Help Man 2008

THE
QUARTERLY JOURNAL
OF ECONOMICS
Vol. CXXIII May 2008 Issue 2
Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

ESTIMATING TRADE FLOWS: TRADING PARTNERS AND
TRADING VOLUMES*
ELHANAN HELPMAN
MARC MELITZ
YONA RUBINSTEIN
We develop a simple model of international trade with heterogeneous firms

that is consistent with a number of stylized features of the data. In particular, the
model predicts positive as well as zero trade flows across pairs of countries, and
it allows the number of exporting firms to vary across destination countries. As
a result, the impact of trade frictions on trade flows can be decomposed into the
intensive and extensive margins, where the former refers to the trade volume per
exporter and the latter refers to the number of exporters. This model yields a gen-
eralized gravity equation that accounts for the self-selection of firms into export
markets and their impact on trade volumes. We then develop a two-stage estima-
tion procedure that uses an equation for selection into trade partners in the first
stage and a trade flow equation in the second. We implement this procedure para-
metrically, semiparametrically, and nonparametrically, showing that in all three
cases the estimated effects of trade frictions are similar. Importantly, our method
provides estimates of the intensive and extensive margins of trade. We show that
traditional estimates are biased and that most of the bias is due not to selection
but rather due to the omission of the extensive margin. Moreover, the effect of the
number of exporting firms varies across country pairs according to their charac-
teristics. This variation is large and particularly so for trade between developed
and less developed countries and between pairs of less developed countries.
I. INTRODUCTION
Estimation of international trade flows has a long tradi-
tion. Tinbergen (1962) pioneered the use of gravity equations in
* We thank Costas Arkolakis, Robert Barro, Moshe Buchinsky, Zvi Eckstein,
Gene Grossman, Bo Honore, Larry Katz, Marcelo Moreira, Ariel Pakes, Jim Powell,
Manuel Trajtenberg, Zhihong Yu, and three referees for comments. Dror Brenner
and Brent Neiman provided superb research assistance. Helpman thanks the NSF
for financial support. Melitz thanks the NSF and the Sloan Foundation for financial
support and the International Economics Section at Princeton University for its
hospitality.

C 2008 by the President and Fellows of Harvard College and the Massachusetts Institute of
Technology.
The Quarterly Journal of Economics, May 2008
441
442 QUARTERLY JOURNAL OF ECONOMICS
empirical specifications of bilateral trade flows in which the vol-

ume of trade between two countries is proportional to the product
of an index of their economic size, and the factor of proportion-
ality depends on measures of trade resistance between them.
Among the measures of trade resistance, he included geographic
distance, a dummy for common borders, and dummies for Com-
monwealth and Benelux memberships. Tinbergens specification
has been widely used, simply because it provides a good fit to most

data sets of regional and international trade flows. And over time,
his approach has been supplemented with theoretical underpin-
nings and better estimation techniques.1 The gravity equation has
dominated empirical research in international trade; it has been
used to estimate the impact on trade flows of international bor-
ders, preferential trading blocs, currency unions, and membership
in the WTO, as well as the size of home-market effects.2
All the above-mentioned studies estimate the gravity equa-
tion on samples of countries that have only positive trade flows
between them. We argue in this paper that, by disregarding coun-
tries that do not trade with each other, these studies give up impor-
tant information contained in the data, and they produce biased
estimates as a result. We also argue that standard specifications
of the gravity equation impose symmetry that is inconsistent with
the data and that this too biases the estimates. To correct these
biases, we develop a theory that predicts positive as well as zero
trade flows between countries and use the theory to derive esti-
mation procedures that exploit the information contained in data
sets of trading and nontrading countries alike.3
The next section briefly reviews the evolution of the volume of
trade among the 158 countries in our sample and the composition
of country pairs according to their trading status.4 Three features
stand out. First, about half of the country pairs do not trade with
1. See, for example, Anderson (1979), Helpman and Krugman (1985),

Helpman (1987), Feenstra (2002), and Anderson and van Wincoop (2003).
2. See McCallum (1995) for the study that triggered an extensive debate on the
role of international borders, as well as Wei (1996), and Anderson and van Wincoop
(2003), and Evans (2003). Feenstra (2003, Chap. 5) provides an overview of this
debate. Also, see Frankel (1997) on preferential trading blocs, Rose (2000) and
Tenreyro and Barro (2003) on currency unions, Rose (2004) on WTO membership,
and Davis and Weinstein (2003) on the size of home-market effects.
3. Evenett and Venables (2002), Anderson and van Wincoop (2004), and Have-
man and Hummels (2004) all highlight the prevalence of zero bilateral trade flows
and suggest theoretical interpretations for them. We provide a theoretical frame-
work that jointly determines both the set of trading partners and their trade
volumes, and we develop estimation procedures for this model.
4. See Appendix I for data sources.
ESTIMATING TRADE FLOWS 443
one another.5 Second, the rapid growth of world trade from 1970 to
1997 was predominantly due to the growth of the volume of trade
among countries that traded with each other in 1970 rather than
due to the expansion of trade among new trade partners.6 Third,
the average volume of trade at the end of the period between
pairs of countries that exported to one another in 1970 was much
larger than the average volume of trade at the end of the period of
country pairs that did not. Nevertheless, we show in Section VI

that the volume of trade between pairs of countries that traded
with one another was significantly influenced by the fraction of
firms that engaged in foreign trade and that this fraction varied
systematically with country characteristics. Therefore the inten-
sive margin of trade was substantially driven by variations in the
fraction of trading firms but not by new trading partners.7
We develop in Section III the theoretical model that moti-
vates our estimation procedures. This is a model of international
trade in differentiated products in which firms face fixed and vari-
able costs of exporting, along the lines suggested by Melitz (2003).
Firms vary by productivity, and only the more productive firms
find it profitable to export. Moreover, the profitability of exports
varies by destination; it is higher for exports to countries with
higher demand levels, lower variable export costs, and lower fixed
export costs. Positive trade flows from country j to country i thus
aggregate exports over varying distributions of firms. Each dis-
tribution is bounded by a marginal exporter in j who just breaks
even by exporting to i. Country j firms with higher productivity
levels generate positive profits from exports to i.
This model has a number of implications for trade flows. First,
no firm from country j may be productive enough to profitably
export to country i. The model is therefore able to predict zero
exports from j to i for some country pairs. As a result, the model
is consistent with zero trade flows in both directions between some
countries, as well as zero exports from j to i but positive exports
5. We say that a country pair i and j do not trade with one another if i does
not export to j and j does not export to i.
6. Felbermayr and Kohler (2006) report that prior to 1970 new trade flows
contributed substantially to the growth of world trade.
7. The role of the number of exported products, as opposed to exports per prod-
uct, has been found to be important in a number of studies. To illustrate, Hummels
and Klenow (2005) find that 60% of the greater export of larger economies in their
sample of 126 exporting countries is due to variation in the number of exported
products, and Kehoe and Ruhl (2002) find that during episodes of trade liberal-
ization in 18 countries a large fraction of trade expansion was driven by trade in
goods that were not traded before.
from i to j for some country pairs. Both types of trade patterns

exist in the data. Second, the model predicts positivethough
asymmetrictrade flows in both directions for some country pairs,
which are also needed to explain the data. And finally, the model
generates a gravity equation.
Our derivation of the gravity equation generalizes the
Anderson and van Wincoop (2003) equation in two ways. First,
it accounts for firm heterogeneity and fixed trade costs and thus

predicts an extensive margin for trade flows. Second, it accounts
for asymmetries between the volume of exports from j to i and the
volume of exports from i to j. Both are important for data analysis.
We also develop a set of sufficient conditions under which more
general forms of the Andersonvan Wincoop equations aggregate
trade flows across heterogeneous firms facing both fixed and vari-
able trade costs.
Section IV develops the empirical framework for estimating
the gravity equation derived in Section III. We propose a two-
stage estimation procedure. The first stage consists of estimating
a Probit equation that specifies the probability that country j ex-
ports to i as a function of observable variables. The specification of
this equation is derived from the theoretical model and an explicit
introduction of unobservable variations. Predicted components of
this equation are then used in the second stage to estimate the
gravity equation in log-linear form. We show that this procedure
yields consistent estimates of the parameters of the gravity equa-
tion, such as the marginal impact of distance between countries
on their exports to one another.8 It simultaneously corrects for
two types of potential biases: a sample selection bias and a bias
from potential asymmetries in the trade flows between pairs of
countries. The latter bias is due to an omitted variable that mea-
sures the impact of the number (fraction) of exporting firms, that
is, the extensive margin of trade. Because this procedure is easy
to implement, it can be effectively used in many applications.
Our theoretical model has firm heterogeneity, yet we do not
need firm-level data to estimate the gravity equation. This prop-
erty results from the fact that the characteristics of the marginal
exporters to different destinations can be identified from the vari-
ation in features of the destination countries and of observable
bilateral trade costs. As a result, there exist sufficient statistics,
8. We also show that consistency requires the use of separate country fixed
effects for exporters and importers, as proposed by Feenstra (2002).
which can be computed from aggregate data, that predict the se-
lection of heterogeneous firms into export markets and their asso-
ciated aggregate trade volumes.9 This is an important advantage
of our approach, which extracts from country-level data informa-
tion that would normally require firm-level data. Although more
firm-level data sets have become available over time, it is not yet
possible to pool them together into a comprehensive data set that
can be used for cross-country estimation purposes.

Section V shows that variables that are commonly used in
gravity equations also affect the probability that two countries
trade with each other. This provides evidence for a potential bias
in the standard estimates. The extent of this bias is then studied
in Sections VI and VII. In Section VI, we estimate the model on a
partial sample of countries for which we have data on regulation
of entry costs, which we use as the excluded variables in the two-
stage estimation procedure. We argue that these variables satisfy
the exclusion restrictions on theoretical grounds. In Section VII,
we use this reduced sample to test for the validity of other po-
tential excluded variables, which are available for virtually all
country pairs, representing a substantial increase in sample size.
We show that an index for common religion (across country pairs)
satisfies the exclusion restrictions for this sample. We then reesti-
mate our model on the full sample of countries using this common
religion index as the excluded variable. This approximately dou-
bles the number of usable observations. This substantial increase
in sample size is the main motivation behind our construction of
the religion variable in the first place.
In both Sections VI and VII, we implement three estimation
methods, progressively relaxing some parameterization assump-
tions: nonlinear least squares, semiparametric, and nonparame-
tric. The nonlinear least squares (NLS) version of the two-stage
procedure uses functional forms derived from the theoretical
model under the assumption that productivity follows a truncated
Pareto distribution. We show that the corrections for the selection
9. Eaton and Kortum (2002) apply a similar principle to determine an aggre-

gate gravity equation across heterogeneous Ricardian sectors. As in our model,
the predicted trade volume reflects an extensive margin (number of sectors/goods
traded) and an intensive one (volume of trade per good/sector). However, Eaton
and Kortum do not model fixed trade costs and the possibility of zero bilateral
trade flows. Unlike our equations, theirs are subject to the criticism raised by
Haveman and Hummels (2004). Bernard et al. (2003) use direct information
on U.S. plant-level sales, productivity, and export status to calibrate a model
that is then used to simulate the extensive and intensive margins of bilateral
trade flows.
and omitted variable biases have a measurable downward impact

on the estimated coefficients. Moreover, the extent of this bias
is not sensitive to the use of the alternative excluded variables.
The nature and extent of this bias is further confirmed when we
estimate the model in the other two alternative ways: first with a
semiparametric method, where we replace the truncated Pareto
distribution for firm productivity with a general distribution ap-
proximated by a polynomial fit, and second with a nonparametric

method, which further relaxes the joint normality assumption for
the unobserved trade costs. In both cases, we obtain results very
similar to our fully parametrized NLS specification. An additional
advantage of the latter two methods is that they can be easily
implemented using OLS in the second stage.
A number of additional insights from our estimates are dis-
cussed in Section VIII. First, we show that most of the bias is due
to the omitted correction for the extensive margin of trade and
not due to the selection bias. In fact, the selection bias is economi-
cally negligible though statistically (strongly) significant. Second,
we show that the asymmetric impact of the extensive margin of
trade is important in explaining the asymmetries in trade flows
observed in the data. Finally, we show that the biases not only
are large, but also systematically vary with the characteristics of
trade partners. For this purpose we perform a counterfactual ex-
ercise in which trade frictions are reduced. A reduction in these
frictions induces trade among country pairs that did not trade be-
fore and raises trade volumes among country pairs with existing
trade relations. When countries are partitioned by income (high
versus low), we find that the impact of reduced trade frictions dif-
fers substantially across country pairs according to these income
levels. The elasticity of trade with respect to such frictions can
vary by a factor of three. That is, it can be three times larger for
some country pairs than for others. This highlights both the size,
and also the large variations in the biases across country pairs.
Section IX concludes.
II. A GLANCE AT THE DATA

