American Economic Association
Does Trade Cause Growth?
Author(s): Jeffrey A. Frankel and David Romer
Source: The American Economic Review, Vol. 89, No. 3 (Jun., 1999), pp. 379-399
Published by: American Economic Association
Stable URL: http://www.jstor.org/stable/117025
Accessed: 31-12-2017 15:06 UTC
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
American Economic Association is collaborating with JSTOR to digitize, preserve and extend
access to The American Economic Review
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
Does Trade Cause Growth?
By JEFFREY A. FRANKEL AND DAVID ROMER*
Examining the correlation between trade and income cannot identify the direction
of causation between the two. Countries' geographic characteristics, however, have
important effects on trade, and are plausibly uncorrelated with other determinants
of income. This paper therefore constructs measures of the geographic component
of countries' trade, and uses those measures to obtain instrumental variables
estimates of the effect of trade on income. The results provide no evidence that
ordinary least-squares estimates overstate the effects of trade. Further, they suggest
that trade has a quantitatively large and robust, though only moderately statistically
significant, positive effect on income. (JEL F43, 040)
This paper is an empirical investigation of the
impact of international trade on standards of
living. From Adam Smith's discussion of specialization
and the extent of the market, to the
debates about import substitution versus exportled
growth, to recent work on increasing returns
and endogenous technological progress, economists
interested in the determination of standards
of living have also been interested in
trade. But despite the great effort that has been
devoted to studying the issue, there is little
persuasive evidence concerning the effect of
trade on income.
To see the basic difficulty in trying to estimate
trade's impact on income, consider a
cross-country regression of income per person
on the ratio of exports or imports to GDP (and
other variables). Such regressions typically find
a moderate positive relationship.' But this relationship
may not reflect an effect of trade on
* Frankel: Kennedy School of Government, Harvard
University, Cambridge, MA 02138; Romer: Department of
Economics, University of California, Berkeley, CA 94720.
We are grateful to Susanto Basu, Michael Dotsey, Steven
Durlauf, William Easterly, Robert Hall, Lars Hansen,
Charles Jones, N. Gregory Mankiw, Maurice Obstfeld,
Glenn Rudebusch, Paul Ruud, James Stock, Shangjin Wei,
Richard Zeckhauser, and the referees for helpful comments;
to Teresa Cyrus for outstanding research assistance and
helpful comments; and to the National Science Foundation
for financial sunnort.
income. The problem is that the trade share may
be endogenous: as Elhanan Helpman (1988),
Colin Bradford, Jr. and Naomi Chakwin (1993),
Rodrik (1995a), and many others observe,
countries whose incomes are high for reasons
other than trade may trade more.
Using measures of countries' trade policies in
place of (or as an instrument for) the trade share
in the regression does not solve the problem.2
For example, countries that adopt free-market
trade policies may also adopt free-market domestic
policies and stable fiscal and monetary
policies. Since these policies are also likely to
affect income, countries' trade policies are
likely to be correlated with factors that are omitted
from the income equation. Thus they cannot
be used to identify the impact of trade (Xavier
Sala-i-Martin, 1991).
This paper proposes an alternative instrument
for trade. As the literature on the gravity model
of trade demonstrates, geography is a powerful
determinant of bilateral trade (see, for example,
Hans Linneman, 1966, Frankel et al., 1995, and
Frankel, 1997). And as we show in this paper,
the same is true for countries' overall trade:
simply knowing how far a country is from other
countries provides considerable information
' See, for example, Michael Michaely (1977), Gershon
Feder (1983), Roger C. Kormendi and Philip G. Meguire
(1985), Stanley Fischer (1991, 1993), David Dollar (1992),
Ross Levine and David Renelt (1992), Sebastian Edwards
(1993), Ann Harrison (1996), and the ordinary least-squares
regressions in Section II of this paper. Edwards (1995) and
Dani Rodrik (1995b) survey the literature.
2 For examples of this approach, see J. Bradford De
Long and Lawrence H. Summers (1991), Fischer (1991,
1993), Dollar (1992), William Easterly (1993), Edwards
(1993), Jong-Wha Lee (1993), Jeffrey D. Sachs and Andrew
Warner (1995), and Harrison (1996).
379
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
380 THE AMERICAN ECONOMIC REVIEW JUNE 1999
abo-ut the amount that it trades. For example, the
fact that New Zealand is far from most other
countries reduces its trade; the fact that Belgium
is close to many of the world's most populous
countries increases its trade.
Equally important, countries' geographic
characteristics are not affected by their incomes,
or by government policies and other factors that
influence income. More generally, it is difficult
to think of reasons that a country's geographic
characteristics could have important effects on
its income except through their impact on trade.
Thus, countries' geographic characteristics can
be used to obtain instrumental variables estimates
of trade's impact on income. That is the
goal of this paper.3
The remainder of the paper contains two
main sections. The first describes the impact of
geographic characteristics on trade in more detail
and uses geographic variables to construct
an instrument for international trade. As we
discuss there, there is one important complication
to our basic argument that geographic variables
can be used to construct an instrument for
international trade in a cross-country income
regression. Just as a country's income may be
influenced by the amount its residents trade with
foreigners, it may also be influenced by the
amount its residents trade with one another.
And just as geography is an important determinant
of international trade, it is also an important
determinant of within-country trade. In
particular, residents of larger countries tend to
engage in more trade with their fellow citizens
simply because there are more fellow citizens to
trade with. For example, Germans almost surely
trade more with Germans than Belgians do with
Belgians. This suggests a second geographybased
test of the impact of trade on income: we
can test whether within-country trade raises income
by asking whether larger countries have
higher incomes.
The reason that this issue complicates our test
of the impact of international trade is that country
size and proximity to other countries are
negatively correlated. Because Germany is
larger than Belgium, the average German is
farther from other countries than the average
Belgian is. Thus in using proximity to estimate
international trade's effect on income, it is
necessary to control for country size. Similarly,
in using country size to test whether withincountry
trade raises income, it is necessary to
control for international trade.
To construct the instrument for international
trade, we first estimate a bilateral trade
equation and then aggregate the fitted values
of the equation to estimate a geographic component
of countries' overall trade. In contrast
to conventional gravity equations for bilateral
trade, our trade equation includes only geographic
characteristics: countries' sizes, their
distances from one another, whether they
share a border, and whether they are landlocked.
This ensures that the instrument
depends only on countries' geographic characteristics,
not on their incomes or actual
trading patterns. We find that these geographic
characteristics are important determinants
of countries' overall trade.
The second main section of the paper employs
the instrument to investigate the impact of
trade on income. We estimate cross-country regressions
of income per person on international
trade and country size by instrumental variables
(IV), and compare the results with ordinary
least-squares (OLS) estimates of the same equations.
There are five main findings.
First, we find no evidence that the positive
association between international trade and income
arises because countries whose incomes
are high for other reasons engage in more trade.
On the contrary, in every specification we consider,
the IV estimate of the effect of trade is
larger than the OLS estimate, often by a considerable
margin. Section II, subsection E, investigates
possible reasons that the IV estimate
exceeds the OLS one.
Second, the point estimates suggest that the
impact of trade is substantial. In a typical specification,
the estimates imply that increasing the
ratio of trade to GDP by one percentage point
3Lee (1993) also uses information on countries' distances
from one another to construct a measure of their
propensity to trade. His approach differs from ours in two
major respects. First, his measure is based not only on
countries' geographic characteristics, but also on their actual
trade patterns; thus it is potentially correlated with other
determinants of income. Second, he does not investigate the
relationship between income and his measure of the propensity
to trade, but only the relationship between income
and the interaction of his measure with indicators of distortionary
trade policies.
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
VOL. 89 NO. 3 FRANKEL AND ROMER: DOES TRADE CAUSE GROWTH? 381
raises income per person by between one-half
and two percent.
Third, the estimates also imply that increased
size raises income. This supports the
hypothesis that greater within-country trade
raises income.
Fourth, the large estimated positive effects of
trade and size are robust to changes in specification,
sample, and construction of the
instrument.
Fifth, the impacts of trade and size are not
estimated very precisely. The null hypothesis that
these variables have no effect is typically only
marginally rejected at conventional levels. As a
result, the estimates still leave considerable uncertainty
about the magnitudes of their effects.
I. Constructing the Instrument
A. Background
Our basic ideas can be described using a
simple three-equation model. First, average income
in country i is a function of economic
interactions with other countries ("international
trade" for short), economic interactions within
the country ("within-country trade"), and other
factors:
(1) ln Yi = a + /3T, + yWi + se.
Here Yi is income per person, Ti is international
trade, Wi is within-country trade, and si reflects
other influences on income. As the vast literature
on trade describes, there are many channels
through which trade can affect income-notably
specialization according to comparative advantage,
exploitation of increasing returns from
larger markets, exchange of ideas through communication
and travel, and spread of technology
through investment and exposure to new goods.
Because proximity promotes all of these types
of interactions, our approach cannot identify the
specific mechanisms through which trade affects
income.
The other two equations concem the determinants
of intemational and within-country trade. International
trade is a function of a country's proximity to
other countries, Pi, and other factors:
(2) Ti = + + Pi + 6i.
Similarly, within-country trade is a function of
the country's size, Si, and other factors:
(3) Wi = n + Asi + vi.
The residuals in these three equations,
and vi, are likely to be correlated. For e
countries with good transportation systems, or
with government policies that promote competition
and reliance on markets to allocate resources,
are likely to have high international
and within-country trade given their geographic
characteristics, and high incomes given their
trade.
