 0.09 7.34 School 1.06 0.29 3.66
Constant 5.67 1.09 5.19 Constant 9.50 1.78 5.35

Growth Growth
I/GDP 0.12 0.02 5.56 I/GDP 0.07 0.04 1.79
PopGro 0.03 0.15 0.21 PopGro -0.38 0.28 -1.35
School 0.07 0.05 1.44 School -0.12 0.10 -1.13
Constant -0.01 0.01 -1.47 Constant 0.02 0.01 1.95

Level: 0.80 R
Level: 0.52
Growth: 0.48 R
Growth: 0.27

Intermediate Sample: 72 Obs Western: 22 Obs
Level Coef. Std.Err. Z Level Coef. Std.Err. Z
I/GDP 0.60 0.11 5.41 I/GDP 0.40 0.32 1.24
PopGro -1.85 0.35 -5.35 PopGro -2.47 0.97 -2.55
School 0.83 0.09 8.81 School 0.28 0.37 0.76
Constant 7.78 1.04 7.47 Constant 4.35 2.93 1.48

Growth Growth
I/GDP 0.12 0.02 5.29 I/GDP 0.05 0.04 1.23
PopGro 0.03 0.15 0.21 PopGro -0.31 0.27 -1.13
School 0.05 0.05 1.02 School -0.05 0.11 -0.44
Constant -0.01 0.01 -1.04 Constant 0.02 0.01 1.35

Level: 0.84 R
Level: 0.45
Growth: 0.43 R
Growth: 0.12


Bernanke Table 5 with DeLong Protestant Variable Added

Dependent Variables: Top of Each Table - Log GDP per working-age person, 1995 (Level)
Bottom of Each Table - Change in log GDP per working-age person, 1960-1995 (Growth)

Non-Oil Sample: 86 Obs OECD Sample: 21 Obs
Level Coef. Std.Err. Z Level Coef. Std.Err. Z
I/GDP 0.46 0.09 4.92 I/GDP 0.20 0.25 0.81
PopGro -1.88 0.38 -4.91 PopGro -1.13 0.47 -2.39
School 0.63 0.07 8.46 School 0.72 0.28 2.53
Prot 0.41 0.16 2.59 Prot 0.25 0.10 2.59
Constant 6.79 1.04 6.54 Constant 8.89 1.56 5.68

Growth Growth
I/GDP 0.10 0.02 5.33 I/GDP 0.05 0.03 1.91
PopGro -0.17 0.15 -1.17 PopGro -0.41 0.19 -2.21
School 0.10 0.05 2.19 School 0.06 0.07 0.81
Prot -0.01 0.00 -2.43 Prot -0.01 0.00 -5.22
Constant 0.00 0.00 -0.34 Constant 0.02 0.01 2.28

Level: 0.85 R
Level: 0.63
Growth: 0.49 R
Growth: 0.68

Intermediate Sample: 70 Obs Western: 22 Obs
Level Coef. Std.Err. Z Level Coef. Std.Err. Z
I/GDP 0.53 0.11 5.03 I/GDP 0.31 0.24 1.31
PopGro -1.52 0.37 -4.15 PopGro -1.71 0.74 -2.32
School 0.81 0.09 9.22 School 0.06 0.28 0.21
Prot 0.36 0.15 2.47 Prot 1.11 0.26 4.23
Constant 8.39 1.05 8 Constant 5.45 2.19 2.49

Growth Growth
I/GDP 0.10 0.02 5.12 I/GDP 0.05 0.04 1.13
PopGro -0.20 0.15 -1.3 PopGro -0.24 0.27 -0.89
School 0.09 0.05 1.76 School -0.06 0.11 -0.51
Prot -0.01 0.00 -2.5 Prot 0.01 0.01 0.87
Constant 0.00 0.01 -0.1 Constant 0.01 0.01 1.3

Level: 0.86 R
Level: 0.70
Growth: 0.45 R
Growth: 0.15


Bernanke Table 5 with Barro Democracy Variable Added

Dependent Variables: Top of Each Table - Log GDP per working-age person, 1995 (Level)
Bottom of Each Table - Change in log GDP per working-age person, 1960-1995 (Growth)