Figure I depicts the empirical extent of zero trade flows. In
this figure, all possible country pairs are partitioned into three
categories. The top portion represents the fraction of country pairs
that do not trade with one another; the bottom portion represents
those that trade in both directions (they export to one another);

FIGURE I
Distribution of Country Pairs Based on Direction of Trade
Note. Constructed from 158 countries.
and the middle portion represents those that trade in one direction
only (one country imports from, but does not export to, the other
country). As is evident from the figure, by disregarding countries
that do not trade with each other or trade only in one direction,
one disregards close to half of the observations. We show below
that these observations contain useful information for estimating
international trade flows.10
Figure II shows the evolution of the aggregate real volume of
exports of all 158 countries in our sample and of the aggregate
real volume of exports of the subset of country pairs that exported
to one another in 1970. The difference between the two curves
represents the volume of trade of country pairs that either did not
trade or traded in one direction only in 1970. It is clear from this
figure that the rapid growth of trade, at an annual rate of 7.5%
on average, was mostly driven by the growth of trade between
countries that traded with each other in both directions at the
beginning of the period. In other words, the contribution to the
10. Silva and Tenreyro (2006) also argue that zero trade flows can be used in
the estimation of the gravity equation, but they emphasize a heteroscedasticity
bias that emanates from the log-linearization of the equation rather than the
selection and asymmetry biases that we emphasize. Moreover, the Poisson method
that they propose to use yields similar estimates on the sample of countries that
have positive trade flows in both directions and the sample of countries that have
positive and zero trade flows. This finding is consistent with our finding that the
selection bias is rather small.

FIGURE II
Aggregate Volume of Exports of All Country Pairs and of Country Pairs That
Traded in Both Directions in 1970
growth of trade of countries that started to trade after 1970 in

either one or both directions was relatively small.
Combining this evidence with the evidence from Figure I,
which shows a relatively slow growth of the fraction of trading
country pairs, suggests that bilateral trading volumes of coun-
try pairs that traded with one another in both directions at the
beginning of the period must have been much larger than the bi-
lateral trading volumes of country pairs that either did not trade
with each other or traded in one direction only at the beginning of
the period. Indeed, at the end of the period the average bilateral
trade volume of country pairs of the former type was about 35
times larger than the average bilateral trade volume of country
pairs of the latter type. This suggests that the enlargement of the
set of trading countries did not contribute in a major way to the
growth of world trade.11
11. This contrasts with the sector-level evidence presented by Evenett and
Venables (2002). They find a substantial increase in the number of trading partners
at the three-digit sector level for a selected group of 23 developing countries. We
conjecture that their country sample is not representative and that most of their
new trading pairs were originally trading in other sectors. And this also contrasts
III. THEORY
Consider a world with J countries, indexed by j = 1, 2, . . . , J.
Every country consumes and produces a continuum of products.
Country js utility function is
1/
uj = x j (l) dl , 0<<1,
lB j

where x j (l) is its consumption of product l and B j is the set of
products available for consumption in country j. The parameter
determines the elasticity of substitution across products, which
is = 1/(1 ). This elasticity is the same in every country.
Let Y j be the income of country j, which equals its expendi-
ture level. Then country js demand for product l is
p j (l) Y j
(1) x j (l) = ,
Pj1
where p j (l) is the price of product l in country j and Pj is the

countrys ideal price index, given by
1/(1)
(2) Pj = p j (l)1 dl .
lB j
This specification implies that every product has a constant de-

mand elasticity .
Some of the products consumed in country j are domestically
produced while others are imported. Country j has a measure N j
of firms, each one producing a distinct product. The products pro-
duced by country- j firms are also distinct from the products pro-
duced by country-i firms for i = j. As a result, there are Jj=1 N j
products in the world economy.
A country- j firm produces one unit of output with a cost-
minimizing combination of inputs that cost c j a, where a mea-
sures the number of bundles of the countrys inputs used by the
firm per unit output and c j measures the cost of this bundle. The
cost c j is country-specific, reflecting differences across countries
in factor prices, whereas a is firm-specific, reflecting productivity
with the finding that changes in the number of trading products has a measurable
impact on trade flows (see Kehoe and Ruhl [2002] and Hummels and Klenow
[2005]).
differences across firms in the same country. The inverse of a, 1/a,

represents the firms productivity level.12 We assume that a cu-
mulative distribution function G(a) with support [aL, aH ] describes
the distribution of a across firms, where aH > aL > 0. This distri-
bution function is the same in all countries.13
We assume that a producer bears only production costs when
selling in the home market. That is, if a country- j producer with
coefficient a sells in country j, the delivery cost of its product is

c j a. If, however, this same producer seeks to sell its product in
country i, there are two additional costs it has to bear: a fixed
cost of serving country i, which equals c j fi j , and a transport cost.
As is customary, we adopt the melting iceberg specification and
assume that i j units of a product have to be shipped from country
j to i for one unit to arrive. We assume that f j j = 0 for every j
and fi j > 0 for i = j, and j j = 1 for every j and i j > 1 for i = j.
Note that the fixed cost coefficients fi j and the transport cost
coefficients i j depend on the identity of the importing and export-
ing countries, but not on the identity of the exporting producer.
In particular, they do not depend on the producers productivity
level.
There is monopolistic competition in final products. Because
every producer of a distinct product is of measure zero, the de-
mand function (1) implies that a country- j producer with an
input coefficient a maximizes profits by charging the mill price
p j (a) = c j a/. This is a standard markup pricing equation, with a
smaller markup associated with a larger elasticity of demand. If
this country- j producer of a product l sells to consumers in country
i, it then sets a delivered price (in country i) equal to
cja
(3) p j (l) = i j .

As a result, the associated operating profits from these sales to
country i are
1
i j c j a
i j (a) = (1 ) Yi c j fi j .
Pi
12. See Melitz (2003) for a discussion of a general equilibrium model of trad-
ing countries in which firms are heterogeneous in productivity. We follow his
specification.
13. The as only capture relative productivity differences across firms in a
country. Aggregate productivity differences across countries are subsumed in the
c j s.
Evidently, these operating profits are positive for sales in the do-
mestic market because f j j = 0. Therefore all N j producers sell in
country j. But sales in country i = j are profitable only if a ai j ,
where ai j is defined by i j (ai j ) = 0, or14
1
i j c j ai j
(4) (1 ) Yi = c j fi j .
Pi

It follows that only a fraction G(ai j ) of country js N j firms ex-
port to country i. For this reason the set Bi of products available
in country i is smaller than the total set of products produced
in the world economy. In addition, it is possible for G(ai j ) to be
zero: no firm from country j finds it profitable to export to country
i. This happens whenever ai j aL: the least productive firm that
can profitably export to country i has a coefficient a below the sup-
port of G(a). We explicitly consider these cases that explain zero
bilateral trade volumes. If ai j were larger than aH , then all firms
from country j would export to i. However, given the pervasive
firm-level evidence on the coexistence of exporting and nonex-
porting firms, even within narrowly defined sectors, we disregard
this possibility.
We next characterize bilateral trade volumes. Let
a
ij
aL a
1
dG (a) for ai j aL
(5) Vi j =
0 otherwise.
The demand function (1) and pricing equation (3) then imply that
the value of country is imports from j is
1
c j i j
(6) Mi j = Yi N j Vi j .
Pi
This bilateral trade volume equals zero when ai j aL, because

Vi j = 0 under these circumstances. Using the definition of Vi j and
(2), we also obtain
c j i j 1
(7) Pi1 = N j Vi j .

j=1
14. Note that ai j + as fi j 0.

Equations (4)(7) provide a mapping from the income levels Yi ,

the numbers of firms Ni , the unit costs ci , the fixed costs fi j , and
the transport costs i j to the bilateral trade flows Mi j .
We show in Appendix II that, together with equality of in-
come and expenditure, equations (4)(7) can be used to derive a
generalized version of Anderson and van Wincoops (2003) gravity
equation with third-country effects. This generalization applies
when transport costs are symmetric (i j = ji i, j) and Vi j can

be multiplicatively decomposed into three components: one that
depends only on importer characteristics, a second that depends
only on exporter characteristics, and a third that depends on the
country pair characteristics but is symmetric for that country pair.
This decomposability holds in Anderson and van Wincoops model.
Importantly, however, there are other cases of interest with pos-
itive fixed export costs and an extensive margin of trade that
also satisfy the generalized gravity equation. Yet even this more
generalized version of the gravity equation cannot explain the
documented pattern of zero trade flows and the bilateral trade
asymmetries (see Appendix II for details). Thus, in order to gain
as much flexibility as possible in the empirical application, we
develop in the next section an estimation procedure that builds
directly on equations (4)(7), which allow for asymmetric bilateral
trade flows, including zeros.
IV. EMPIRICAL FRAMEWORK

We begin by formulating a fully parametrized estimation pro-
cedure for this model, which delivers our benchmark results. We
then progressively loosen these parametric restrictions and re-
estimate the model. In all cases, we obtain similar results that
are consistent with the analysis of the baseline scenario.
In the baseline specification, we assume that firm productiv-
ity 1/a is Pareto distributed, truncated to the support [aL, aH ].
Thus, we assume G(a) = (ak aLk )/(akH aLk ), k > ( 1). As previ-
ously highlighted, we allow for ai j < aL for some i j pairs, induc-
ing zero exports from j to i (i.e., Vi j = 0 and Mi j = 0). This frame-
work also allows for asymmetric trade flows, Mi j = M ji , which
may also be unidirectional, with M ji > 0 and Mi j = 0, or M ji = 0
and Mi j > 0. Such unidirectional trading relationships are empir-
ically common and can be predicted using our empirical method.
Moreover, asymmetric trade frictions are not necessary to induce
such asymmetric trade flows when productivity is drawn from a
truncated Pareto distribution.
Our assumptions imply that Vi j can be expressed as (see (5))
kaLk+1
Vi j = Wi j ,
(k + 1) akH aLk
where
k+1
ai j
(8) Wi j = max 1, 0 ,
aL

and ai j is determined by the zero profit condition (4). Note that
both Vi j and Wi j are monotonic functions of the proportion of ex-
porters from j to i, G(ai j ). The export volume from j to i, given by
(6), can now be expressed in log-linear form as
mi j = ( 1) ln ( 1) ln c j + n j + ( 1) pi + yi
+ (1 ) ln i j + vi j ,
where lowercase variables represent the natural logarithms of

their respective uppercase variables. i j captures variable trade
costs: costs that affect the volume of firm-level exports. We as-
sume that these costs are stochastic due to i.i.d. unmeasured
trade frictions ui j , which are country-pair specific. In particular,

let i1
j Di j eui j , where Di j represents the (symmetric) distance
between i and j, and ui j N(0, u2 ).15 Then the equation of the
bilateral trade flows mi j yields the estimating equation
(9) mi j = 0 + j + i di j + wi j + ui j ,
where i = ( 1) pi + yi is a fixed effect of the importing coun-

try and j = ( 1) ln c j + n j is a fixed effect of the exporting
country.16
Equation (9) highlights several important differences with
the gravity equation, as derived, for example, by Anderson and
van Wincoop (2003). The most important difference is the addi-
tion in our formulation of the new variable wi j , which controls for
the fraction of firms (possibly zero) that export from j to i. This
15. In the following derivations, we use distance as the only source of ob-
servable variable trade costs. It should nevertheless be clear how this approach
generalizes to a matrix of observable bilateral trade frictions paired with a vector
of elasticities .
16. We replace vi j with wi j , and therefore 0 now also contains the log of
the constant multiplier in Vi j . If tariffs are not directly controlled for, then the
importers fixed effect will subsume an average tariff level. Similarly, average
export taxes will show up in the exporters fixed effect.
variable is a function of the cutoff ai j , which is determined by

other explanatory variables (see (4)). When wi j is not included on
the right-hand side, the coefficient on distance (or any other co-
efficient on a potential trade barrier) can no longer be interpreted
as the elasticity of a firms trade with respect to distance (or other
trade barriers), which is the way in which such trade barriers are
almost always modeled in the literature that follows the new
trade theory. Instead, the estimation of the standard gravity equa-

tion confounds the effects of trade barriers on firm-level trade with
their effects on the proportion of exporting firms, which induces
an upward bias in the estimated coefficient .
Another bias is introduced into the estimation of equation
(9) when country pairs with zero trade flows are excluded. This
selection effect induces a positive correlation between the unob-
served ui j s and the trade barrier, di j s; country pairs with large
observed trade barriers (high di j ) that trade with each other are
likely to have low unobserved trade barriers (high ui j ). Although
this induces a downward bias in the trade barrier coefficient, our
empirical results show that this effect is dominated by the upward
bias generated by the endogenous number of exporters.
Last, we emphasize again that in our formulation, bilateral
trade flows need not be balanced, even when all bilateral trade
barriers are symmetric. First and foremost, wi j can be asymmet-
ric. We document later in Section VIII that such asymmetries are
empirically important and substantial. Second, the importer fixed
effects may differ from the exporter fixed effects for given coun-
tries. This substantiates the use of directional trade flows and sep-
arate fixed effects for the exporting and the importing countries.
IV.A. Firm Selection into Export Markets