The key identifying assumption of our analysis
is that countries' geographic characteristics
(their Pi's and Si's) are uncorrelated with the
residuals in equations (1) and (3). Proximity and
size are not affected by income or by other
factors, such as government policies, that affect
income. And as we observe in the introduction,
it is difficult to think of important ways that
proximity and size might affect income other
than through their impact on how much a country's
residents interact with foreigners and with
one another.
Given the assumption that P and S are uncorrelated
with s, data on Y, T, W, P, and S
would allow us to estimate equation (1) by
instrumental variables: P and S are correlated
with T and W [by (2) and (3)], and are uncorrelated
with e (by our identifying assumption).
Unfortunately, however, there are no data on
within-country trade. Ideally, we would want
data comparable to measures of international
trade. That is, we would want a measure of the
value of the exchange of all goods and services
among individuals within a country, both across
and within firms. This measure would probably
be many times GDP for most countries. But no
such measure exists.
To address this problem, we substitute (3)
into (1) to obtain
(4) ln Yi
=a + fTi + y(Tq + ASi + vi) + si
- (a + yr) + 3Ti + yASi
+ (YVi + Si)This
content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
382 THE AMERICAN ECONOMIC REVIEW JUNE 1999
Our identifying assumption implies that
Si are uncorrelated with the composite re
yvi + si. Thus (4) can be estimated by ins
mental variables, with Pi and Si (and the
stant) as the instruments.
Note that estimation of (4) yields an estimate
not only of ,3, international trade's impact
on income, but also of yA, country size's
impact on income. Since the two components
of this coefficient are not identified separately,
one cannot obtain an estimate of y, the
effect of within-country trade on income. But
as long as A is positive-that is, as long as
larger countries have more within-country
trade-the sign of y is the same as the sign of
yA. Thus, although we cannot estimate the
magnitude of the impact of within-country
trade on income, we can obtain evidence
about its sign.
As we argue in the introduction (and verify
below), Pi and Si are negatively correlated: the
larger a country is, the farther its typical resident
is from other countries. Thus if we do not
control for size in (4), Pi will be negatively
correlated with the residual, and thus will not be
a valid instrument. Intuitively, smaller countries
may engage in more trade with other countries
simply because they engage in less withincountry
trade. This portion of the geographic
variation in international trade cannot be used to
identify trade's impact on income. Similarly, if
we fail to control for Ti in (4), Si will be
negatively correlated with the residual. Thus,
our approach requires us to examine the impacts
of both international trade and country size.
To estimate (4), we need data on four variables:
income (Y), international trade (T), size
(S), and proximity (P). We measure the first
three in the usual ways. Our measure of income
is real income per person. Following standard
practice, we measure international trade as imports
plus exports divided by GDP. This is
clearly an imperfect measure of economic interactions
with other countries, an issue we return
to in Section II, subsection E.4 Finally,
theory provides little guidance about the best
measure of size. We therefore use two natural
measures, population and area, both in logs. In
interpreting the results concerning size, we focus
on the sum of the coefficients on log population
and log area. Thus we consider the
impact of an increase in population and area
together, with no change in population density.
Such a change clearly increases the scope for
within-country trade.5
To measure proximity, we need an appropriate
weighted average of distance or ease of
exchange between a given country and every
other country in the world. To choose the
weights to put on the different counties, we
begin by estimating an equation for bilateral
trade as a function of distance, size, and so on.
We then use the estimated equation to find fitted
values of trade between countries i and j as a
share of i's GDP. Finally, we aggregate over j
to obtain a geographic component of country i's
total trade, Ti. It is this geographic component
of T1 that we use as our measure of proximity.
The remainder of this section describes the specifics
of how we do this.
B. The Bilateral Trade Equation
Work on the gravity model of bilateral trade
shows that trade between two countries is negatively
related to the distance between them and
positively related to their sizes, and that a loglinear
specification characterizes the data fairly
well. Thus, a minimal specification of the bilateral
trade equation is
(5) ln(Tij/GDPi) a0 + aIln Dij + a2ln Si
+ a3ln Sj+ eij
where Tij is bilateral trade between countries i
and j (measured as exports plus imports), Di1 is
the distance between them, and Si and S3 are
measures of their sizes.
This specification omits a considerable amount
of geographic inforation about trade. The equation
that we estimate therefore differs from (5) in
4 Using the log of the trade share instead of the level
does not affect our conclusions. Examining import and
export shares separately yields, not surprisingly, the result
that the geographic components of the two shares are almost
identical. Thus our approach cannot separate the effects of
imports and exports.
' Throughout the paper, we use the labor force as our
measure of population in computing income per person and
in measuring country size. Using total population instead
makes little difference to the results.
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
VOL. 89 NO. 3 FRANKEL AND ROMER: DOES TRADE CAUSE GROWTH? 383
three ways. First, as described above, we include
two measures of size: log population and log area.
Second, whether countries are landlocked and
whether they have a common border have important
effects on trade; we therefore include dummy
variables for these factors. And third, a large part
of countries' trade is with their immediate neighbors.
Since our goal is to identify geographic
influences on overall trade, we therefore include
interaction terms of all of the variables with the
common-border dummy.
The fact that we are measuring trade relative
to country i's GDP means that we are already
including a measure of i's size. We therefore do
not constrain the coefficients on the population
measures for the two countries, or the coefficients
on the area measures, to be equal. We do,
however, constrain the coefficients on the landlocked
dummies, and their interactions with the
common-border dummy, to be equal for i and
j.6 Thus, the equation we estimate is
(6) ln(T1j/GDPi)
= ao + alln Di1 + a2lnNi + a3ln Ai
+ a4lnNj + a5lnAj + a6(Li + Lj)
+ a7Bij + a8Bijln Dij + a9Bi;ln Ni
+ aloBiiln Ai + a I jBiln Nj
+ a12Bi1ln Aj + a13Bi (Li + Lj) + eij,
where N is population, A is area, L is a dumnmy
for landlocked countries, and B is a dummy for
a common border between two countries.
C. Data and Results
We use the same bilateral trade data as
Frankel et al. (1995) and Frankel (1997); the
data are originally from the IFS Direction of
Trade statistics. They are for 1985, and cover
trade among 63 countries. Following these papers,
we drop observations where recorded bilateral
trade is zero. Distance is measured as the
great-circle distance between countries' principal
cities. The information on areas, common
borders, and landlocked countries is from Rand
McNally (1993). Finally, the data on population
are from the Penn World Table.7
The results are shown in Table 1. The first
column shows the estimated coefficients and
standard errors on the variables other than the
common-border dummy and its interactions.
These estimates are shown in the second
column.8
The results are generally as expected. Distance
has a large and overwhelmingly significant
negative impact on bilateral trade; the
estimated elasticity of trade with respect to distance
is slightly less (in absolute value) than
-1. Trade between country i and country j is
strongly increasing in j's size; the elasticity with
respect to j's population is about 0.6. In addition,
trade (as a fraction of i's GDP) is decreasing
in i's size and in j's area. And if one of the
countries is landlocked, trade falls by about a
third.
Because only a small fraction of country pairs
share a border, the coefficients on the commonborder
variables are not estimated precisely.
Nonetheless, the point estimates imply that
sharing a border has a considerable effect on
trade. Evaluated at the mean value of the variables
conditional on sharing a border, the estimates
imply that a common border raises trade
by a factor of 2.2. The estimates also imply that
the presence of a common border alters the
effects of the other variables substantially. For
example, the estimated elasticity with respect to
country j's population across a shared border is
0.47 rather than 0.61, and the estimated elasticity
with respect to distance is -0.70 rather than
-0.85.
Most importantly, the regression confirms
6 Allowing them to differ changes the results only triv-
ially.
7 We use Mark 5.6 of the table, which is distributed by
the National Bureau of Economic Research. This is an
updated version of the data described in Robert Summers
and Alan Heston (1991). The capital city is used as the
principal city, except for a small number of cases where the
capital is far from the center of the country (in terms of
population). In these cases, a more centrally located large
city is chosen. For the United States, for example, Chicago
rather than Washington is used as the principal city.
8 The coefficient on the commnon-border variable itself is
therefore shown as the coefficient on the interaction of the
common-border dummy with the constant.
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
384 THE AMERICAN ECONOMIC REVIEW JUNE 1999
TABLE 1-THE BILATERAL TRADE EQUATION
Variable Interaction
Constant -6.38 5.10
(0.42) (1.78)
Ln distance -0.85 0.15
(0.04) (0.30)
Ln population -0.24 -0.29
(country i) (0.03) (0.18)
Ln area -0.12 -0.06
(country i) (0.02) (0.15)
Ln population 0.61 -0.14
(country j) (0.03) (0.18)
Ln area -0.19 -0.07
(country j) (0.02) (0.15)
Landlocked -0.36 0.33
(0.08) (0.33)
Sample size 3220
R 2 0.36
SE of regression 1.64
Notes: The dependent variable is ln(T i/GDP,). The fir
column reports the coefficient on the variable listed, and th
second column reports the coefficient on the variable's
interaction with the common-border dummy. Standard errors
are in parentheses.
that geographic variables are major determinants
of bilateral trade. The R2 of the regression
is 0.36. The next step is to aggregate across
countries and see if geographic variables are
also important to overall trade.9
D. Implications for Aggregate Trade
To find the implications of our estimates for
the geographic component of countries' overall
trade, we aggregate the fitted values from the
bilateral trade equation. That is, we first rewrite
equation (6) as
(7) ln(Tij/GDPi) = a'Xij + eij,
where a is the vector of coefficients in (6) (ao,
al, ..., al 3), and Xii is the vector of right-hand
side variables (1, In Di1, ..., Bij[Li + L>]). Our
estimate of the geographic component of country
i's overall trade share is then
(8) ^ E
jui
That is, our est
of country i's trade is the sum of the estimated
geographic components of its bilateral trade with
each other country in the world.'0
All that is needed to perform the calculations
in equation (8) are countries' populations and
geographic characteristics. We therefore take
the sum in (8) not just over the countries covered
by the bilateral trade data set, but over all
countries in the world.1' Similarly, we are able
to find the constructed trade share, T, for all
countries, not just those for which we have
bilateral trade data. Since our income regressions
will also require data on trade and income,
however, we limit our calculation of T to the
countries in the Penn World Table. Thus we
compute T for 150 countries.