Non-Oil Sample: 77 Obs OECD Sample: 21 Obs
Level Coef. Std.Err. Z Level Coef. Std.Err. Z
I/GDP 0.40 0.11 3.76 I/GDP 0.02 0.24 0.07
PopGro -1.93 0.41 -4.69 PopGro 0.05 0.63 0.09
School 0.68 0.09 7.74 School 0.42 0.33 1.3
Democ 0.40 0.20 1.99 Democ 1.46 0.50 2.9
Constant 6.46 1.09 5.93 Constant 10.01 1.52 6.6

Growth Growth
I/GDP 0.10 0.02 4.82 I/GDP 0.07 0.04 1.89
PopGro -0.14 0.16 -0.9 PopGro -0.48 0.40 -1.18
School 0.10 0.05 2.03 School -0.08 0.13 -0.61
Democ -0.01 0.01 -1.66 Democ -0.01 0.02 -0.4
Constant 0.00 0.01 0.4 Constant 0.03 0.02 1.58

Level: 0.82 R
Level: 0.65
Growth: 0.45 R
Growth: 0.27

Intermediate Sample: 65 Obs Western: 22 Obs
Level Coef. Std.Err. Z Level Coef. Std.Err. Z
I/GDP 0.54 0.13 4.13 I/GDP 0.38 0.32 1.2
PopGro -1.69 0.41 -4.12 PopGro -2.13 1.01 -2.11
School 0.81 0.10 8.45 School 0.21 0.37 0.56
Democ 0.27 0.21 1.32 Democ 0.51 0.54 0.96
Constant 7.82 1.14 6.85 Constant 4.62 2.88 1.6

Growth Growth
I/GDP 0.10 0.02 4.41 I/GDP 0.05 0.04 1.2
PopGro -0.20 0.17 -1.16 PopGro -0.27 0.28 -0.96
School 0.07 0.05 1.32 School -0.05 0.11 -0.5
Democ -0.01 0.01 -1.58 Democ 0.00 0.01 0.34
Constant 0.01 0.01 0.8 Constant 0.01 0.01 0.92

Level: 0.83 R
Level: 0.47
Growth: 0.38 R
Growth: 0.12

Bernanke Table 5 with Hall & Jones GADP “Institution” Variable Added

Dependent Variables: Top of Each Table - Log GDP per working-age person, 1995 (Level)
Bottom of Each Table - Change in log GDP per working-age person, 1960-1995 (Growth)

Non-Oil Sample: 86 Obs OECD Sample: 21 Obs
Level Coef. Std.Err. Z Level Coef. Std.Err. Z
I/GDP 0.20 0.09 2.32 I/GDP 0.04 0.19 0.22
PopGro -0.93 0.35 -2.67 PopGro -0.50 0.38 -1.31
School 0.58 0.06 9.19 School 0.12 0.26 0.46
GADP 2.17 0.32 6.88 GADP 2.42 0.47 5.17
Constant 7.33 0.83 8.83 Constant 6.98 1.27 5.48

Growth Growth
I/GDP 0.10 0.02 4.22 I/GDP 0.07 0.03 2.52
PopGro 0.03 0.16 0.17 PopGro -0.74 0.22 -3.31
School 0.07 0.05 1.42 School 0.18 0.10 1.78
GADP 0.00 0.01 0.32 GADP -0.07 0.02 -4.29
Constant -0.01 0.01 -0.95 Constant 0.06 0.01 4.85

Level: 0.89 R
Level: 0.78
Growth: 0.46 R
Growth: 0.60

Intermediate Sample: 70 Obs Western: 22 Obs
Level Coef. Std.Err. Z Level Coef. Std.Err. Z
I/GDP 0.27 0.11 2.47 I/GDP 0.06 0.23 0.27
PopGro -0.85 0.35 -2.46 PopGro -1.79 0.67 -2.69
School 0.74 0.08 9.34 School -0.08 0.26 -0.31
GADP 1.83 0.33 5.47 GADP 2.56 0.50 5.14
Constant 8.31 0.88 9.48 Constant 3.02 2.00 1.51

Growth Growth
I/GDP 0.10 0.03 4.01 I/GDP 0.04 0.04 0.83
PopGro 0.00 0.17 0.01 PopGro -0.25 0.27 -0.93
School 0.05 0.05 0.9 School -0.08 0.11 -0.72
GADP 0.00 0.01 0.32 GADP 0.02 0.01 1.14
Constant -0.01 0.01 -0.7 Constant 0.01 0.01 0.81

Level: 0.89 R
Level: 0.75
Growth: 0.24 R
Growth: 0.16

You're Reading a Free Preview

/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->