The selection of firms into export markets, represented by
the variable Wi j , is determined by the cutoff value of ai j , which
is implicitly defined by the zero profit condition (4). We define a
related latent variable Zi j as

1
(1 ) Pi c ji j Yi aL1
(10) Zi j = .
c j fi j
This is the ratio of variable export profits for the most pro-
ductive firm (with productivity 1/aL) to the fixed export costs
(common to all exporters) for exports from j to i. Positive ex-
ports are observed if and only if Zi j > 1. In this case Wi j is a
(k+1)/(1)
monotonic function of Zi j ; that is, Wi j = Zi j 1 (see (4)
and (8)). As with the variable trade costs i j , we assume that
the fixed export costs fi j are stochastic due to unmeasured trade
frictions i j that are i.i.d., but may be correlated with the ui j s.
Let fi j exp( EX, j + I M,i + i j i j ), where i j N(0, 2 ), I M,i
is a fixed trade barrier imposed by the importing country on all
exporters, EX, j is a measure of fixed export costs common across
all export destinations, and i j is an observed measure of any ad-

ditional country-pair specific fixed trade costs.17 Using this speci-
fication together with ( 1) ln i j di j ui j , the latent variable
zi j ln Zi j can be expressed as
(11) zi j = 0 + j + i di j i j + i j ,
where i j ui j + i j N(0, u2 + 2 ) is i.i.d. (yet correlated with
the error term ui j in the gravity equation), j = ln c j + EX, j
is an exporter fixed effect, and i = ( 1) pi + yi I M,i is an im-
porter fixed effect. Although zi j is unobserved, we observe the
presence of trade flows. Therefore zi j > 0 when j exports to i, and
zi j = 0 when it does not. Moreover, the value of zi j affects the
export volume.
Define the indicator variable Ti j to equal 1 when country j
exports to i and 0 when it does not. Let i j be the probability
that j exports to i, conditional on the observed variables. Because
we do not want to impose 2 u2 + 2 = 1, we divide (11) by the
standard deviation and specify the Probit equation
i j = Pr(Ti j = 1 | observed variables)

(12) = 0 + j + i di j i j ,
where () is the cdf of the unit-normal distribution, and every
starred coefficient represents the original coefficient divided by
.18 Importantly, this selection equation has been derived from
a firm-level decision, and it therefore does not contain the unob-
served and endogenous variable Wi j that is related to the fraction
of exporting firms. Moreover, the Probit equation can be used to
derive consistent estimates of Wi j .
17. As with variable trade costs, it should be clear how this derivation can be
extended to a vector of observable fixed trade costs.
18. By construction, the error term ij i j / is distributed unit normal. The
Probit equation (12) distinguishes between observable trade barriers that affect
variable trade costs (di j ) and fixed trade costs ( fi j ). In practice, some variables
may affect both. Their coefficients in (12) then capture the combined effect of these
barriers.
Let i j be the predicted probability of exports from j to i, using

the estimates from the Probit equation (12), and let zij = 1 (i j )
be the predicted value of the latent variable zij zi j / . Then a
consistent estimate for Wi j can be obtained from

(13) Wi j = max Zij 1, 0 ,
where (k + 1)/( 1).

IV.B. Consistent Estimation of the Log-Linear Equation
Consistent estimation of (9) requires controls for both the en-
dogenous number of exporters (via wi j ) and the selection of country
pairs into trading partners (which generates a correlation between
the unobserved ui j and the independent variables). We thus need
estimates for E[wi j | ., Ti j = 1] and E[ui j | ., Ti j = 1]. Both terms
depend on ij E[ij | ., Ti j = 1]. Moreover, E[ui j | ., Ti j = 1] =
corr (ui j , i j )(u/ )ij . Since ij has a unit normal distribution,
a consistent estimate ij is obtained from the inverse Mills ratio,
that is, ij = (zij )/(zij ). Therefore z ij zij + ij is a consistent
estimate for E[zij | ., Ti j = 1] and w ij ln{exp[(zij + ij )] 1} is
a consistent estimate for E[wi j | ., Ti j = 1] (see (13)). We therefore
can estimate (9) using the transformation

mi j = 0 + j + i di j + ln exp zij + ij 1 + u ij + ei j ,
(14)
where u corr (ui j , i j )(u/ ) and ei j is an i.i.d. error term sat-

isfying E[ei j | ., Ti j = 1] = 0. Because (14) is nonlinear in , we
estimate it using nonlinear least squares.
The use of ij to control for E[ui j | ., Ti j = 1] is the standard
Heckman (1979) correction for sample selection. This addresses
the biases generated by the unobserved country-pair level shocks
ui j and i j . However, this does not correct for the biases generated
by the underlying unobserved firm-level heterogeneity. The lat-
ter biases are corrected by the additional control zij (along with
the functional form determined by our theoretical assumptions).
Used alone, the standard Heckman (1979) correction would only
be valid in a world without firm-level heterogeneity, or where
such heterogeneity was not correlated with the export decision.
Thus, all firms are identically affected by trade barriers and coun-
try characteristics and make the same export decisionsor make
export decisions that are uncorrelated with trade barriers and
country characteristics. This misses the potentially important ef-

fect of trade barriers and country characteristics on the share of
exporting firms. In a world with firm-level heterogeneity, a larger
fraction of firms export to more attractive export destinations.19
Our empirical results highlight the overwhelming contribution of
this channel relative to the standard correction for sample selec-
tion, which ignores firm-level heterogeneity.
To summarize, our theoretical framework delivers two equa-

tions, (11) and (14), which can be estimated in two stages.
Although the theoretical model allows for arbitrary varia-
tion in bilateral variable and fixed trade costs, for estima-

tion purposes we restrict these variations to i1 j Di j eui j
and fi j exp( EX, j + I M,i + i j i j ), respectively. These re-
strictions make it possible to identify and , which are important
parameters, but they do not make it possible to infer every param-
eter of the model. For example, we cannot separately identify the
elasticity of demand . Evidently, it is necessary to impose more
restrictions in order to gain additional identification.20
Before describing the empirical results, we pause to note that
our distributional assumptions on the joint normality of the un-
observed trade costs and the Pareto distribution of firm-level pro-
ductivity affect the functional form of the trade flow equation (14)
via the functional form of the two additional controls for firm het-
erogeneity (w ij ) and sample selection ( ij ). After presenting our
main results, we will describe a number of alternative specifi-
cations that relax these assumptions, yet generate very similar
estimates. They illustrate the robustness of the findings in our
baseline specification.
V. TRADITIONAL ESTIMATES
Traditional estimates of the gravity equation use data on
country pairs that trade in at least one direction. The first col-
umn in Table I provides a representative estimate of this sort for
all bilateral trade flows reported in 1986 from a set of 158 coun-
tries (the full list is reported in Appendix I). Note that instead of
constructing symmetric trade flows by combining exports and im-
ports for each country pair, we use the unidirectional trade value
19. Eaton, Kortum, and Kramarz (2004) find that more French firms export
to larger foreign markets, and Bernard, Jensen, and Schott (2005) find a similar
pattern for U.S. firms. Our model is consistent with these findings.
20. See, for example, Eaton and Kortum (2002) and Anderson and van
Wincoop (2003) for ways to estimate this elasticity.
TABLE I
BENCHMARK GRAVITY AND SELECTION INTO TRADING RELATIONSHIPS
1986 1980s
(Probit) (Probit) (Probit)

Variables mi j Ti j mi j Ti j mi j Ti j
Distance 1.176 0.263 1.201 0.246 1.200 0.246

(0.031) (0.012) (0.024) (0.008) (0.024) (0.008)
0.458 0.148 0.366 0.146 0.364 0.146

Land border
(0.147) (0.047) (0.131) (0.032) (0.131) (0.032)
Island 0.391 0.136 0.381 0.140 0.378 0.140
(0.121) (0.032) (0.096) (0.022) (0.096) (0.022)
Landlock 0.561 0.072 0.582 0.087 0.581 0.087
(0.188) (0.045) (0.148) (0.028) (0.147) (0.028)
Legal 0.486 0.038 0.406 0.029 0.407 0.028
(0.050) (0.014) (0.040) (0.009) (0.040) (0.009)
Language 0.176 0.113 0.207 0.109 0.203 0.108
(0.061) (0.016) (0.047) (0.011) (0.047) (0.011)
Colonial ties 1.299 0.128 1.321 0.114 1.326 0.116
(0.120) (0.117) (0.110) (0.082) (0.110) (0.082)
Currency union 1.364 0.190 1.395 0.206 1.409 0.206
(0.255) (0.052) (0.187) (0.026) (0.187) (0.026)
FTA 0.759 0.494 0.996 0.497 0.976 0.495
(0.222) (0.020) (0.213) (0.018) (0.214) (0.018)
Religion 0.102 0.104 0.018 0.099 0.038 0.098
(0.096) (0.025) (0.076) (0.016) (0.077) (0.016)
WTO (none) 0.068 0.056
(0.058) (0.013)
WTO (both) 0.303 0.093
(0.042) (0.013)
Observations 11,146 24,649 110,697 248,060 110,697 248,060
R2 0.709 0.587 0.682 0.551 0.682 0.551
Notes. Exporter, importer, and year fixed effects. Marginal effects at sample means and pseudo R2 reported
for Probit. Robust standard errors (clustering by country pair).
+ Significant at 10%.
Significant at 5%.
Significant at 1%.
and introduce both importing and exporting country fixed effects.