E. The Quality of the Instrument
Figure 1 is a scatterplot of the true overall
trade share, T, against the constructed share, T.
The figure shows that geographic variables account
for a major part of the variation in overall
trade. The correlation between T and T is 0.62.
As column (1) of Table 2 shows, a regression of
9 The standard errors reported in Table 1 are conventional
OLS standard errors. It is likely that the residuals of
the bilateral trade equation are not completely independent,
and thus that the reported standard errors are too low. But as
described in Section II, subsection B, uncertainty about the
parameters of the bilateral trade equation contributes only a
small amount to the standard errors of the cross-country
income regressions that we ultimately estimate. For example,
doubling the variance-covariance matrix of the estimated
parameters of the bilateral trade equation increases
the standard error of the coefficient on the trade share in our
baseline cross-country regression [column (2) of Table 3]
by less than 10 percent.
'0The expectation of T i/GDPi conditional on Xii is
actually equal to eaXij times E[eeJi]. Since we are modeling
eiq as homoskedastic, however, E[eeii] is the same for all
observations, and thus multiplies tj by a constant. This has
no implications for the subsequent analysis, and is therefore
omitted for simplicity.
" For convenience, we omit a handful of countries with
populations less than 100,000: Antigua and Barbuda,
Greenland, Kiribati, Liechtenstein, the Marshall Islands,
Micronesia, Monaco, Nauru, St. Kitts and Nevis, and San
Marino. In addition, for countries that are not in the Penn
World Table, we have data on population but not on the
labor force. To estimate the labor force for these countries,
we multiply their populations by the average ratio of the
labor force to population amnong the countries in the same
continent that are in the Penn World Table. We use the Penn
World Table's definitions of the continents.
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
VOL. 89 NO. 3 FRANKEL AND ROMER: DOES TRADE CAUSE GROWTH? 385
350 -
300 -
250 -
200 Actual
Trade a
Share (percent) ? o
150 - o a o
50 } .
0 50 100 150 200 250 300
Constructed Trade Share (percent)
FIGURE 1. ACTUAL VERSUS CONSTRUCTED TRADE SHARE
T on a constant and t yields a coefficient on
of essentially one and a t-statistic of 9.5.
As described in subsection A of this section,
however, the component of the constructed trade
share that is correlated with country size cannot be
used to estimate trade's impact on income: smaller
countries may engage in more international trade
but in less within-country trade. That is, our identification
of trade's impact on income will come
from the component of the excluded exogenous
variable (the constructed trade share) that is uncorrelated
with the other exogenous variables (the
size measures).
The constructed trade share is in fact highly
correlated with country size. For example, the
five countries with the smallest constructed
shares all have areas over 1,000,000 square
miles, and the five with the largest constructed
shares all have areas under 10,000 square miles.
A regression of the constructed trade share on a
constant, log population, and log area yields
negative and significant coefficients on both
size measures and an R2 of 0.45.
Thus in examining whether geographic variables
provide useful information about intemational trade,
we need to ask whether they provide infonnation
beyond that contained in country size. Columns (2)
and (3) of Table 2 therefore compare a regression of
the actual trade share on a constant and the two size
measures with a regression that also includes our
constructed trade share. As expected, size has a negative
effect on trade. Area is highly significant, while
population is moderately so. The coefficient on the
constiucted trade share falls by slighdy more than
half when the size controls are added.
The important message of columns (2) and (3),
TABLE 2-THE RELATION BETWEEN ACTUAL AND
CONSTRUCTED OVERALL TRADE
(1) (2) (3)
Constant 46.41 218.58 166.97
(4.10) (12.89) (18.88)
Constructed trade share 0.99 0.45
(0.10) (0.12)
Ln population -6.36 -4.72
(2.09) (2.06)
Ln area -8.93 -6.45
(1.70) (1.77)
Sample size 150 150 150
R2 0.38 0.48 0.52
SE of regression 36.33 33.49 32.19
Notes: The dependent variable is the actual trade share.
Standard errors are in parentheses.
however, is that the constructed trade share still
contains a considerable amount of information
about actual trade. For example, its t-statistic in
column (3) is 3.6; this corresponds to an F-statistic
of 13.1. As the results in the next section show,
this means that the constructed trade share contains
enough information about actual trade for IV
estimation to produce only moderate standard errors
for the estimated impact of trade. Furthermore,
the results of Douglas Staiger and James H.
Stock (1997), Charles R. Nelson and Richard
Startz (1990), and Alastair R. Hall et al. (1996)
imply that these first-stage F-statistics are large
enough that the finite-sample bias of instrumental
variables-which biases the IV estimate toward
the OLS estimate-is unlikely to be a serious
problem in our IV regressions.
Figure 2, Panel A, shows the partial association
between the actual and constructed trade shares
controlling for the size measures. The figure
shows that although the relationship is not as
strong as the simple relationship shown in
Figure 1, it is still positive. The figure also shows
that there are two large outliers in the relationship:
Luxembourg, which has an extremely high fitted
trade share given its size, and Singapore, which
has an extremely high actual trade share given its
size. Figure 2, Panel B, therefore shows the scatterplot
with these two observations omitted. Again
there is a definite positive relationship.12
12 When these two observations are dropped from the
regression in column (3) of Table 2, the coefficient on the
constructed trade share rises to 0.69, but the t-statistic falls
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
386 THE AMERICAN ECONOMIC REVIEW JUNE 1999
200 'so
-
100 - n
Actual Trade 50 - a
Share (percent) 0 6 Ca
-50
-100 - ___ _ , _ ................ .
-50 o 50 0 o 50 200 250
Constructed Trade Share (percent)
A. Full Sample
125 -
100 -
75 - 3
13 O
50 - a
Actual Trade 25 - ?? ?
Share (percent) a B e s D
-25 s
-50 s o
o coo
-75.- -__________
-25 0 25 50
Constructed Trade Share (percent)
B. Luxembourg and Singapore Omitted
FIGURE 2. PARTIAL ASSOCIATION BETWEEN ACTUAL AND
CONSTRUCTED TRADE SHARE
The five countries with the smallest constructed
trade shares controlling for size are
Western Samoa, Tonga, Fiji, Mauritius, and
Vanuatu. All five are geographically isolated
islands. Of the five, Western Samoa and
Tonga have large negative shares given their
size; Fiji and Mauritius have moderate negative
shares; and Vanuatu has a moderate positive
share. At the other extreme, the five
countries with the highest constructed trade
shares controlling for size are Luxembourg,
Belize, Jordan, Malta, and Djibouti. Of these
five, Luxembourg and Jordan have large positive
actual trade shares controlling for size,
and the other three have moderate positive
shares. In sum, countries' geographic characteristics
do provide considerable information
about their overall trade.
II. Estimates of Trade's Effect on Income
A. Specifications and Samples
This section uses the instrument constructed
in Section I to investigate the relationship between
trade and income. The dependent variable
in our regressions is log income per person.
The data are for 1985 and are from the Penn
World Table. Our basic specification is
(9) Iln Y,=a + bT,
+ clln Ni + c21n Ai + ui,
where Yi is income per person in country i, Ti
the trade share, and Ni and Ai are population
and area. This specification differs from the one
derived from the simple model in Section I,
subsection A, only by including two measures
of size rather than one [see equation (4)].
We do not include any additional variables in
the regression. There are of course many other
factors that may affect income. But our argument
about the appropriateness of using
geographic characteristics to construct an instrument
for trade implies that there is no strong
reason to expect additional independent determinants
of income to be correlated with our
instrument; thus they can be included in the
error term. In addition, if we included other
variables, the estimates of trade's impact on
income would leave out any effects operating
through its impact on these variables. Suppose,
for example, that increased trade raises the rate
of return on domestic investment and therefore
increases the saving rate. Then by including the
saving rate in the regression, we would be omitting
trade's impact on income that operates via
saving.
As an alternative to including control variables,
in subsection D of this section, we decompose
countries' incomes in various ways
and ask how trade affects each component. In
doing so, we can obtain information about the
channels through which trade affects income.
For example, we ask how trade affects both income
at the beginning of the sample and growth
to 3.2 (corresponding to an F-statistic of 10.1). This
F-statistic remains large enough that the finite-sample bias
of the IV estimates is still likely to be small.
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
VOL. 89 NO. 3 FRANKEL AND ROMER: DOES TRADE CAUSE GROWTH? 387
over the sample period.13
Appendix Table Al reports the basic data
used in the tests. It lists, for each country in the
sample, its actual trade share in 1985, its constructed
trade share, its area and 1985 population,
and its income per person in 1985.14
We focus on two samples. The first is the full
set of 150 countries covered by the Penn World
Table. Our instrument is only moderately correlated
with trade once we control for size, and
much of the variation is among the smallest
countries in the sample. Thus it is important to
consider a relatively broad sample. And as we
describe below, the results for this sample are
robust to the exclusion of outliers and of observations
where the data are potentially the most
subject to error.