With these fixed effects every country pair is represented twice:
one time for exports from i to j and another time for exports from
j to i.21 Nevertheless, the results in Table I are similar to those
obtained with symmetric trade flows and a unique country fixed
effect. They show that country j exports more to country i when
the two countries are closer to each other, they both belong to the
21. Among the 158 157 = 24,806 possible bilateral trading relationships,
there are only 11,146 (less than half) positive trade flows.
same regional free trade agreement (FTA), they share a common

language, they have a common land border, they are not islands,
they share the same legal system, they share the same currency,
or one country has colonized the other. The probability that two
randomly drawn persons, one from each country, share the same
religion raises export volumes.22 Details on the construction of all
the variables are provided in Appendix I.
We next estimate a Probit equation for the presence of a

trading relationship using the same explanatory variables as the
initial gravity specification (the specification follows (12), with ex-
porter and importer fixed effects). The marginal effects, evaluated
at the sample means, are reported in column (2).23 These results
clearly show that the very same variables that impact export vol-
umes from j to i also impact the probability that j exports to i. In
almost all cases, the impact goes in the same direction. The effect
of a common border is the only exception: it raises the volume
of trade but reduces the probability of trading. We attribute this
finding to the effect of territorial border conflicts that suppress
trade between neighbors. In the absence of such conflicts, common
land borders enhance trade. We also note that a common religion
strongly affects the formation of trading relationships (its effect is
similar to that of a common language, increasing the probability
of trade by 10% for the typical country pair). Overall, this evi-
dence strongly suggests that disregarding the selection equation
of trading partners biases the estimates of the export equation, as
we have argued in Section IV.
These results and their consequences are not specific to 1986.
We repeat the same regressions increasing the sample years to
cover all of the 1980s, adding year fixed effects. The results in
columns (3) and (4) are very similar to those in the first two
columns. As expected, the standard errors are reduced (all stan-
dard errors are robust to clustering by country pairs). Adding
the time variation also allows the identification of the effects of
changing country characteristics. We use this additional source
of variation to investigate the effects of WTO/GATT member-
ship (hereafter summarized as WTO) on trade volumes as well as
the formation of bilateral trade relationships. We thus repeat the
22. The common religion variable is not used in traditional gravity equations.
We have constructed it especially for use in our two-stage estimation procedure,
as explained in the following sections.
23. The sample size is reduced from 158 157 = 24,806 to 24,649 because
Congo did not export to anyone in 1986, and an exporter fixed effect cannot be
estimated.
same regressions for the 1980s, adding bilateral controls when-

ever both countries or neither country is a member of WTO. As
emphasized by Subramanian and Wei (2007), the use of unidirec-
tional trade data and separate exporter and importer fixed effects
substantially increases the statistically significant positive effect
of WTO membership on trade volumes.24 Our theoretical frame-
work provides a justification for this estimation strategy when
bilateral trade flows are asymmetric. Furthermore, we also find

that WTO membership has a very strong and significant effect on
the formation of bilateral trading relationships. The coefficients
in column (6) show that, for any country pair, joint WTO member-
ship has an impact on the probability of trade similar to common
language or colonial ties.25
In reporting results for the 1980s, we aim to show that our
choice of 1986 for the cross-section study does not affect the
estimates. In other words, there is nothing special about 1986.
And moreover, because this is mostly a methodological paper, we
do not think that the choice of year is particularly important. Yet
1986 has the added advantage that it allows us to compare our
results with French firm-level export data by destination reported
in Eaton, Kortum, and Kramarz (2004) (see below).
VI. TWO-STAGE ESTIMATION

We now turn to the second stage estimation of the trade
flow equation (14). As we describe in Section IV, this requires
a first-stage Probit selection equation (12) such as that reported
in Table I, which yields a predicted probability of export i j (and
thus the additional w ij and ij controls). Because we do not want
the identification of our second stage estimates to rely on the
normality assumption for the unobserved trade costs, we also
need to select valid excluded variables for that second stage (we
will also relax these distributional assumptions through the use
of nonparametric methods). Our theoretical model suggests that
trade barriers that affect fixed trade costs but do not affect vari-
able (per-unit) trade costs satisfy this exclusion restriction. We
now describe the construction of such variables.
24. Rose (2004) reports a significant though smaller effect of WTO member-
ship on trade volumes using symmetric trade flow data and a unique set of country
fixed effects.
25. When two countries both join the WTO, their probability of trade increases
by 15%.
We start with country-level data on the regulation costs of

firm entry, collected and analyzed by Djankov et al. (2002). These
entry costs are measured via their effects on the number of days,
the number of legal procedures, and the relative cost (as a per-
centage of GDP per capita) for an entrepreneur to legally start op-
erating a business.26 We surmise (and confirm empirically) that
they also affect the costs faced by exporting firms to/from that
country, and that these costs are magnified when both export-

ing and importing countries impose high regulatory hurdles. By
their nature, these measures affect firm-level fixed rather than
variable costs of trade. We therefore construct an indicator for
high fixed-cost trading country pairs, consisting of country pairs
in which both the importing and exporting countries have en-
try regulation measures above the cross-country median. One
variable uses the sum of the number of days and procedures
above the median (for both countries) whereas the other uses the
sum of the relative costs above the median (again for both coun-
tries).27 By construction, these bilateral variables reflect regula-
tion costs that should not depend on a firms volume of exports to
a particular country, and therefore satisfy the requisite exclusion
restrictions.28
Using these additional variables for our first stage estima-
tion of selection into trading relationships entails a substantial
drop in sample size. First, 42 of 158 countries do not have any
available regulation cost data.29 Second, among the remaining
countries, 8 of them export to everyone, and Japan imports from
26. Unfortunately, historic data were not available. For this reason, we use
the data for 1999. See Djankov et al. (2002) for details.
27. Recall that these relative costs are measured as a percentage of GDP
per capita, so these cost measures can be compared across countries. We could
also have separated the number of days and procedures into separate variables,
but we found that the jointly defined indicator variable has substantially more
explanatory power.
28. Variable (per-unit) export costs at the country level could potentially be
correlated with the fixed regulation costs associated with trade. However, our first
stage estimation also includes country fixed effects. These correlated country-level
variable costs would then have to interact in the same pattern as the fixed costs
across country pairs in order to generate a correlation at the country level that is
left uncontrolled by the country fixed effects. This possibility is substantially more
remote than the potential correlation at the country level.
29. These 42 countries are Afghanistan, Bahamas, Bahrain, Barbados, Belize,
Bermuda, Brunei, Cayman Islands, Comoros, Cuba, Cyprus, Djibouti, Equatorial
Guinea, French Guiana, Gabon, Gambia, Greenland, Guadeloupe, Guinea-Bissau,
Guyana, Iceland, Iraq, Kiribati, North Korea, Liberia, Libya, Maldives, Malta,
Mauritius, Myanmar, New Caledonia, Qatar, Reunion, Seychelles, Somalia, St.
Kitts, Sudan, Suriname, Trinidad-Tobago, Turks Caicos, Western Sahara, and
Zaire.
everyone.30 Fixed exporter (and in the case of Japan, importer) ef-

fects can thus not be estimated, and all the observations with that
particular exporter (or importer) are dropped. Third, the num-
ber of observations decreases with the square of the number of
dropped countries. Jointly, these factors account for the halving
of the available observations. This substantial decrease has led
us to statistically test the validity of the exclusion restriction for
additional bilateral trade barriers available for our full sample of

countries (see following section). For now, the most relevant issue
for our estimation purposes is that the additional cost variables
have substantial explanatory power for the formation of trading
relationships. This is strongly confirmed by the results in the first
column of Table II. We reran the same Probit equation (based on
(12)) as previously reported in Table I, adding our two cost mea-
sures. The results for all the explanatory variables from Table I
are roughly similar, and the two cost variables are economically
and statistically significant.
We next estimate our fully parametrized trade flow equa-
tion (14) using nonlinear least squares (NLS). We use the esti-
mates of the Probit equation for the reduced sample to construct
ij = (zij )/(zij ) and w ij () = ln{exp[(zij + ij )] 1} for all coun-
try pairs with positive trade flows.31 The former controls for the
sample selection bias whereas the latter controls for unobserved
firm heterogeneity, that is, the effect of trade frictions and coun-
try characteristics on the proportion of exporters. We first report
the results from a benchmark gravity equation without these con-
trols in the second column of Table II and then report our NLS
results in the third column. The standard errors are bootstrapped
based on sampling (500 times) all available countries with replace-
ment and using all the potential country pairs from that country
sample. Both the nonlinear coefficient for w ij and the linear
30. These 8 countries are Japan, Hong Kong, France, Germany, Italy, the
Netherlands, the United Kingdom, and Sweden.
31. Recall that zij = 1 (i j ). The characteristics of our data induces a compli-
cation associated with this transformation: Our sample includes a relatively small
number of country pairs whose characteristics are such that their probability of
trade i j is indistinguishable from 1. We therefore cannot infer any differences in
the zij s among this subgroup of country pairs based on their probability of trade
(whose binary realization is the only relevant datum we observe). Hence, we assign
the same zij to those country pairs with an estimated i j > .9999999, equivalent
to an estimated i j at this cutoff. This censoring affects 4.3% of the 6,602 country
pairs with positive trade flows from the Probit estimation on the reduced sample
(12,198 country pairs).
TABLE II
BASELINE RESULTS
1986 reduced sample
mi j
Indicator variables
(Probit)
Variables Ti j Benchmark NLS Polynomial 50 bins 100 bins

0.213 1.167 0.813 0.847 0.755 0.789

Distance
(0.016) (0.040) (0.049) (0.052) (0.070) (0.088)
Land border 0.087 0.627 0.871 0.845 0.892 0.863
(0.072) (0.165) (0.170) (0.166) (0.170) (0.170)
Island 0.173 0.553 0.203 0.218 0.161 0.197
(0.078) (0.269) (0.290) (0.258) (0.259) (0.258)
Landlock 0.053 0.432 0.347 0.362+ 0.352+ 0.353+
(0.050) (0.189) (0.175) (0.187) (0.187) (0.187)
Legal 0.049 0.535 0.431 0.434 0.407 0.418
(0.019) (0.064) (0.065) (0.064) (0.065) (0.065)
Language 0.101 0.147+ 0.030 0.017 0.061 0.036
(0.021) (0.075) (0.087) (0.077) (0.079) (0.083)
Colonial ties 0.009 0.909 0.847 0.848 0.853 0.838
(0.130) (0.158) (0.257) (0.148) (0.152) (0.153)
Currency union 0.216 1.534 1.077 1.150 1.045 1.107
(0.038) (0.334) (0.360) (0.333) (0.337) (0.346)
FTA 0.343 0.976 0.124 0.241 0.141 0.065
(0.009) (0.247) (0.227) (0.197) (0.250) (0.348)
Religion 0.141 0.281 0.120 0.139 0.073 0.100
(0.034) (0.120) (0.136) (0.120) (0.124) (0.128)
Regulation 0.108 0.146
costs (0.036) (0.100)
R. costs (days 0.061 0.216+
& proc.) (0.031) (0.124)
(from w ij ) 0.840
(0.043)
ij 0.240 0.882
(0.099) (0.209)
z ij 3.261
(0.540)
z i2j 0.712
(0.170)
z i3j 0.060
(0.017)
Observations 12,198 6,602 6,602 6,602 6,602 6,602
R2 0.573 0.693 0.701 0.704 0.706
Notes. Exporter and importer fixed effects. Marginal effects at sample means and pseudo R2 reported
for Probit. Regulation costs are excluded variables in all second stage specifications. Bootstrapped standard
errors for NLS; robust standard errors (clustering by country pair) elsewhere.
+Significant at 10%.
Significant at 5%.
Significant at 1%.
coefficient for ij are precisely estimated. The remaining results

for the linear coefficients clearly demonstrate the importance of an
unmeasured heterogeneity bias in the estimated effects of trade
barriers: higher trade volumes are not just the direct consequence
of lower trade barriers; they also represent a greater proportion of
exporters to a particular destination. Consequently, the measures
of the effects of trade frictions in the benchmark gravity equa-
tion are biased upward, as they confound the true effect of these

frictions with their indirect effect on the proportion of exporting
firms.32 As highlighted in Table II, these biases are substantial.
The coefficient on distance drops roughly by a third, indicating a
much smaller effect of distance on firm-level (hence product-level)
trade.33 The effects of a currency union and colonial ties on firm
or product level trade are also substantially reduced. The bias
for the effect of FTAs is even more severe, as its coefficient drops
by almost an order of magnitude and becomes insignificant. The
measured effect of a common language is also strongly affected; it
becomes insignificant (the benchmark coefficient is significant at
the 5.2% level) and is precisely estimated around zero. Similarly
for common religion: it becomes insignificant. This suggests that
FTAs, a common language, and a common religion predominantly
reduce the fixed costs of trade: they have a great influence on a
firms choice of export location, but not on its export volume once
the exporting decision has been made.
We now progressively relax the parameterization assump-
tions that determined our functional forms. First, we relax the
assumption governing the distribution of firm heterogeneity, and
hence the form of the control function w ij () for z ij in the trade
flow equation (14). That is, we drop the Pareto assumption for
G(.) and revert to the general specification for Vi j in (5). Using
(4) and (10), vi j (zi j ) is now an arbitrary (increasing) function
of zi j . We then directly control for E[Vi j | ., Ti j = 1] using (z ij ),
which we approximate with a polynomial in z ij . This replaces
w ij ln{exp[(z ij )] 1} in (14).34 As the nonlinearity induced by
w ij is eliminated, we now estimate the second stage using OLS.
32. The effect of a land border is an exception because it negatively affects