Our second sample is the 98-country sample
considered by N. Gregory Mankiw et al. (1992).
The countries in this sample generally have
more reliable data; they are also generally
larger, and thus less likely to have their incomes
determined by idiosyncratic factors. In addition,
data limitations require that we employ a
smaller sample when we examine the channels
through which trade affects income.
B. Basic Results
Table 3 reports the regressions. Column (1) is
an OLS regression of log income per person on a
constant, the trade share, and the two size measures.
The regression shows a statistically and
economically significant relationship between
trade and income. The t-statistic on the trade share
is 3.5; the point estimate implies that an increase
in the share of one percentage point is associated
with an increase of 0.9 percent in income per
person. The regression also suggests that, controlling
for international trade, there is a positive
(though only marginally significant) relation between
country size and income per person; this
supports the view that within-country trade is benTABLE
3-TRADE AND INCOME
(1) (2) (3) (4)
Estimation OLS IV OLS IV
Constant 7.40 4.96 6.95 1.62
(0.66) (2.20) (1.12) (3.85)
Trade share 0.85 1.97 0.82 2.96
(0.25) (0.99) (0.32) (1.49)
Ln population 0.12 0.19 0.21 0.35
(0.06) (0.09) (0.10) (0.15)
Ln area -0.01 0.09 -0.05 0.20
(0.06) (0.10) (0.08) (0.19)
Sample size 150 150 98 98
R 2 0.09 0.09 0.11 0.09
SE of
regression 1.00 1.06 1.04 1.27
First-stage F
on excluded
instrument 13.13 8.45
Notes: The dependent variable is log income per person in
1985. The 150-country sample includes all countries for
which the data are available; the 98-country sample includes
only the countries considered by Mankiw et al. (1992).
Standard errors are in parentheses.
eficial. The point estimates imply that increasing
both population and area by one percent raises
income per person by 0.1 percent.
Column (2) reports the IV estimates of the
same equation. The trade share is treated as
endogenous, and the constructed trade share is
used as an instrument. 15 The coefficient on trade
rises sharply. That is, the point estimate suggests
that examining the link between trade and
income using OLS understates rather than overstates
the effect of trade. The estimates now
imply that a one-percentage-point increase in
the trade share raises income per person by 2.0
percent. In addition, the hypothesis that the IV
coefficient is zero is marginally rejected at conventional
levels (t = 2.0). The coefficient is
1' Fischer (1993) uses a similar approach to investigate
the effects of inflation. Frankel et al. (1996) and the working
paper version of this paper (Frankel and Romer, 1996)
investigate the effects of controlling for physical and human
capital accumulation and population growth, and find that
this does not change the character of the results.
14 The other data used in the analysis are available from
the authors on request.
'5 Throughout, the standard errors for the IV regressions
account for the fact that the instrument depends on the
parameters of the bilateral trade equation. That is, the variance-covariance
matrix of the coefficients is estimated as
the usual IV formnula plus (a&/aA))Q(afI/a)', where 6 is the
vector of estimated coefficients from the cross-country income
regression, a is the vector of estimated coefficients
from the bilateral trade equation, and Q is the estimated
variance-covariance matrix of a. In all cases, this additional
term makes only a small contribution to the standard errors.
In the regression in column (2) of Table 3, for example, this
correction increases the standard error on the trade share
from 0.91 to 0.99.
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
388 THE AMERICAN ECONOMIC REVIEW JUNE 1999
much less precisely estimated under IV than
under OLS, however. As a result, the hypothesis
that the IV and OLS estimates are equal cannot
be rejected (t = 1.2).16
Moving from OLS to IV also increases the
estimated impact of country size. The estimated
effect of raising both population and area by one
percent is now to increase income per person by
almost 0.3 percent. This estimate is marginally
significantly different from zero (t = 1. 8).
One interesting aspect of the results concerning
size is that the coefficient on area is positive.
One might expect increased area, controlling for
population, to reduce within-country trade and
thus lower income. One possibility is that the
positive coefficient is due to sampling error: the
t-statistic on area is slightly less than one. Another
is that greater area has a negative impact
via decreased within-country trade, but a larger
positive impact via increased natural resources.
It is because of this possibility that we focus on
the sum of the coefficients on log population
and log area in our discussion. As described
above, this sum shows the effects of increased
size with population density held constant. In
addition, as we show below, using population
alone to measure size has no major impact on
the results.
Columns (3) and (4) consider the 98-country
sample. This change has no great effect on the
OLS estimates of the coefficients on trade and
size, but raises both the IV estimates and their
standard errors considerably. The t-statistics for
the OLS estimates fall moderately, while those
for the IV estimates change little.
Frankel et al. (1996) and the working paper
version of this paper (Frankel and Romer, 1996)
consider a second approach to constructing the
instrument for trade. In addition to using information
on the proximity of a country's trading
partners, these papers use some information
about the partners' incomes. Specifically, they
include measures of physical and human capital
accumulation and population growth. As those
papers explain, one can argue that the factor
accumulation of a country's trading partners is
uncorrelated with determinants of the country's
income other than its trade and factor accumulation:
once one accounts for the impact of a
country's own factor accumulation on its income,
the most evident channel through which
its partners' factor accumulation may affect its
income is by increasing the partners' incomes,
and thus increasing the amount the country
trades. Thus if one controls for the country's
own factor accumulation, its partners' factor
accumulation can be used in constructing the
instrument.
Not surprisingly, using more information in
constructing the instrument increases the precision
of the IV estimates of trade's effect on
income. But the estimated effect of trade is not
systematically different when one moves to the
alternative instrument; this supports the argument
that it is a valid instrument. The overall
results are similar to those we obtain with our
basic instrument: the IV estimates of trade's
impact on income are much larger than the OLS
estimates, and are marginally significantly different
from zero.
C. Robustness
A natural question is whether the results are
robust. We consider robustness along four
dimensions.
First, as described above, Luxembourg and
Singapore are major outliers in the relationship
between the actual and constructed trade shares.
Dropping either or both of these observations,
however, does not change the basic pattern of
the results. The most noticeable change occurs
when Luxembourg is dropped from the baseline
regressions in columns (1) and (2) of Table 3.
This has effects similar to those of moving to
the 98-country sample (which does not include
Luxembourg): the OLS estimate of the effect of
trade changes little, but the IV estimate and its
standard error rise. Similarly, adding Luxembourg
to the regressions in columns (3) and (4)
of Table 3 moderately reduces the IV estimate
of the effect of trade.
16 That is, we perform a Hausman test (Jerry A. Hausman,
1978) of the hypothesis that actual trade is uncorrelated
with the residual, and thus that OLS is unbiased.
Under the null, asymptotically the OLS and IV estimates of
trade's impact differ only because of sampling error. As a
result, the difference between the two estimates divided by
the standard error of the difference is distributed asymptotically
as a standard normal variable. Furthermore, under the
null the variance of the difference between the IV and OLS
estimates is just the difference in their variances; that is, the
standard error of the difference is the square root of the
difference in the squares of their standard errors.
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
VOL. 89 NO. 3 FRANKEL AND ROMER: DOES TRADE CAUSE GROWTH? 389
A second concern is the possibility that systematic
differences among parts of the world are
driving the results. That is, it could be that our
IV estimates of the impact of trade arise because
the countries in certain regions of the world
have systematically higher constructed trade
shares given their size and also have systematically
higher incomes. In this case, our findings
might be the result not of trade, but of other
features of those regions.
To address this concem, we reestimate the regressions
in Table 3 including a dummy variable
for each continent. This modification substantially
increases the standard errors of the IV estimates of
the impact of trade on income. As a result, the
estimates are no longer significantly different
from zero. In addition, the IV estimate of the
impact of trade for the 150-country sample [column
(2) of Table 3] falls to only moderately above
the OLS estimate. When Luxembourg is dropped
from the sample, however, the IV estimate retums
to being much larger than the OLS estimate. For
the 98-country sample, in contrast, inclusion of the
continent dummies raises the IV estimate slightly
and lowers the OLS estimate slightly.
As an alternative way of considering the impact
of differences across parts of the world, we
follow Robert E. Hall and Charles I. Jones
(1999) and include countries' distance from the
equator as a control variable. This variable may
reflect the impact of climate, or it may be a
proxy for omitted country characteristics that
are correlated with latitude. With this approach,
the IV estimate of trade and size's effects are
virtually identical to the OLS estimate for the
full sample, and only moderately larger than the
OLS estimate for the 98-country sample. Thus
there is some evidence that systematic differences
among regions are important to our finding
that the IV estimates of trade's effects
exceed the OLS estimates. Even in this case,
however, there is still no evidence that the OLS
estimates overstate trade's effects.
Third, our data are highly imperfect in various
ways. Potentially most important, for the major
oil-producing countries much of measured GDP
represents the sale of existing resources rather
than genuine value added. Since these countries
have among the highest measured incomes in the
sample, they have the potential to affect the results
substantially. In fact, however, they are not important
to our findings: they are already absent
from the 98-country sample, and excluding them
from the 150-country sample changes the estimates
only slightly. As a more general check of
the possible importance of data problems, we use
the Penn World Table's summary assessments of
the quality of countries' data to exclude the countries
with the poorest data from the sample. Specifically,
we exclude all countries whose data are
assigned a grade of "D." This reduces the 150country
sample to 99, and the 98-country sample
to 77. These reductions in the sample sizes moderately
increase the standard errors of both the
OLS and IV estimates of the impact of trade. The
point estimates change little, however.