the probability of trade.
33. Several studies have documented that the effect of distance in gravity
models is overstated because distance is correlated with other trade frictions (such
as lack of information). The same issue applies here and would even further reduce
the directly measured effect of distance.
34. Recall that wi j and vi j differ only by a constant term.
In practice, we have found no noticeable changes from expand-

ing (z ij ) beyond a cubic polynomial. The results from this sec-
ond stage estimation (the first stage Probit remains unchanged)
are reported in the fourth column of Table II. These results are
very similar to the NLS estimates.35 In other words, the Pareto
distribution does not appear to unduly constrain our baseline
specification.
We further relax the joint normality assumption for the un-

observed trade costs, and hence the Mills ratio functional form
for the selection correction. This naturally precludes the separa-
tion of the effects of the latter from the firm heterogeneity effects.
However, we can still jointly control for these effects with a flexi-
ble nonparametric functional form and thus obtain our key results
for the intensive-margin contribution of the various trade barri-
ers. The first stage estimation remains the same except that we
now can use any cumulative distribution function instead of the
normal distribution. We have experimented with the Logit and
t-distributions with various low degrees of freedom and found
that the resulting predicted probabilities i j are strikingly simi-
lar. For this reason, we no longer use the normality assumption to
recover the z ij and ij . Instead, we work directly with the predicted
probabilities i j .
In order to approximate as flexibly as possible an arbitrary
functional form of the i j s, we use a large set of indicator variables.
We partition the obtained i j s into a number of bins with equal
observations and assign an indicator variable to every bin. We
then replace the w ij and ij controls from the NLS estimation or
the z ij and ij controls from the polynomial estimation with this
set of indicator variables. We report results with both 50 and 100
bins, to ensure a large degree of flexibility.36 The results are in
the last two columns of Table II. Here, we use the predicted prob-
abilities from the baseline Probit, but these results are virtually
unchanged when switching to a Logit or a t-distribution in the first
stage. Evidently, all three estimation methods yield very similar
results.
35. Here, we report the robust standard errors controlling for clustering at
the country-pair level but do not correct for the generated regressors in the second
stage. We experimented with bootstrapping the standard errors, as performed
for the NLS specification, but this barely affected any of them. No coefficient
significance test (at the 1%, 5%, or 10% level) was affected.
36. As with the polynomial approximation, this specification is now linear,
and we thus use OLS.
VII. AN ALTERNATIVE EXCLUDED VARIABLE

Although the use of regulation cost variables has advantages,
it also has a drawback: it substantially reduces the number of
usable observations, as we explained in Section V (from 24,649 to
12,198 for the first stage, and from 11,156 to 6,602 for the second
stage). For this reason it is desirable to find at least one other
variable that satisfies the exclusion restrictions, which can be
used for estimation with the full sample of countries. We argue in

this section that our religion variable is suitable for this purpose.37
Once we have reliable excluded variables, such as our regu-
lation cost variables, we can test whether any additional variable
satisfies the exclusion restrictions. The key is for this variable to
be correlated with the zi j s but not be correlated with the resid-
ual of the second stage equation that has been estimated with the
reliable excluded variables (the reliable excluded variables are be-
lieved to satisfy the exclusion restrictions on theoretical grounds).
In our case this means that the residuals from the trade flow equa-
tion should be uncorrelated with this variable. We argue that our
common religion variable satisfies these requirements.
That the religion variable satisfies the first requirement is
evident from the Probit equation, in which religion has a positive
and significant effect on the probability of exporting (see Tables I
and II). A simple test of the second requirement is provided in
Table II. As is evident from the standard errors, one cannot reject
the hypothesis that the coefficient on religion equals zero in each
and every case. In other words, religion is not correlated with
the second stage residuals.38 To further enhance confidence in
common religion as the excluded variable, we rerun all the second
stage specifications from Table II, dropping the regulation cost
variables and using religion as the excluded variable. The results
of this estimation procedure applied to the reduced sample are
reported in the left-hand-side panel of Table III. Evidently, they
are very similar to the results in Table II.
37. We also experimented using the common language variable as the ex-
cluded variable. We obtained results almost identical to those using religion as the
excluded variable.
38. We also performed a chi-squared test with one overidentification restric-
tion (see Wooldridge [2002]) using all three excluded variables (the two regulations
of entry costs variables and common religion). However, since the second stage
residuals are no longer normally distributed after correcting for sample selection,
this test is only asymptotically valid. Still, in all specifications, we cannot reject
the hypothesis that all three variables are uncorrelated with the second stage
residuals.
TABLE III
ALTERNATE EXCLUDED VARIABLES
1986 reduced sample 1986 full sample

Indicator variables Indicator variables
Variables NLS Polynomial 50 bins 100 bins NLS Polynomial 50 bins 100 bins
Distance 0.822 0.853 0.751 0.731 0.798 0.862 0.671 0.623

(0.048) (0.051) (0.069) (0.089) (0.039) (0.041) (0.059) (0.076)
Land border 0.878 0.855 0.903 0.907 0.834 0.786 0.894 0.924
(0.169) (0.164) (0.166) (0.167) (0.132) (0.144) (0.147) (0.150)
Island 0.204 0.219 0.171 0.142 0.169 0.200+ 0.091 0.074
(0.291) (0.258) (0.265) (0.266) (0.120) (0.118) (0.119) (0.121)
Landlock 0.348 0.361+ 0.347+ 0.344+ 0.447 0.482 0.437 0.439
(0.176) (0.188) (0.190) (0.192) (0.172) (0.186) (0.186) (0.186)
Legal 0.439 0.442 0.424 0.418 0.387 0.385 0.350 0.345
(0.066) (0.064) (0.065) (0.066) (0.048) (0.049) (0.050) (0.050)
ESTIMATING TRADE FLOWS
Language 0.018 0.005 0.060 0.068 0.023 0.045 0.044 0.062

(0.085) (0.077) (0.079) (0.085) (0.062) (0.061) (0.064) (0.068)
Colonial ties 0.835 0.839 0.837 0.830 1.001 1.038 0.960 0.929
(0.251) (0.147) (0.149) (0.148) (0.204) (0.116) (0.117) (0.119)
Currency union 1.034 1.102 1.021 0.984 1.023 1.106 0.977 0.960
(0.361) (0.334) (0.341) (0.353) (0.273) (0.261) (0.265) (0.270)
467

468
TABLE III
(CONTINUED)
1986 reduced sample 1986 full sample

Indicator variables Indicator variables
Variables NLS Polynomial 50 bins 100 bins NLS Polynomial 50 bins 100 bins
FTA 0.154 0.254 0.161 0.200 0.380 0.457 0.050 0.091

(0.225) (0.199) (0.250) (0.337) (0.182) (0.162) (0.165) (0.210)
(from w ij ) 0.827 0.871
(0.043) (0.028)
ij 0.198 0.823 0.372 1.131
(0.099) (0.211) (0.069) (0.138)
z ij 3.229 3.602
(0.538) (0.386)
z i2j 0.709 0.782
(0.169) (0.123)
z i3j s 0.061 0.064
(0.017) (0.013)
Observations 6,602 6,602 6,602 6,602 11,146 11,146 11,146 11,146
QUARTERLY JOURNAL OF ECONOMICS
R2 0.700 0.702 0.705 0.721 0.722 0.723
Notes. mi j is dependent variable throughout. Exporter and importer fixed effects. Religion is excluded variable in all second stage specifications. Bootstrapped standard errors
for NLS; robust standard errors (clustering by country pair) elsewhere.
Significant at 5%.
Significant at 1%.

Once religion is accepted as a legitimate excluded variable,

we can use it to estimate the model on the full sample of countries
instead of the smaller sample that has been used so far, thereby
roughly doubling the number of usable observations. The results
of this estimation are reported in the right-hand-side panel of Ta-
ble III for all three estimation methods (the benchmark gravity
and Probit selection results were already reported in Table I). The
magnitudes of these coefficients remain comparable despite the

substantial increase in sample size.39 Similar to the reduced sam-
ple estimates, we find that heterogeneity matters; higher trade
volumes are driven both by direct effects of lower trade barriers
and by greater proportions of exporters. As a result, estimated
trade frictions in the benchmark gravity equation are biased up-
ward, confounding the true effects with the indirect effects on the
fraction of exporting firms.40 As is evident from Table III, these bi-
ases are substantial; the coefficients on distance, currency union,
colonial ties, language, and FTAs drop significantly.41
The substantial increase in country coverage allows us to
study how these biases vary with the characteristics of the
country pairs, which we explore in our counterfactual analysis
in the following section. For now, we also take advantage of
the inclusion of French exports in the full sample to compare
our estimates for the extensive margin of French exports to the
direct measure of the number of French exporting firms across
destinations reported by Eaton, Kortum, and Kramarz (2004) for
1986. In our model, wi j is an increasing function of the fraction
of firms exporting from country j to country i. Our estimates of
wi j for j = France should therefore be positively correlated with
the number of French exporters to country i reported by Eaton,
Kortum, and Kramarz (2004). We check this using our estimates
for both w ij from the NLS specification and for (z ij ) from the
polynomial approximation. These correlations are extremely high
in both cases: 77% for w ij and 78% for (z ij ).42
39. The effects of FTAs are estimated to be significantly higher in the NLS
and polynomial approximation specifications, though still substantially lower than
in the benchmark estimates.
40. The effect of a land border is again an exception, because it negatively
affects the probability of trade.
41. In the working paper version (Helpman, Melitz, and Rubinstein 2007),
we also report results for the 1980s. They show that 1986 is not exceptional in
terms of the full sample estimates. The coefficient for joint membership in the
WTO drops substantially, but remains statistically and economically significant.
42. Because wi j is a logged value, we compute the correlation using the loga-
rithm of the number of exporting firms.
VIII. ADDITIONAL INSIGHTS

We now use the full sample estimates from the previous sec-
tion to examine several aspects of the results in further detail.
VIII.A. Decomposing the Biases

Our second stage estimate addresses two different sources
of bias for standard gravity equations: a selection bias that arises

from the pairing of countries into exporterimporter relationships,
and an unobserved heterogeneity bias that results from the varia-
tion in the fraction of firms that export from a source to a destina-
tion country. To examine the relative importance of these biases,
we now estimate two specifications of the second-stage equation,
one controlling for unobserved heterogeneity only, the other con-
trolling for selection only.
The results are reported in Table IV. The first two columns
report the benchmark equation and our second stage NLS esti-
mates from the full sample from Tables I and III. The differences
in the estimated coefficients of these two equations represent the
joint outcome of the two biases. As we discussed, all the coeffi-
cients, with the exception of the land border effect, are lower in
absolute value in the second column. We then implement a sim-
ple linear correction for unobserved heterogeneity by only adding
zij = 1 (i j ) as an additional regressor to the standard gravity
specification (here, we do not correct for the sample selection bias
via ij ).43 The results reported in the third column clearly show
that this unobserved heterogeneity (the proportion of exporting
firms) addresses almost all the biases in the standard gravity
equation. The coefficients and standard errors for all the observed
trade barriers are very similar to those obtained in our second
stage nonlinear estimation.
In the fourth column, we correct only for the selection bias (the
standard two-stage Heckman (1979) selection procedure) by intro-
ducing the Mills ratio ij as an additional regressor to the bench-
mark specification. Although the estimated coefficient on ij is pos-
itive and significant, the remaining coefficients are very similar
to those obtained in the benchmark specification of column (1).44
43. In this exercise we want to ensure a simple monotonic transformation of

zij , so we do not add any higher order terms.
44. This is consistent with the finding of Silva and Tenreyro (2006), whose
application of a Poisson estimation method on a sample that consists of positive
trade flows and a sample that includes zeros as well yields similar results.
TABLE IV
BIAS DECOMPOSITION
1986 full sample

Firm Heckman
Variables Benchmark NLS heterogeneity selection
Distance 1.176 0.798 0.769 1.214