Finally, one can imagine reasons that virtually
all the variables used in finding the geographic
component of countries' trade might
have some endogenous component that is correlated
with the error term in the income equation.
For example, whether countries have
access to an ocean may be endogenous in the
truly long run, and may be determined in part by
other forces that affect income. 17 Similarly,
population is endogenous in the very long run.
To check that no single variable that could
conceivably be endogenous is driving the results,
we redo the construction of the instrument
and the regressions in Table 3 in five ways:
omitting the landlocked variable from the bilateral
trade equation; excluding population from
this equation; omitting all interactions with the
common-border dummy from this equation; using
total population rather than the labor force
both in measuring countries' sizes and in computing
income per person; and excluding area
from both equations.
None of these changes has a major effect on
the results. Although the changes sometimes
affect the IV estimates noticeably, in every case
they remain much larger than the OLS estimates.
Moreover, there is no systematic tendency
for the changes to reduce the IV
'7 The potential endogeneity of characteristics of borders
other than whether countries are landlocked is unlikely to
cause serious difficulties, for two reasons. First, the location
of borders is largely determined by forces other than govemnment
policies and other determinants of current income;
that is, the endogenous component of borders appears small.
Second, because our estimates control for within-country
trade, what is key to our estimates is the overall distribution
of population, not the placement of country borders.
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
390 THE AMERICAN ECONOMIC REVIEW JUNE 1999
estimates; indeed, there are several cases where
the changes raise the estimate considerably, or
reduce its standard error considerably. Thus, our
findings do not hinge on the use of any one of
these variables in constructing the fitted trade
shares.
D. The Channels Through Which Trade
Affects Income
The results thus far provide no information
about the mechanisms through which trade
raises income. To shed some light on this issue,
we decompose income and examine trade's impact
on each component.
We consider two decompositions of income.
The first follows Hall and Jones (1999). Suppose
output in country i is given by
(10) Yi = K [ee(si)AjNj]l a,
where K and N are capital and labor, S is
workers' average years of schooling, 4(Q) gives
the effects of schooling, and A is a productivity
term. Equation (10) could be used to decompose
differences in output per worker into the contributions
of capital per worker, schooling, and
productivity. As Hall and Jones note, however,
an increase in A leads to a higher value of K for
a given investment rate. Following Peter Klenow
and Andres Rodriguez-Clare (1997), they
therefore rewrite (10) as
( 11) Yi = (Kj1Yj)a(-)e4(si)AjNj.
Dividing both sides by Ni and taking logs yields
a
(12) ln(Y1/Nj)=1 ln(Ki/Yi)
+ 4(Sf) + ln Ai.
Equation (12) expresses the log of output per
worker as the sum of the contributions of capital
depth (reflecting such factors as investment and
population growth), schooling, and productivity.
Hall and Jones set a = 1/3 and let 4() be a
piecewise linear function with coefficients
based on microeconomic evidence. This allows
them to measure each component of (12) other
than In Ai directly from the data, and then find
In Ai as a residual.
Our second decomposition of income is simpler.
We write log output per worker in 1985 as
the sum of its value at the beginning of the
sample (1960) plus the change during the sample
period:
(1 3) 1n(Y /Nj) 1985
In(Yj1Nj) 1960
+ [ln( Yi/Ni) 1985- ln( Y1/Nj) 1960]
For both decompositions, we regress each
component of income on a constant, the trade
share, and the size measures. Again, we consider
both OLS and IV. Since the decompositions
cannot be performed for the full sample,
we consider only the 98-country sample. The
results are reported in Table 4. The first three
pairs of columns show the results for the Hall
and Jones decomposition, and the remaining
two pairs show the results for the decomposition
into initial income and subsequent growth.
With both decompositions, the estimates
using instrumental variables suggest that
trade increases income through each component
of income. For the first decomposition,
the estimated impacts of trade on physical
capital depth and schooling are moderate, and
its estimated impact on productivity is large.
The estimates imply that a one-percentagepoint
increase in the trade share raises the
contributions of both physical capital depth
and schooling to output by about one-half of
a percentage point, and the contribution of
productivity to output by about two percentage
points. For the second decomposition,
trade's estimated effects on both initial income
and subsequent growth are large. Here
the estimates imply that a one-percentagepoint
increase in the trade share raises both
initial income and the change over the sample
period by about one and a half percentage
points. Further, in every case, the estimates
suggest that country size, controlling for international
trade, is beneficial. And in every
case the IV estimates of the effects of international
trade and country size are substantially
larger than the OLS estimates.
The standard errors of the IV estimates are
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
VOL. 89 NO. 3 FRANKEL AND ROMER: DOES TRADE CAUSE GROWTH? 391
TABLE 4-TRADE AND THE COMPONENTS OF INCOME
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
Dependent a
variable 1 - a ln(Ki/Yi) C(Si) InAi ln(Y/N)1960 A ln(Y/N)
Estimation OLS IV OLS IV OLS IV OLS IV OLS IV
Constant -0.72 -1.29 0.10 -0.37 7.47 3.05 7.45 4.27 -0.50 -2.65
(0.34) (0.93) (0.30) (0.81) (0.74) (2.84) (1.03) (3.07) (0.39) (1.66)
Trade share 0.36 0.59 0.18 0.37 0.27 2.04 0.38 1.66 0.45 1.31
(0.10) (0.36) (0.08) (0.31) (0.21) (1.10) (0.29) (1.19) (0.11) (0.65)
Ln population 0.02 0.04 0.06 0.07 0.21 0.32 0.09 0.17 0.12 0.18
(0.03) (0.04) (0.03) (0.03) (0.06) (0.11) (0.09) (0.12) (0.03) (0.06)
Ln area 0.04 0.07 -0.01 0.01 -0.13 0.08 -0.02 0.13 -0.03 0.07
(0.02) (0.05) (0.02) (0.04) (0.05) (0.14) (0.07) (0.15) (0.03) (0.08)
Sample size 98 98 98 98 98 98 98 98 98 98
R 2 0.13 0.13 0.09 0.08 0.14 0.06 0.03 0.02 0.24 0.20
SE of
regression 0.32 0.33 0.28 0.29 0.69 0.92 0.96 1.06 0.36 0.47
First-stage F
on excluded
instrument 8.45 8.45 8.45 8.45 8.45
Note: Standard errors are in parenthes
large, however. For productivity and for growth
over the sample period, the estimates of trade's
effect are marginally significantly different
from zero (t-statistics of 1.8 and 2.0, respectively).
For the other dependent variables, the tstatistics
are between 1.2 and 1.6. Similarly, the
t-statistic for the sum of the coefficients on
population and area is 1.9 for productivity, 2.0
for growth, and between 1.3 and 1.5 for the
other dependent variables. Thus, although the
estimates suggest that international and withincountry
trade raise income through several different
channels, they do not allow us to
determine their contributions through each
channel with great precision.
The results also provide no evidence that
OLS is biased. The IV and OLS estimates of
trade's impact never differ by a statistically
significant amount. Indeed, the t-statistic for the
null that the estimates are equal exceeds 1.5
only once.18
E. Why Are the IV Estimates Greater Than
the OLS Estimates?
Both the simple model presented in Section I,
subsection A, and prevailing views about the
association between trade and income suggest
that IV estimates of trade's impact on income
will be less than OLS estimates. There are four
main reasons. First, countries that adopt freetrade
policies are likely to adopt other policies
18 The result that the IV estimates of the effects of trade
and country size on each component of income are positive
is extremely robust to the exclusion of outliers, the addition
of continent dummies and latitude, the omission of countries
with the most questionable data, the use of less information
in constructing the instrument, and expanding the
sample to include as many countries as possible (132 countries
for the Hall and Jones decomposition, 127 countries for
the decomposition into initial income and subsequent
growth). The one exception is that in the IV regression with
4(Si) as the dependent variable, both the coefficient on
trade and the sum of the coefficients on population and area
are very slightly negative when latitude is included as a
control. Likewise, the finding that the IV estimates of the
effects of trade and size on productivity, initial income, and
growth over the sample are larger than the OLS estimates is
extremely robust. Again there is only a single exception:
when latitude is included, the IV estimates of the impact of
trade and of the combined effect of population and area on
initial income are slightly smaller than the OLS estimates.
The result that the IV estimates of trade's and size's effects
on capital depth and schooling are larger than the OLS
estimates, on the other hand, is only moderately robust: for
several of our robustness checks, the IV estimates are
smaller. The difference is never large, however.
Finally, Hall and Jones's data are for 1988 rather than
1985. This is not important, however: redoing the basic
regressions in Table 3 for the 98-country sample using 1988
income per worker changes the results only trivially.
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
392 THE AMERICAN ECONOMIC REVIEW JUNE 1999
that raise income. Second, countries that are
wealthy for reasons other than trade are likely to
have better infrastructure and transportation
systems. Third, countries that are poor for reasons
other than low trade may lack the institutions
and resources needed to tax domestic
economic activity, and thus may have to rely on
tariffs to finance government spending. And
fourth, increases in income coming from
sources other than trade may increase the variety
of goods that households demand and shift
the composition of their demand away from
basic commodities toward more processed,
lighter weight goods. All of these considerations
would lead to positive correlation between
trade and the error term in an OLS
regression, and thus to upward bias in the OLS
estimate of trade's effects. And since there is no
reason to expect correlation between proximity
and these various omitted country characteristics,
there is no reason to expect the IV estimate
to suffer a similar bias. But we find that the IV
estimate is almost always considerably larger
than the OLS estimate. Although the difference
is not statistically significant, the fact that it is
large and positive is surprising.
The OLS estimate is determined by the partial
association between income and trade,
while the IV estimate is determined by the partial
association between income and the component
of trade correlated with the instrument.
Thus, mechanically, the fact that the OLS estimate
is smaller than the IV estimate means that
income's partial association with the component
of trade that is not correlated with the
instrument is weaker than its partial association
with the component that is correlated. Figure 3
presents these two partial associations. Since
Luxembourg and Singapore are outliers in the
relationship between trade and the instrument,
they are omitted. Panel A of the figure shows
the positive partial association between income
and the component of trade correlated with the
instrument; it is this association that underlies
the positive coefficient on trade in the basic IV
regression [column (2) of Table 3]. Panel B of
the figure shows that there is also a positive
partial association between income and the
component of trade not correlated with the instrument.
This association, though statistically
significant, is smaller than the partial association
in Panel A (b = 0.75, with a standard error
2.0
1.5~ ~ ~ 0
1.0 -d
0.5 -o O ?
Ln Income
-0.5
-1.0 O
- 1.5 - 0?gODO
-2.0
-2.5 -20
-I0 0 t o 20 30
Trade Share (percent)
A. lnocte and the Component of the Trade Share Correlated with the Instrument
2.0
1.5 n 0 n0
I.0 n
0.5 St i h s l a
Ln Income 0 00 t
-0.5
-1.0 0
-2.0 -
-2.5-60
-30 0 30 60 90 120 150
Trade Share (percent)
B. Income and the Component of the Trade Share Un
FIGURE 3. PARTIAL AsSOCIATIONs BETWEEN INCOME AND
THE TRADE SHARE
of 0 e26). It is this smaller relationship that
causes the OLS estimate to be less than the IV
estimate.
The figures show that the difference between
the OLS and IV estimates is not due to a handful
of observations. (Recall that including Luxembourg
and Singapore reduces the IV estimate of
trade's effects. Thus the only major outliers are
not the source of the gap between the OLS and
IV estimates, but in fact reduce it.) In addition,
examining which specific countries are important
to the estimates suggests that there is no
simple omitted country characteristic that is
driving the results.
The observations that are responsible for
the high estimated IV coefficient are those at
the upper right and lower left of Panel A. The
most influential observations at the upper
right are Qatar, Belgium, the United States,
and Israel. The most influential observations
at the lower left are Malawi, Lesotho, Burundi,
and Haiti. Similarly, the observations
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
VOL. 89 NO. 3 FRANKEL AND ROMER: DOES TRADE CAUSE GROWTH? 393
that are lowering the OLS estimate relative to
the IV estimate are those in the lower right
and upper left of Panel B. The most influential
observations at the lower right are Qatar,
Trinidad and Tobago, Hong Kong, and Syria,
while the most influential observations at the
upper left are Lesotho, Mauritania, Togo, and
Belize.
Thus, in both cases the low-income countries
are, not surprisingly, disproportionately
from sub-Saharan Africa. But there is no evident
characteristic other than proximity that
distinguishes the countries with low proximity
and low income from those with high
proximity and high income. Similarly, there is
no clear characteristic other than trade that
differentiates the countries with low residual
trade and high income from those with high
residual trade and low income. Together with
our earlier results, this suggests that the finding
that the IV estimate exceeds-or at least
does not fall short of-the OLS estimate is
robust and not the result of omitted country
characteristics.
There are two leading explanations of the
fact that the IV estimate of trade's impact
exceeds the OLS estimate. The first is that it
is due to sampling variation. That is, although
there no reason to expect systematic correlation
between the instrument and the residual,
it could be that by chance they are positively
correlated. The principal evidence supporting
this possibility is that the differences between
the IV and OLS estimates, though quantitatively
large, are well within the range that can
arise from sampling error. In our baseline
regressions [columns (1) and (2) of Table 3],
the t-statistic for the null hypothesis that the
OLS and IV estimates are equal is 1.2 (p =
0.25). And in the variations on the baseline
regression that we consider, this t-statistic
never exceeds 2, and it is almost always less
than 1.5. Moreover, in a few cases it is essentially
zero. Thus, if one believes that theory
provides strong grounds for believing that
OLS estimates are biased up, our IV estimates
do not provide a compelling reason for changing
this belief.
The second candidate explanation of the
finding that the IV estimates exceed the OLS
estimates is that OLS is in fact biased down.
The literal shipping of goods between countries
does not raise income. Rather, trade is a
proxy for the many ways in which interactions
between countries raise income-specialization,
spread of ideas, and so on. Trade
is likely to be highly, but not perfectly, correlated
with the extent of such interactions.
Thus, trade is an imperfect measure of income-enhancing
interactions among countries.
And since measurement error leads to
downward bias, this would mean that OLS
would lead to an understatement of the effect
of income-enhancing interactions.'9
To explore this idea, consider the following
simple model. Average income in country i is
given by
(14) ln Yi = a + bIj - b2I2i + 4- bNINi
+ cSi + ei,
where I1, I2. * - IN are measures of the various
ways in which interactions among countries affect
income. This equation has the same form as
(4), except that trade has been replaced by II,
I2, - IN. We want to rewrite this model in a
way that expresses the idea that trade is a noisy
measure of income-promoting interactions. To
do this, we define T* = (bIIIi + b2I2i ? ?
bNINi)/b, where b = Var(bjI + b2I2 + +
bNIN)fCov(blIl + b2I2 + ?-- + bNIN, T) and
where T is again trade. With this definition, we
can rewrite (14) as
(15) ln Y, = a + bT?+ cSi + ei.
It is straightforward to check that with these
definitions, T - T* is uncorrelated with T*.
Thus we can write
(16) Ti=T + ui,
where ui is uncorrelated wit
tions (15)-(16) express the idea that actual trade
is a noisy measure of income-promoting inter-
actions.
With this formulation, the relationship between
income and actual trade is
9 We are grateful to an anonymous referee and several
seminar participants for suggesting this possibility.
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
394 THE AMERICAN ECONOMIC REVIEW JUNE 1999
(17) ln Yi=a + bTi + cSi + (ei-bui).
From (16), T and u are positively correlated. If
this correlation is strong enough, it will cause
the overall correlation between T and the residual
in (17), e-bu, to be negative, and thus cause
OLS to yield a downward biased estimate of b.
And again this bias will not carry over to the IV
estimates-proximity and other geographic
variables are likely to be strongly correlated
with income-enhancing interactions, and there
is no evident reason for them to be correlated
with the idiosyncratic factors that cause trade to
move differently from those interactions.
In the univariate case, the expectation of
the OLS estimate of b would be given by
E[bOLS] = (VT*/VT)b, where VT* and VT are
the variances of T* and T. When size is
controlled for, the expression is analogous:
VTs
(18) E[bOLS] = VT b,
VTIS
where VT*IS and VTIs are the conditional variances
of T* and T given size.
This discussion implies that trade must be a
very noisy proxy for income-promoting interactions
to account for the gap between the OLS and
IV estimates. In our baseline regression for the full
sample [columns (1) and (2) of Table 3], the OLS
estimate of b is less than half the IV estimate.
Thus for the gap between the two estimates to
arise from the fact that trade is an imperfect proxy
for income-enhancing interactions, the majority of
the variation in trade uncorrelated with size would
have to be uncorrelated with income-promoting
interactions. To put it differently, u in (16) would
have to have a standard deviation of 22 percentage
points.
We conclude that the most plausible explanation
of the bulk of the gap between the IV and
OLS estimates is simply sampling error. This
implies that our most important finding is not
that the IV estimates of trade's effects exceed
the OLS estimates, but rather that there is no
evidence that the IV estimates are lower. In
addition, it implies that our IV estimates may be
substantially affected by sampling error, and
thus that the OLS estimates are likely to be
more accurate estimates of trade's actual impact
on income.
Illo Conclusion
This paper investigates the question of how
international trade affects standards of living.
Although this is an old question, it is a difficult
one to answer. The amounts that countries trade
are not determined exogenously. As a result,
correlations between trade and income cannot
identify the effect of trade.
This paper addresses this problem by focusing
on the component of trade that is due to
geographic factors. Some countries trade more
just because they are near well-populated countries,
and some trade less because they are isolated.
Geographic factors are not a consequence
of income or government policy, and there is no
likely channel through which they affect income
other than through their impact on a country's
residents' interactions with residents of
other countries and with one another. As a
result, the variation in trade that is due to geographic
factors can serve as a natural experiment
for identifying the effects of trade.
The results of the experiment are consistent
across the samples and specifications we consider:
trade raises income. The relation between
the geographic component of trade and
income suggests that a rise of one percentage
point in the ratio of trade to GDP increases
income per person by at least one-half percent.
Trade appears to raise income by spurring
the accumulation of physical and human
capital and by increasing output for given
levels of capital.
The results also suggest that within-country
trade raises income. Controlling for intemational
trade, countries that are larger-and that
therefore have more opportunities for trade
within their borders-have higher incomes. The
point estimates suggest that increasing a country's
size and area by one percent raises income
by one-tenth of a percent or more. And the
estimates suggest that within-country trade, like
international trade, raises income both through
capital accumulation and through income for
given levels of capital.