(0.031) (0.039) (0.038) (0.031)
0.458 0.834 0.855 0.436

Land border
(0.147) (0.132) (0.142) (0.149)
Island 0.391 0.169 0.164 0.425
(0.121) (0.120) (0.118) (0.120)
Landlock 0.561 0.447 0.433 0.565
(0.188) (0.172) (0.187) (0.187)
Legal 0.486 0.387 0.381 0.488
(0.050) (0.048) (0.049) (0.050)
Language 0.176 0.023 0.023 0.223
(0.061) (0.062) (0.060) (0.061)
Colonial ties 1.299 1.001 0.979 1.311
(0.120) (0.204) (0.119) (0.123)
Currency union 1.364 1.023 0.996 1.391
(0.255) (0.273) (0.260) (0.257)
FTA 0.759 0.380 0.314+ 0.737
(0.222) (0.182) (0.168) (0.235)
Religion 0.102
(0.096)
(from w ij ) 0.871
(0.028)
ij 0.372 0.265
(0.069) (0.070)
z ij 0.892
(0.051)
Observations 11,146 11,146 11,146 11,146
R2 0.709 0.716 0.710
Notes. mi j is dependent variable throughout. Exporter and importer fixed effects. Religion is excluded
variable in all second stage specifications. Bootstrapped standard errors for NLS; robust standard errors
(clustering by country pair) elsewhere.
Significant at 5%.
Significant at 1%.
Thus, the bias corrections implemented in our second stage

estimation are dominated by the influence of unobserved firm
heterogeneity rather than sample selection. This finding sug-
gests that although aggregate country-pair shocks do have a
significant effect on trade patterns, they only negligibly affect the
TABLE V
ASYMMETRIES
1986 full sample

mi j mji
Variable Ti j T ji NLS Polynomial
i j ji 0.999
(0.0169)

w ij w ji 1.187 1.251
(0.042) (0.266)
v(z ij ) v(z ji ) 1.012 0.703
(0.035) (0.143)
Country fixed effects No No Yes No Yes
Observations 12,246 4,517 4,517 4,517 4,517
R2 0.219 0.153 0.324 0.157 0.325
Significant at 1%.
responsiveness of trade volumes to observed trade barriers.45

The results in column (3) clearly show that this is not the case
for the effects of unobserved heterogeneity: the latter would
affect trade volumes even were all country pairs trading with one
another, because it operates independently of the selection effect.
Neglecting to control for this unobserved heterogeneity induces
most of the biases exhibited in the standard gravity specification.
VIII.B. Evidence on Asymmetric Trade Relationships

As was previously mentioned, our model predicts asymmetric
trade flows between countries. These asymmetries can be extreme,
with trade predicted in only one direction, as also reflected in the
data. More nuanced, trade can be positive in both directions, but
with a net trade imbalance. Do these predicted asymmetries have
explanatory power for the direction of trade flows and net bilateral
trade balances? The answer is an overwhelming yes, as evidenced
by the results reported in Table V. The first part of the table shows
the results of the OLS regression of Ti j T ji on i j ji (based on
the Probit results for 1986).46 Note that the regressand, Ti j T ji ,
takes on the values 1, 0, and 1, depending on the direction of
45. This finding also highlights the important information conveyed by the
nontrading country pairs. If such zero trade values were just the outcome of cen-
soring, then a Tobit specification would provide the best fit to the data. This is just
a more restrictive version of the selection model, which is rejected by the data in
favor of the specification incorporating firm heterogeneity.
46. Recall that Ti j is the indicator variable for positive trade from j to i.
trade between i and j (it is 0 if trade flows in both directions or

if the countries do not trade at all). The magnitude of the regres-
sor i j ji measures the models prediction for an asymmetric
trading relationship, while its sign predicts the direction of the
asymmetry. Table V shows that the predicted asymmetries have
a substantial amount of explanatory power; the regressor coeffi-
cient is significant at any conventional level and explains on its
own 22% of the variation in the direction of trade.47 We emphasize

that the regressor is constructed only from the predicted probabil-
ity of export i j , which is a function only of country-level variables
(the fixed effects) and symmetric bilateral measures.
The second part of Table V focuses on bilateral trade flow
asymmetries between country pairs trading in both directions.
It shows the results of the OLS regression of net bilateral trade
mi j mji (the percentage difference between exports and imports)
on alternatively w ij w ji (for the NLS specification) or (z ij )
(z ji ) (for the polynomial approximation). That regressor captures
differences in the proportion of exporting firms. Combined with
the country fixed effects, these variables capture differences in the
number of exporting firms from one country to the other. Again,
we find that this single regressor (using either specification) is
a strong predictor of net bilateral trade. On its own, it explains
15%16% of the variance in net trade, and along with the country
fixed effects it explains 32%33% of that variance.48
VIII.C. Counterfactuals
We have just shown how the fitted values for i j and wij can
explain a large portion of the variation in the direction of trade
and in its extensive margin. We next show how to use these fitted
values to make predictions about the response of trade to changes
in trade costs. For every change in the bilateral trade costs di j , our
model predicts the new pattern of trade, that is, who trades with
whom, and in which direction. In addition, for country pairs that
trade with each other, the model predicts the resulting changes in
the composition of trade flows between the extensive and intensive
margins. These counterfactual predictions can be measured, and
47. This understates the variables explanatory power, because it is continu-

ous and it predicts a discrete variable.
48. Anderson and van Wincoop (2003) account for asymmetric bilateral trade
flows with asymmetric variable bilateral trade costs. In a more general model one
can have both asymmetric bilateral trade costs and asymmetric extensive margins
of trade.
we illustrate their quantitative impact for a reduction in trade

costs associated with distance.
In response to a drop in distance-related trade costs, some
countries start trading with one another. Trade rises for country
pairs that traded before the drop in trade costs, and we report how
the increase in trade can be decomposed into the intensive and ex-
tensive margin. We find that the extensive margin is especially
important in shaping the response of trade flows across coun-

try pairs because it generates substantial heterogeneity across
country pairs. This richness contrasts sharply with the uniform
response implied by the baseline gravity model, which does not
account for the extensive margin of trade (nor does it account for
the creation of new trading relationships).
The computation of these responses involves some techni-
cal details that are explained in Appendix III. Here we report
the results of a particular counterfactual experiment involving
a decrease in the trade costs associated with distance. That is,
we investigate the response of trade for any given country pair,
assuming that the distance between these two country pairs de-
creases by a given percentage. We first focus on country pairs
observed trading and focus on the elasticity of the overall trade
response for each country pair: |mi j mi j |/|di j di j |, where di j now
specifically references the bilateral distance variable.49 Since our
model predicts different response elasticities with the magnitude
of the trade decrease, we report these elasticities for the case of a
10% distance decrease (di j di j = log 0.9), although any percent-
age decrease under 20% would yield virtually identical results.50
As was previously mentioned, the elasticities vary widely
across different country pairs. In order to highlight how these elas-
ticities vary along one important country pair dimensioncountry
incomewe report summary statistics across three groups of
country pairs: NorthNorth, NorthSouth, and SouthSouth,
sorted by GDP per capita.51 These statistics appear in Table IV for
49. To avoid any confusion when discussing larger versus smaller elastic-
ities, we express the elasticities in absolute value. Naturally, for the case of trade
costs, these elasticities are all negative.
50. Larger decreases in trade costs would produce larger elasticities but with
similar qualitative patterns across country pairs.
51. We use 1986 U.S. $15,000 as the cutoff GDP per capita between North and
South. The former group is composed of nineteen countries: Australia, Austria,
Belgium-Luxemburg, Canada, Denmark, Finland, France, Germany, Hong Kong,
Iceland, Italy, Japan, Netherlands, New Zealand, Norway, Sweden, Switzerland,
United Kingdom, and the United States.
TABLE VI
SUMMARY STATISTICS OF THE TRADE ELASTICITY RESPONSE ACROSS COUNTRY PAIRS
Country Number of Nonlinear least squares Polynomial aproximation

pairs country
group pairs Mean S. D. Min Max Mean S. D. Min Max
NN 342 1.292 0.034 1.283 1.642 1.290 0.107 1.141 2.222

NS 4,626 1.404 0.152 1.283 2.949 1.526 0.386 1.141 2.895
SS 6,178 1.698 0.303 1.283 3.777 2.130 0.443 1.141 2.995

Overall 11,146 1.563 0.289 1.283 3.777 1.854 0.517 1.141 2.995
both our NLS and polynomial approximation specifications. Im-

portantly, we emphasize that all the heterogeneity in the elasticity
response is driven by the extensive margin because the elastic-
ity response at the intensive margin is fixed at 0.799 (NLS) and
0.862 (polynomial approximation). Because this extensive mar-
gin response depends fundamentally on the functional forms for
w ij or (z ij ) in terms of z ij , we report the elasticities for both
cases. Although the shape of the functional form for w ij is in part
determined by our theoretical modeling assumptions (see (13)),
the shape of the (z ij ) is entirely data-driven. Reassuringly, both
functions have very similar shapes over the range of z ij , and the
counterfactual distributions of the response elasticity are similar.
The heterogeneous trade responses reported in Table IV show
that these elasticities vary between 1.283 and 3.777 for the NLS
estimates and between 1.141 and 2.995 for the semiparametric
estimates, large variations indeed.52 We visually depict these dis-
tributions across country pairs group in Figure III. The charts
clearly document how the range and distribution of elasticities
vary with country income: the elasticities are highest for South
South trade, lower for NorthSouth trade, and lowest for North
North trade. Thus, when trade costs related to distance fall, our
estimates predict that the response of the extensive margin of
trade is larger for less developed countries.53
52. Of course, departing from the log-linear specification for distance would
yield different elasticities for different changes in trade costs related to distance.
Our main point is that, given a log-linear specification for distance in both stages,
our model still predicts substantial differences in the response elasticity, driven
by the characteristics of the country pairs that jointly determine the extensive
margin of trade.
53. In Helpman, Melitz, and Rubinstein (2007) we also report how many of
the countries that do not trade initially and which pairs start trading when the
trade costs fall. These results suggest that large changes in trade-related costs
are needed to induce nontrading country pairs, involving at least one Southern

FIGURE III
The Distribution of the Distance Elasticity across Country Pairs
IX. CONCLUDING COMMENTS

Empirical explanations of international trade flows have a
long tradition. The gravity equation with various measures of
trade resistance plays a key role in this literature. Indeed, esti-
mates of the impact of trade resistance measures provide impor-
tant information about the roles played by common currencies,
free trade areas, membership in the WTO, and other features of
trading countries. For this reason, it is important to obtain reli-
able estimates of the effects of those trade barriers/enhancers on
international trade flows.
We develop in this paper an estimation procedure that cor-
rects certain biases embodied in the standard gravity estimation
of trade flows. Our approach is driven by theoretical as well as
econometric considerations. On the theoretical side, we develop
a simple model that is capable of explaining empirical phenom-
ena, such as zero trade flows between certain pairs of countries
and larger numbers of exporters to larger destination markets.
We then derive from this theory a two-equation system that can
be estimated with standard data sets. Importantly, this system
country, to trade. Moreover, they are in line with the evidence presented in Figures
I and II that almost all of the increase in world trade flows in the last thirty years
has occurred among countries with trading relationships in 1970.
enables one to decompose the impact on trade volumes of all trade

resistance measures into their intensive (trade volume per firm)
and extensive (number of exporting firms) margin components.
We show how to obtain estimates of this decomposition without
having firm-level data, but rather country-level data that are nor-
mally used to estimate trade flows. The ability to obtain such
a decomposition is important because, in practice, a substantial
proportion of trade adjustment takes place at the extensive mar-