There are two important caveats to these conclusions.
First, the effects are not estimated with
great precision. The hypotheses that the impacts
of trade and size are zero are typically only
marginally rejected at standard significance levels.
In addition, the hypothesis that the estimates
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
VOL. 89 NO. 3 FRANKEL AND ROMER: DOES TRADE CAUSE GROWTH? 395
based on the geographic component of trade are
the same as the estimates based on overall trade
are typically relatively far from rejection. Thus,
although the results bolster the case for the
benefits of trade, they do not provide decisive
evidence for it.
This limitation is probably inherent in the experiment
we are considering. Once country size is
controlled for, geography appears to account for
only a moderate part of the variation in trade.
As a result, geographic variables provide only a
limited amount of information about the relation
between trade and income. Thus, unless additional
portions of overall trade that are unaffected by
other determinants of income can be identified, it
is likely to be difficult to improve greatly on our
estimates of the effects of trade.
The second limitation of the results is that
they cannot be applied without qualification to
the effects of trade policies. There are many
ways that trade affects income, and variations in
trade that are due to geography and variations
that are due to policy may not involve exactly
the same mix of the various mechanisms. Thus,
differences in trade resulting from policy may
not affect income in precisely the same way as
differences resulting from geography.
Nonetheless, our estimates of the effects of
geography-based differences in trade are at least
suggestive about the effects of policy-induced
differences. Our point estimates suggest that the
impact of geography-based differences in trade
are quantitatively large. We also find that the
estimated impact is larger than what one obtains
by naively using OLS, but is not significantly
different from the OLS estimate. These results
are not what one would expect if the positive
correlation between trade and income reflected
an impact of income on trade, or of omitted
factors on both variables. In that sense, our
results bolster the case for the importance of
trade and trade-promoting policies.
APPENDIX
TABLE Al-BASIC DATA
Area
Actual trade Constructed (thousands of Population Income per
Country share trade share square miles) (millions) worker
Algeria 49.66 13.97 919.595 4.859 13434
Angola 69.10 11.51 481.354 3.512 1742
Benin 76.99 42.20 43.483 1.874 2391
Botswana 121.28 24.03 231.800 0.370 6792
Burkina Faso 52.42 14.10 105.870 4.150 940
Burundi 30.82 24.86 10.747 2.539 986
Cameroon 57.67 15.79 183.569 3.831 3869
Cape Verde Islands 118.02 45.11 1.557 0.120 2829
Central African Republic 65.23 15.13 241.313 1.309 1266
Chad 61.43 12.00 495.755 1.791 1146
Comoros 67.06 46.77 0.863 0.181 1400
Congo 112.81 25.77 132.046 0.760 6878
Djibouti 117.06 70.97 8.958 0.105 4647
Egypt 51.97 11.75 386.900 12.719 7142
Ethiopia 34.13 8.44 472.432 18.385 705
Gabon 100.18 30.65 103.346 0.420 9672
Gambia 89.14 52.20 4.093 0.358 1609
Ghana 21.29 18.87 92.100 4.468 2237
Guinea 71.80 23.95 94.926 2.243 1583
Guinea-Bissau 62.74 42.24 13.948 0.425 1354
Ivory Coast 78.19 16.58 124.502 4.030 3740
Kenya 51.69 12.48 224.960 7.980 2014
Lesotho 152.42 20.66 11.720 0.743 2028
Liberia 79.63 29.81 43.000 0.811 2312
Madagascar 30.99 9.90 226.660 4.498 1707
Malawi 54.09 12.67 45.747 3.180 1171
Mali 73.60 12.80 482.077 2.332 1686
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
396 THE AMERICAN ECONOMIC REVIEW JUNE 1999
TABLE Al-Continued.
Area
Actual trade Constructed (thousands of Population Income per
Country share trade share square miles) (millions) worker
Mauritania 141.56 23.44 397.953 0.533 2674
Mauritius 109.10 31.11 0.787 0.577 7474
Morocco 58.50 12.71 172.413 6.714 6427
Mozambique 18.38 11.11 308.642 7.290 1417
Namibia 119.81 21.31 317.818 0.380 8465
Niger 51.27 12.37 489.206 3.343 1098
Nigeria 28.53 8.68 356.700 30.743 2874
Reunion 53.14 39.92 0.969 0.216 7858
Rwanda 30.65 26.20 10.169 3.005 1539
Senegal 70.63 19.87 75.954 2.758 2688
Seychelles 111.95 84.98 0.175 0.029 7058
Sierra Leone 19.15 27.81 27.700 1.372 2411
Somalia 25.64 14.89 246.199 2.774 1574
South Africa 55.43 8.90 471.440 11.240 9930
Sudan 21.34 10.97 967.491 7.121 2436
Swaziland 118.71 56.87 6.704 0.277 5225
Tanzania 21.03 10.97 364.900 10.266 975
Togo 105.52 41.47 21.925 1.277 1516
Tunisia 71.33 23.83 63.379 2.280 8783
Uganda 22.54 12.97 91.343 6.236 1224
Zaire 53.15 8.97 905.365 12.321 1136
Zambia 76.96 13.81 290.586 2.274 2399
Zimbabwe 56.40 11.27 150.699 3.135 3261
Bahamas 124.11 38.03 5.382 0.097 29815
Barbados 130.30 56.10 0.166 0.127 12212
Belize 183.27 87.48 8.866 0.049 8487
Canada 54.48 4.97 3851.809 12.595 31147
Costa Rica 63.19 23.37 19.652 0.920 9148
Dominica 103.09 75.08 0.305 0.030 6163
Dominican Republic 64.24 22.37 18.704 1.912 7082
El Salvador 52.21 28.91 8.260 1.564 5547
Grenada 120.63 81.25 0.133 0.039 4502
Guatemala 24.94 22.04 42.042 2.262 7358
Haiti 38.44 20.44 10.714 2.514 2125
Honduras 54.15 27.58 43.277 1.307 4652
Jamaica 131.89 22.19 4.411 1.059 4726
Mexico 25.74 4.52 761.600 24.669 17036
Nicaragua 36.60 23.46 50.180 0.980 5900
Panama 70.96 23.56 29.761 0.760 10039
Puerto Rico 136.74 22.75 3.515 1.101 21842
St. Lucia 165.77 68.83 0.238 0.057 5317
St. Vincent & Grenadines 152.17 79.41 0.150 0.042 5796
Trinidad & Tobago 61.90 30.33 1.980 0.441 25529
United States 18.01 2.56 3540.939 117.362 33783
Argentina 17.10 5.60 1072.067 10.798 14955
Bolivia 30.27 8.06 424.162 1.978 5623
Brazil 19.34 3.03 3286.470 49.609 10977
Chile 53.85 7.25 292.132 4.303 9768
Colombia 26.33 7.54 439.735 9.433 9276
Ecuador 47.63 11.42 109.484 2.820 9615
Guyana 109.95 25.92 83.000 0.280 3573
Paraguay 49.58 10.43 157.047 1.226 6241
Peru 39.42 7.03 496.222 6.107 8141
Suriname 82.99 30.96 63.251 0.124 10883
Uruguay 47.86 17.07 68.040 1.169 10216
Venezuela 40.76 8.94 352.143 5.789 18362
Bahrain 188.70 71.82 0.240 0.178 22840
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
VOL. 89 NO. 3 FRANKEL AND ROMER: DOES TRADE CAUSE GROWTH? 397
TABLE Al-Continued.
Area
Actual trade Constructed (thousands of Population Income per
Country share trade share square miles) (millions) worker
Bangladesh 25.78 10.31 55.598 27.684 4265
Bhutan 62.54 37.74 17.954 0.575 1504
China 19.44 2.30 3689.631 612.363 2166
Hong Kong 209.52 35.88 0.398 3.516 16447
India 15.04 3.29 1229.737 295.478 2719
Indonesia 42.66 4.47 735.268 62.136 4332
Iran 15.20 10.06 636.293 -13.540 13847
Iraq 49.22 19.14 169.235 4.105 15855
Israel 85.80 54.17 8.020 1.602 21953
Japan 25.54 5.47 143.574 '75.526 18820
Jordan 113.50 68.18 37.297 0.601 15655
Korea, Republic of 67.86 14.36 38.031 16.608 10361
Kuwait 96.45 42.55 6.880 0.640 35065
Laos 13.80 27.32 91.429 1.758 2739
Malaysia 104.69 16.82 128.328 6.217 10458
Mongolia 82.72 13.52 604.829 0.894 3966
Myanmar 13.16 10.74 261.220 16.613 1332
Nepal 31.29 13.26 54.463 6.958 2244
Oman 87.06 34.19 82.030 0.368 31609
Pakistan 34.00 8.04 310.400 28.567 4249
Philippines 45.84 8.84 115.830 19.945 4229
Qatar 80.94 69.56 4.412 0.166 36646
Saudi Arabia 79.97 14.98 865.000 3.652 28180
Singapore 318.07 48.90 0.220 1.189 17986
Sri Lanka 62.93 13.94 25.332 5.786 5597
Syria 37.23 37.44 71.498 2.556 17166
Taiwan 94.62 17.92 13.895 8.262 12701
Thailand 51.20 9.45 198.455 26.793 4751
United Arab Emirates 89.66 33.42 32.000 0.694 38190
Yemen 49.34 16.83 128.560 2.369 6425
Austria 81.27 36.64 32.375 3.528 23837
Belgium 151.34 52.46 11.781 4.071 27325
Bulgaria 85.99 31.12 42.823 4.417 9662
Cyprus 107.57 54.39 3.572 0.310 13918
Czechoslovakia 69.45 21.07 49.383 8.137 7467
Denmark 72.99 30.89 16.631 2.780 23861
Finland 57.50 21.64 130.119 2.493 23700
France 47.17 15.26 211.208 24.882 27064
Germany, West 61.52 18.47 96.010 28.085 27252
Greece 53.97 27.01 50.961 3.800 16270
Hungary 82.32 26.92 35.920 5.195 10827
Iceland 81.83 33.08 39.709 0.127 23256
Ireland 118.84 33.85 26.600 1.342 19197
Italy 46.06 13.97 116.500 22.714 27189
Luxembourg 211.94 281.29 0.999 0.157 30782
Malta 160.86 98.14 0.122 0.119 15380
Netherlands 118.76 35.84 16.041 5.855 28563
Norway 86.00 23.54 125.049 2.043 28749
Poland 35.07 13.84 120.728 19.235 8079
Portugal 77.95 18.78 35.550 4.540 11343
Romania 41.62 18.80 91.699 11.275 4021
Spain 43.51 12.38 194.885 13.732 21169
Sweden 69.02 18.22 173.800 4.238 26504
Switzerland 77.69 32.57 15.941 3.222 29848
Turkey 44.40 11.26 300.947 21.829 7091
United Kingdom 56.87 13.47 94.247 27.684 22981
Soviet Union 18.28 3.68 8600.387 142.801 13700
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
398 THE AMERICAN ECONOMIC REVIEW JUNE 1999
TABLE Al-Continued.