gin, and it is not possible to obtain consistent firm-level data
with export destinations for a large number of countries (which
would be needed for a direct estimation of the extensive margin
component).
Our empirical analysis has been confined to country-level
trade flows, where about half of the observations are zeros. Nat-
urally, the problem of zeros is even more severe at the industry
level. That is, in data sets of sectoral trade flows the fraction
of zeros is much larger. Importantly, our estimation method can
be implemented on such data sets as well. Manova (2006) is an
example of this, highlighting the important contribution of the
extensive margin of trade in explaining the impact of financial
frictions on sectoral trade flows.
A variety of robustness checks show that the resulting es-
timates are not sensitive to the estimation method (parametric,
semiparametric, or nonparametric) nor to the excluded variables
from the first stage of our two-stage estimation procedure. More-
over, these estimates suggest that the biases embodied in the com-
monly used approach are substantial and that they are mostly due
to the omitted control for the extensive margin of trade. Especially
important is our finding that the bias not only is large, but also
substantially varies across country pairs with different charac-
teristics. In particular, the response of the trade flow between one
pair of countries to a given reduction in distance-related trade
frictions (such as transport costs) can be as much as three times
larger than the response of the trade flow between another pair
of countries. We show how these large variations across country
pairs in the response to a given trade friction reduction are driven
by variation in the extensive margin responses.
Finally, we note that our estimation procedure is easy to im-
plement. In addition, it is flexible, because it allows the use of
parametric, semiparametric, and nonparametric specifications. In
other words, the procedure provides the researcher with flexibility
and convenience in individual applications.
APPENDIX I
We describe in this Appendix our data sources.
Trade data: The bilateral trade flows are from Feenstras
World Trade Flows, 19701992 and World Trade Flows, 1980
1997. These data include 183 country titles over the period
1970 to 1997. In some cases Feenstra grouped several countries
into a single title. We excluded twelve such country titles and
three proper countries for which data other than trade flows were

missing. This left usable data for bilateral trade flows among 158
countries. The list of these countries is provided in Table A.1.
For the 158 countries we constructed a matrix of trade flows,
measured in constant 2000 U.S. dollars, using the U.S. CPI. This
matrix represents 158 157 = 24,806 observations, consisting of
exports from country j to country i. Many of these export flows
are zeros.
Country-level data: Population and real GDP per capita have
been obtained from two standard sources: the Penn World Tables
6.1 and the World Banks World Development Indicators.
We used the CIAs World Factbook to construct a number of
variables, which can be classified as follows:54
1. Geography: Latitude, longitude, and whether a country is

landlocked or an island.
2. Institutions: Legal origin, colonial origin, GATT/WTO
membership.
3. Culture: Primary language and religion.
We also used data from Rose (2000) and Glick and Rose (2002),
as presented on Andrew Roses Web site, to identify whether a
country pair belongs to the same currency union or the same FTA.
And we used data from Rose (2004) to identify whether a country
is a member of the GATT/WTO.
Using these data, we constructed country-pair specific vari-
ables, such as the distance between countries i and j, whether
they share a border, the same legal system, the same colonial
origin, or membership in the GATT/WTO (see below).
The construction of the regulation costs of firm entry are de-
scribed in the main text. As previously mentioned, cost data on the
number of days, number of legal procedures, and relative cost (as
percent of GDP per capita) are reported in Djankov et al. (2002).
54. See http://www.cia.gov/cia/publications/factbook/docs/profileguide.html.

TABLE A.1
LIST OF COUNTRIES
Afghanistan Ecuador Korea DPR Romania

Albania Egypt Korea Rep. Rwanda
Algeria El Salvador Kuwait Saudi Arabia
Angola Eq. Guinea Laos Senegal
Argentina Ethiopia Lebanon Seychelles
Australia Fiji Liberia Sierra Leone
Austria Finland Libya Singapore

Bahamas Fm. Czechoslovakia Madagascar Solomon Islds.
Bahrain Fm. USSR Malawi Somalia
Bangladesh Fm. Yugoslavia Malaysia South Africa
Barbados France Maldives Spain
BelgiumLux. French Guiana Mali Sri Lanka
Belize Gabon Malta St Kitts Nevis
Benin Gambia Mauritania Sudan
Bermuda Germany Mauritius Surinam
Bhutan Ghana Mexico Sweden
Bolivia Greece Mongolia Switzerland
Brazil Greenland Morocco Syria
Brunei Guadeloupe Mozambique Taiwan
Bulgaria Guatemala Myanmar Thailand
Burkina Faso Guinea Nepal Togo
Burundi Guinea-Bissau Netherlands Trinidad-Tobago
Cambodia Guyana Neth. Antilles Tunisia
Cameroon Haiti New Caledonia Turkey
Canada Honduras New Zealand Turks and Caicos
Cayman Islds Hong Kong Nicaragua Uganda
Central Africa Hungary Niger United Kingdom
Chad Iceland Nigeria United Arab Em.
Chile India Norway United Rep.
China Indonesia Oman Tanzania
Colombia Iran Pakistan United States
Comoros Islds. Iraq Panama Uruguay
Congo Ireland Papua N.Guinea Venezuela
Costa Rica Israel Paraguay Vietnam
Cote DIvoire Italy Peru Western Sahara
Cuba Jamaica Philippines Yemen
Cyprus Japan Poland Zaire
Denmark Jordan Portugal Zambia
Djibouti Kenya Qatar Zimbabwe
Dominican Rep. Kiribati Reunion
Main Variables:
1. Distance: the distance (in km) between importers i and

exporters j capitals (in logs).
2. Common border: a binary variable that equals one if im-

porter i and exporter j are neighbors that meet a common
physical boundary, and zero otherwise.
3. Island: a binary variable that equals one if both importer
i and exporter j are islands, and zero otherwise.
4. Landlocked: a binary variable that equals one if both ex-
porting country j and importing country i have no coastline
or direct access to sea, and zero otherwise.

5. Colonial ties: a binary variable that equals one if importing
country i ever colonized exporting country j or vice versa,
and zero otherwise.
6. Currency union: a binary variable that equals one if im-
porting country i and exporting country j use the same
currency or if within the country pair money was inter-
changeable at a 1:1 exchange rate for an extended period
of time (see Rose [2000, 2004] and Glick and Rose [2002]),
and zero otherwise.
7. Legal system: a binary variable that equals one if the im-
porting country i and exporting country j share the same
legal origin, and zero otherwise.
8. Religion: (% Protestants in country i % Protestants in
country j) + (% Catholics in country i % Catholics in coun-
try j) + (% Muslims in country i % Muslims in country j).
9. FTA: a binary variable that equals one if exporting
country j and importing country i belong to a common
regional trade agreement, and zero otherwise.
10. WTO: a vector of two dummy variables: the first binary
variable equals one if both exporting country j and import-
ing country i do not belong to the GATT/WTO, and zero
otherwise; the second binary variable equals one if both
countries belong to the GATT/WTO, and zero otherwise.
11. Entry costs: a binary indicator that equals one if the sum
of the number of days and procedures to form a business
is above the median for both the importing country i and
exporting country j, or if the relative cost (as percent of
GDP per capita) of forming a business is above the median
in the exporting country j and the importing country i,
and zero otherwise.
APPENDIX II
We derive in this Appendix a gravity equation with third-
country effects, which generalizes Anderson and van Wincoops
(2003) equation, and we show that their equation applies when-

ever i j = ji for every country pair and Vi j can be decomposed
in a particular way. We then discuss some limitations of their
formulation.
Equality of income and expenditure implies Yi = Jj=1 M ji .
That is, the value of country is exports to all countries, including
sales to home residents Mii , equals the value of country is output.
Equation (6) then implies

c 1
j
hj 1
(15) Yj = Nj Yh Vhj .
h
Ph
Using this expression we can rewrite the bilateral trade volume

(6) as

1
i j
Yi Y j Pi
Vi j
(16) Mi j =
,
Y J hj
1
h=1 Ph V hj sh

where Y = Jj=1 Y j is world income and sh = Yh/Y is the share of
country h in world income.
We next show that if Vi j is decomposable in a particular way,
and transport costs are symmetric (i.e., i j = ji for all i and j),
then (16) yields the generalized gravity equation that has been
derived by Anderson and van Wincoop (2003). Their specifica-
tion satisfies these conditions. Importantly, however, there are
other cases of interest, less restrictive than the Anderson and
van Wincoop specification, that satisfy them too. Therefore, our
derivation of the gravity equation shows that it applies under
wider circumstances and, in particular, when there is productivity
heterogeneity across firms and firms bear fixed costs of exporting.
Under these circumstances only a fraction of the firms export:
those with the highest productivity. Finally, note that our general
formulationwithout decomposabilityis more relevant for em-
pirical analysis because, unlike previous formulations, it enables
bilateral trade flows to equal zero. This flexibility is important be-
cause, as we have explained in the Introduction, there are many
zero bilateral trade flows in the data.
Consider the following:
DECOMPOSABILITY ASSUMPTION. Vi j is decomposable as follows:
1
Vi j = I M,i EX, j i j ,
where I M,i depends only on the parameters of the importing

country, EX, j depends only on the parameters of the exporting
country, and i j = ji for all i, j.
In this decomposition, only the symmetric terms i j depend on the
joint identity of the importing and exporting countries; all other
parameters do not.
To illustrate circumstances in which the decomposability as-
sumption is satisfied, first consider a situation where the fixed

costs fi j are very small, so that ai j > aH for all i, j. That is, the
lowest productivity level that makes exporting profitable, 1/ai j ,
is lower than the lowest productivity level in the support of G(),
1/aH . Under these circumstances all firms export and Vi j is the
same for every country pair i, j.55 Alternatively, suppose that pro-
ductivity 1/a has a Pareto distribution with shape k and aL = 0.
That is, G(a) = (a/aH )k for 0 a aH . Moreover, let either fi j de-
pend only on the identity of the exporter, so that fi j = f j , or let the
fixed costs be symmetric, so that fi j = f ji . Then Vi j satisfies the
decomposability assumption and in every country j only a fraction
of firms export to country i.56
Using the decomposability property and symmetry require-
ments i j = ji and i j = ji , we obtain57
1
Mi j i j i j
(17) = si s j ,
Y Qi Q j
where the values of Q j are solved from
jh jh 1
(18) Q1
j = sh.
h
Qh
This is essentially the Anderson and van Wincoop (2003) system.

Evidently, the solution of the Q j s depends only on income shares
a
55. More precisely, Vi j = a H a1 dG(a).
L
56. Under these conditions Vi j = k(ai j )k+1 /(aH )k(k + 1) and either ai j =
[c j f j /(1 )]1/(1) /(i j c j / Pi ), so that f j becomes part of EX, j whereas i j be-
comes part of i j , or ai j = [c j fi j /(1 )]1/(1) /(i j c j / Pi ), so that fi j and i j become
part of i j .
57. Decomposability allows us to rewrite (16) as
1
Yi Y j i j i j
(20) Mi j = ,
Y Qi Q j
and transport costs and possibly on a constant in Vi j that is em-

bodied in the i j s. However, an upward shift of this constant raises
the product Qi Q j proportionately and therefore has no effect on
Mi j . Therefore, imports of country i from j as a share of world in-
come, which equal imports of country j from i as a share of world
income, depend only on the structure of trade costs and the size
distribution of countries. Bilateral imports as a fraction of world
income are proportional to the product of the two countries shares

in world income, with the factor of proportionality depending on
the structure of trading costs and the worldwide distribution of
relative country size.
The decomposability assumption is too restrictive, however.
It implies that if imports of country i from j equal zero, that
is, Vi j = 0, then one of the s ( I M,i , EX, j , or i j ) must be infinite,
because > 1. In other words, some trade costs, either at the coun-
try or bilateral level, must be infinite in order to explain zero trade
flows. Our framework, which does not rely on this decomposability
assumption, is much more general, as it can explain the prevalent
zero trade flows based on finite trade costs (which can then be es-
timated). Furthermore, the gravity specification (17) based on the
decomposability assumption cannot explain the asymmetries in
bilateral trade flows (which must then stem from country fixed ef-
fects). In the case of zero bilateral trade in only one direction, this
would impose either that the importer does not import from any
other country or that the exporter does not export to any other
where Qi = Pi / I M,i and

hj hj 1
(21) Q1
j = sh .
Qh
h
In addition, (7) and (15) imply

chihih 1 1
Qi1 = Nh EX,h ,

h

c 1 1 1
j
sj = N j EX, j Q j .