Area
Actual trade Constructed (thousands of Population Income per
Country share trade share square miles) (millions) worker
Yugoslavia 57.88 25.82 39.449 10.475 11417
Australia 35.28 4.07 2966.150 7.391 28960
Fiji 89.13 18.56 7.078 0.232 9840
New Zealand 65.25 8.19 103.884 1.438 26039
Papua New Guinea 94.52 10.17 178.704 1.660 3374
Solomon Islands 123.60 25.12 10.954 0.088 5109
Tonga 102.25 43.40 0.288 0.030 6022
Vanuatu 123.33 30.86 4.707 0.042 5707
Western Samoa 92.17 32.77 1.093 0.050 5388
Notes: Actual trade share-Ratio of imports plus exports to GDP, 1985 (Penn World Table, Mark 5.6, Series OPEN).
Constructed trade share-Aggregated fitted values of bilateral trade equation with geographic variables. (See text, Section
I, subsections B-D.)
Area-Rand McNally (1993).
Population-Economically active population, 1985 (Penn World Table, Mark 5.6, constructed from real GDP per capita
(RGDPCH), real GDP per worker (RGDPW), and total population (POP): RGDPCH*POP/RGDPW).
Income per worker-Real GCDP per worker, 1985; 1985 intemational prices (dollars) (Penn World Table, Mark 5.6, Series
RGDPCH).
REFERENCES
.Bradford, Colin, Jr. and Chakwin, Naomi. "Alternative
Explanations of the Trade-Output
Correlation in East Asian Economies."
OECD Development Centre Technical Paper
No. 87, August 1993.
De Long, J. Bradford and Summers, Lawrence H.
"Equipment Investment and Economic
Growth." Quarterly Journal of Economics,
May 1991, 106(2), pp. 445-502.
Dollar, David. "Outward-Oriented Developing
Economies Really Do Grow More Rapidly:
Evidence from 95 LDCs, 1976-1985." Economic
Development and Cultural Change,
April 1992, 40(3), pp. 523-44.
Easterly, William. "How Much Do Distortions
Affect Growth?" Journal of Monetary Economics,
November 1993, 32(2), pp. 187-212.
Edwards, Sebastian. "Openness, Trade Liberalization,
and Growth in Developing Countries."
Journal of Economic Literature.
September 1993, 31(3), pp. 1358-93.
" 'Trade Policy, Exchange Rates, and
Growth," in Rudiger Dombusch and Sebastian
Edwards, eds., Reforn, recovery, and growth:
Latin America and the Middle East. Chicago:
University of Chicago Press, 1995, pp. 13-49.
Feder, Gershon. "On Exports and Economic
Growth." Journal of Development Economics,
February-April 1983, 12(1-2), pp. 59-
73.
Fischer, Stanley. "Growth, Macroeconomics,
and Development," in Olivier Jean Blanchard
and Stanley Fischer, eds., NBER macroeconomics
annual 1991. Cambridge, MA: MIT
Press, 1991, pp. 329-64.
______ "The Role of Macroeconomic Factors
in Growth." Journal of Monetary Economics,
December 1993, 32(3), pp. 485-512.
Frankel, Jeffrey A. Regional trading blocs in the
world trading system. Washington, DC: Institute
of International Economics, 1997.
Frankel, Jeffrey A. and Romer, David. "Trade
and Growth: An Empirical Investigation."
National Bureau of Economic Research
(Cambridge, MA) Working Paper No. 5476,
March 1996.
Frankel, Jeffrey A.; Romer, David and Cyrus,
Teresa L. "Trade and Growth in East Asian
Countries: Cause and Effect?" National
Bureau of Economic Research Working Paper
(Cambridge, MA) No. 5732, August
1996.
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms
VOL. 89 NO. 3 FRANKEL AND ROMER: DOES TRADE CAUSE GROWTH? 399
Frankel, Jeffrey A.; Stein, Ernesto and Wei,
Shang-jin. "Trading Blocs and the Americas:
The Natural, the Unnatural, and the Supernatural."
Journal of Development Economics,
June 1995, 47(1), pp. 61-95.
Hall, Alastair R.; Rudebusch, Glenn D. and Wilcox,
David W. "Judging Instrument Relevance
in Instrumental Variables Estimation." International
Economic Review, May 1996, 37(2),
pp. 283-98.
Hall, Robert E. and Jones, Charles L. "Why Do
Some Countries Produce So Much More Output
per Worker than Others?" Quarterly Journal
of Economics, February 1999, 114(1), pp.
83-116.
Harrison, Ann. "Openness and Growth: A TimeSeries,
Cross-Country Analysis for Developing
Countries." Journal of Development
Economics, March 1996, 48(2), pp. 419--47.
Hausman, Jerry A. "Specification Tests in
Econometrics." Econometrica, November
1978, 46(6), pp. 1251-71.
Helpman, Elhanan. "Growth, Technological
Progress, and Trade." National Bureau of
Economic Research (Cambridge, MA) Reprint
No. 1145, 1988.
Klenow, Peter and Rodriguez-Clare, Andres.
"The Neoclassical Revival in Growth Economics:
Has It Gone Too Far?" in Ben S.
Bernanke and Julio J. Rotemberg, eds.,
NBER macroeconomics annual 1997. Cambridge,
MA: MIT Press, 1997, pp. 73-103.
Kormendi, Roger C. and Meguire, Philip G. "Macroeconomic
Determinants of Growth: Cross-Country
Evidence." Journal of Monetary Economics,
September 1985, 16(2), pp. 141-63.
Lee, Jong-Wha. "International Trade, Distortions
and Long-Run Economic Growth." IMF
Staff Papers, June 1993, 40(2), pp. 299-328.
Levine, Ross and Renelt, David. "A Sensitivity
Analysis of Cross-Country Growth Regressions."
American Economic Review, September
1992, 82(4), pp. 942-63.
Linneman, Hans. An econometric study of international
trade flows. Amsterdam: NorthHolland,
1966.
Mankiw, N. Gregory; Romer, David and Weil,
David N. "A Contribution to the Empirics of
Economic Growth." Quarterly Journal of
Economics, May 1992, 107(2), pp. 407-37.
Michaely, Michael. "Exports and Growth: An
Empirical Investigation." Journal of Development
Economics, March 1977, 4(1), pp.
49-53.
Nelson, Charles R. and Startz, Richard. "Some
Further Results on the Exact Small Sample
Properties of the Instrumental Variable Estimator."
Econometrica, July 1990, 58(4), pp.
967-76.
Rand McNally. Quick reference world atlas.
Chicago: Rand McNally, 1993.
Rodrik, Dani. "Getting Interventions Right:
How South Korea and Taiwan Grew Rich."
Economic Policy, April 1995a, (20), pp.
53-97.
. "Trade and Itidustrial Policy Reform,"in
Jere Behrman and T. N. Srinivasan,
eds., Handbook of development economics,
Vol. 3B. Amsterdam: North-Holland, 1995b,
pp. 2925-82.
Sachs, Jeffrey D. and Warner, Andrew. "Economic
Reform and the Process of Global
Integration." Brookings Papers on Economic
Activity, 1995, (1), pp. 1-95.
Sala-i-Martin, Xavier. "Comment," in Olivier
Jean Blanchard and Stanley Fischer, eds.,
NBER macroeconomics annual 1991. Cambridge,
MA: MIT Press, 1991, pp. 368-
78.
Staiger, Douglas and Stock, James H. "Instrumental
Variables Regression with Weak Instruments."
Econometrica, May 1997, 65(3),
pp. 557-86.
Summers, Robert and Heston, Alan. "The Penn
World Table (Mark 5): An Expanded Set of
International Comparisons, 1950-1988."
Quarterly Journal of Economics, May 1991,
106(2), pp. 327-68.
This content downloaded from 131.211.208.19 on Sun, 31 Dec 2017 15:06:10 UTC
All use subject to http://about.jstor.org/terms