Therefore,
jh jh 1
(22) Q1
j = sh .
h
Qh
Equations (21) and (22) together with symmetry conditions i j = ji and i j = ji

then imply that Q j = Q j for every j. As a result, (20) and (21) yield the equations
in the text.
country. This is clearly inconsistent with the data. As we have

explained in the introduction, most countries trade only with a
fraction of the countries in the world economy: neither with all of
them nor with none of them. In the case of positive trade flows in
both directions, (17) imposes that all bilateral trade asymmetries
stem from the country fixed-effects. This is also inconsistent with
the observed pattern of trade, as documented in the second panel
of Table V. Furthermore, that table documents that those asym-

metries are highly correlated with the asymmetric pattern of zero
trade flows (which would be inconsistent with (17)). Indeed, this is
the main logic behind our more general theoretical model and em-
pirical implementation: the decision to export to a foreign country
is not independent of the volume of exports, and thus that the
pattern of trading partners and trading volumes must jointly be
analyzed. For these reasons we use the less restrictive equations
(4)(7) for estimation purposes.
APPENDIX III
We explain in this Appendix the computation of the counter-
factuals in Section VIII. To this end, consider an observed change
in the bilateral trade costs from di j to di j .58 The new predicted esti-
mates of the probability of trade i j and zij = 1 (i j ) are obtained
in a straightforward way from the first stage estimated Probit
equation by replacing di j with di j . We next need to obtain a con-
sistent estimate of zij conditional on the observed trade status of
j and i (trade or no trade) when trade costs are di j , given that
we do not observe the trade status under the new trade costs di j .
This will replace z ij in our equations. Originally we were only con-
cerned with computing z ij for country pairs with active trade, that
is, with Ti j = 1. But now we also need to consider country pairs
that do not trade under costs di j but might trade under costs di j .
For this reason we need to examine two cases.
III. A. Country Pairs Observed Trading

First, we note that the unobserved trade costs ij are not
affected by the change in trade costs di j .59 If we knew whether a
58. As in our previous derivations, di j can represent any given observable

variable trade cost.
59. That is, we seek a ceteris paribus counterfactual prediction for a direct
change in di j .
country pair traded under di j , say Tij , then we could construct a

new estimate for ij , say ij , conditional on both Ti j and Tij . Absent
this additional information, our best estimate for ij is conditional
on Ti j and is still given by ij = E[ij | ., Ti j = 1] = (zij )/(zij ).
Thus, when Ti j = 1, our best estimate for zij is given by

z ij = E zij | ., Ti j = 1 = zij + zij / zij .

Again, note that the new distance cost di j is used to compute
the new zij but not the bias correction for ij . If z ij < 0, then we
predict that j no longer exports to i. Because z ij > 0, this can only
happen when di j > di j (a scenario we will not explicitly consider).
If z ij > 0, then we predict that the country pair continues to trade
(this must be the case when di j < di j ). This new value of z ij can
then be used in conjunction with the second stage estimates to
predict the response of trade flows at the extensive margin. In the
case of the NLS estimation, this is w ij = ln{exp[(z ij )] 1} (and
(z ij ) for the polynomial approximation). The overall predicted
trade response mi j is given by the fitted value from the estimated
second stage equation (14) using the new values for z ij and di j :
(19) mi j = 0 + j + i + di j + w ij + u ij .
In the case of the polynomial approximation, 0 + w ij is replaced

by (z ij ).
III. B. Country Pairs Not Observed Trading

We now show how our model can be used to determine which
nontrading country pairs are predicted to start trading under
costs di j , and the associated new predicted trade flow. The first
stage yields a predicted i j and zij for all country pairs under di j ,
including the nontrading country pairs. We now need to obtain
a consistent estimate for zij for these country pairs, conditional
on Ti j = 0. We start by expanding the definition for ij to include
the country pairs that do not trade: ij = E[ij | ., Ti j ]. (This was
previously defined only when Ti j = 1.) When Ti j = 0, this is given
by
(zij )
ij = E ij | ., Ti j = 0 = E ij | ., ij < zij = ,
1 (zij )
because ij is distributed standard normal. Hence, ij , our consis-

tent estimate for E[ij | ., Ti j ], is constructed as

zij

if Ti j = 0,

1 z
ij =
i j

zij

if Ti j = 1.


z
ij
Using this new expanded definition for ij , our previous definition

for z ij = zij + ij now provides a consistent estimate for E[zij | Ti j ],
which now includes the case for country pairs with Ti j = 0. Note
that, by construction, z ij must be negative whenever Ti j = 0 (recall
that z ij > 0 whenever Ti j = 1).
When trade costs change to di j , we obtain a new z ij for
country pairs with Ti j = 0 in a way similar to that for Ti j = 1:
z ij = zij + ij , where we do not adjust ij for the new value of the
trade costs.60 Whenever z ij > 0, our model predicts that j exports
to i under the trade costs di j . For these country pairs, the new
predicted trade flow mi j can be predicted in a similar way to that
for all the other trading country pairs using (19) along with the
newly constructed z ij .
HARVARD UNIVERSITY AND CIFAR

PRINCETON UNIVERSITY, CENTRE FOR ECONOMIC POLICY RESEARCH, AND NATIONAL
BUREAU OF ECONOMIC RESEARCH
BROWN UNIVERSITY
REFERENCES
Anderson, James A., A Theoretical Foundation for the Gravity Equation, Ameri-
can Economic Review, 69 (1979), 106116.
Anderson, James E., and Eric van Wincoop, Gravity with Gravitas: A Solution to
the Border Puzzle, American Economic Review, 93 (2003), 170192.
,Trade Costs, Journal of Economic Literature, 42 (2004), 691751.
Bernard, Andrew B., Jonathan Eaton, J. Bradford Jensen, and Samuel Kortum,
Plants and Productivity in International Trade, American Economic Review,
93 (2003), 12681290.
Bernard, Andrew B., J. Bradford Jensen, and Peter K. Schott, Importers, Ex-
porters, and Multinationals: A Portrait of Firms in the U.S. that Trade Goods,
NBER Working Paper 11404, 2005.
Davis, Donald R., and David E. Weinstein, Market Access, Economic Geography
and Comparative Advantage: An Empirical Test, Journal of International
Economics, 59 (2003), 123.
60. As before, we do not observe a new Tij under di j .

Djankov, Simeon, Rafael La Porta, Florencio Lopez-de-Silanes, and Andrei Shleifer,

The Regulation of Entry, Quarterly Journal of Economics, 117 (2002), 137.
Eaton, Jonathan, and Samuel S. Kortum, Technology, Geography, and Trade,
Econometrica, 70 (2002), 17411779.
Eaton, Jonathan, Samuel Kortum, and Francis Kramarz, Dissecting Trade:
Firms, Industries, and Export Destination, American Economic Review
(Papers and Proceedings), 94 (2004), 150154.
Evans, Carolyn L., The Economic Significance of National Border Effects, Amer-
ican Economic Review, 93 (2003), 12911312.
Evenett, Simon J., and Anthony J. Venables, Export Growth in Developing Coun-
tries: Market Entry and Bilateral Trade Flows, mimeo, London School of
Economics, 2002.

Feenstra, Robert C., Border Effects and the Gravity Equation: Consistent Meth-
ods for Estimation, Scottish Journal of Political Economy, 49 (2002), 491506.
, Advanced International Trade (Princeton, NJ: Princeton University Press,
2003).
Felbermayr, Gabriel J., and Wilhelm Kohler, Exploring the Intensive and Ex-
tensive Margins of World Trade, Review of World Economics, 142 (2006),
642674.
Frankel, Jeffrey A., Regional Trading Blocks in the World Economic System (Wash-
ington, DC: Institute of International Economics, 1997).
Glick, Reuven, and Andrew K. Rose, Does a Currency Union Affect Trade? The
Time Series Evidence, European Economic Review, 46 (2002), 11251151.
Haveman, Jon, and David Hummels, Alternative Hypotheses and the Volume
of Trade: The Gravity Equation and the Extent of Specialization, Canadian
Journal of Economics, 37 (2004), 199218.
Heckman, James J., Sample Selection Bias as a Specification Error, Economet-
rica, 47 (1979), 153161.
Helpman, Elhanan, Imperfect Competition and International Trade: Evidence
from Fourteen Industrial Countries, Journal of the Japanese and Interna-
tional Economics, 1 (1987), 6281.
Helpman, Elhanan, and Paul R. Krugman, Market Structure and Foreign Trade
(Cambridge, MA: The MIT Press, 1985).
Helpman, Elhanan, Marc Melitz, and Yona Rubinstein, Estimating Trade Flows:
Trading Partners and Trading Volumes, NBER Working Paper No. 12927,
2007.
Hummels, David, and Peter J. Klenow, The Variety and Quality of a Nations
Exports, American Economic Review, 95 (2005), 704719.
Kehoe, Timothy J., and Kim J. Ruhl, How Important Is the New Goods Margin in
International Trade? mimeo, University of Minnesota, 2002.
Manova, Kalina, Credit Constraints, Heterogeneous Firms and International
Trade, mimeo, Harvard University, 2006.
McCallum, John, National Borders Matter: CanadaU.S. Regional Trade Pat-
terns, American Economic Review, 85 (1995), 615623.
Melitz, Marc J., The Impact of Trade on Intra-industry Reallocations and Aggre-
gate Industry Productivity, Econometrica, 71 (2003), 16951725.
Rose, Andrew K., One Money One Market: Estimating the Effect of Common
Currencies on Trade, Economic Policy, 15 (2000), 746.
, Do We Really Know That the WTO Increases Trade? American Economic
Review, 94 (2004), 98114.
Silva, J. M. C. Santos, and Silvana Tenreyro, The Log of Gravity, Review of
Economics and Statistics, 88 (2006), 641658.
Subramanian, Arvind, and Shang-Jin Wei, The WTO Promotes Trade, Strongly
But Unevenly, Journal of International Economics, 72 (2007), 151175.
Tenreyro, Silvana, and Robert Barro, Economic Effects of Currency Unions,
National Bureau of Economic Research, Working Paper 9435, 2003.
Tinbergen, Jan, Shaping the World Economy (New York: The Twentieth Century
Fund, 1962).
Wei, Shang-Jin, Intranational Versus International Trade: How Stubborn are
Nations in Global Integration? NBER, Working Paper No. 5531, 1996.
Wooldridge, Jeffrey M., Econometric Analysis of Cross Section and Panel Data
(Cambridge, MA: The MIT Press, 2002).

Help Man 2008

Uploaded by

Copyright:

Available Formats

Help Man 2008

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Help Man 2008

Uploaded by

Copyright:

Available Formats

THE

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

We develop a simple model of international trade with heterogeneous firms

empirical specifications of bilateral trade flows in which the vol-

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

1. See, for example, Anderson (1979), Helpman and Krugman (1985),

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

from i to j for some country pairs. Both types of trade patterns

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

9. Eaton and Kortum (2002) apply a similar principle to determine an aggre-

and omitted variable biases have a measurable downward impact

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

II. A GLANCE AT THE DATA

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

growth of trade of countries that started to trade after 1970 in

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

where p j (l) is the price of product l in country j and Pj is the

This specification implies that every product has a constant de-

differences across firms in the same country. The inverse of a, 1/a,

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

This bilateral trade volume equals zero when ai j aL, because

14. Note that ai j + as fi j 0.

Equations (4)(7) provide a mapping from the income levels Yi ,

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

IV. EMPIRICAL FRAMEWORK

Our assumptions imply that Vi j can be expressed as (see (5))

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

where lowercase variables represent the natural logarithms of

where i = ( 1) pi + yi is a fixed effect of the importing coun-

variable is a function of the cutoff ai j , which is determined by

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

IV.A. Firm Selection into Export Markets

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

Let i j be the predicted probability of exports from j to i, using

where (k + 1)/( 1).

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

where u corr (ui j , i j )(u/ ) and ei j is an i.i.d. error term sat-

country characteristics. This misses the potentially important ef-

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

(Probit) (Probit) (Probit)

Distance 1.176 0.263 1.201 0.246 1.200 0.246

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

and introduce both importing and exporting country fixed effects.

same regional free trade agreement (FTA), they share a common

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

same regressions for the 1980s, adding bilateral controls when-

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

VI. TWO-STAGE ESTIMATION

We start with country-level data on the regulation costs of

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

everyone.30 Fixed exporter (and in the case of Japan, importer) ef-

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

1986 reduced sample

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

coefficient for ij are precisely estimated. The remaining results

Downloaded from http://qje.oxfordjournals.org/ at University of Edinburgh on June 18, 2012

country pair traded under di j , say Tij , then we could construct a

In the case of the polynomial approximation, 0 + w ij is replaced