Eaton 2011

AN ANATOMY OF INTERNATIONAL TRADE: EVIDENCE FROM FRENCH FIRMS
Author(s): Jonathan Eaton, Samuel Kortum and Francis Kramarz

Source: Econometrica , September 2011, Vol. 79, No. 5 (September 2011), pp. 1453-1498
Published by: The Econometric Society
Stable URL: http://www.jstor.com/stable/41237782
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
Wiley and The Econometric Society are collaborating with JSTOR to digitize, preserve and
extend access to Econometrica
This content downloaded from

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
All use subject to https://about.jstor.org/terms
Econometrica, Vol. 79, No. 5 (September, 2011), 1453-1498
AN ANATOMY OF INTERNATIONAL TRADE: EVIDENCE FROM

FRENCH FIRMS
By Jonathan Eaton, Samuel Kortum, and Francis Kramarz1

We examine the sales of French manufacturing firms in 113 destinations, includi
France itself. Several regularities stand out: (i) the number of French firms selling
a market, relative to French market share, increases systematically with market si
(ii) sales distributions are similar across markets of very different size and extent
French participation; (iii) average sales in France rise systematically with selling to
popular markets and to more markets. We adopt a model of firm heterogeneity
export participation which we estimate to match moments of the French data using
method of simulated moments. The results imply that over half the variation ac
firms in market entry can be attributed to a single dimension of underlying firm h
erogeneity: efficiency. Conditional on entry, underlying efficiency accounts for m
less of the variation in sales in any given market. We use our results to simulate the
fects of a 10 percent counterfactual decline in bilateral trade barriers on French fir
While total French sales rise by around $16 billion (U.S.), sales by the top decile
firms rise by nearly $23 billion (U.S.). Every lower decile experiences a drop in sa
due to selling less at home or exiting altogether.
Keywords: Export destinations, firms, efficiency, sales distribution, simulated m

ments.
1. INTRODUCTION
We exploit detailed data on the exports of French firms to confront a

new generation of trade theories. These theories resurrect technological het-
erogeneity as the force that drives international trade. In Eaton and Kortum
(2002), differences in efficiencies across countries in making different goods
determine aggregate bilateral trade flows. Since they focused only on aggre-
gate data, underlying heterogeneity across individual producers remains hid-
den. Subsequently, Melitz (2003) and Bernard, Eaton, Jensen, and Kortum
(2003; henceforth BEJK) have developed models in which firm heterogeneity
explicitly underlies comparative advantage. An implication is that data on indi-
vidual firms can provide another window on the determinants of international
trade.
On the purely empirical side, a literature has established a number of regu-
larities about firms in trade.2 Another literature has modeled and estimated the
!We thank Costas Arkolakis, Christopher Flinn, Jeremy Fox, Xavier d'Haultfoeuille, and
Azeem Shaikh for extremely useful comments, and Dan Lu and Sebastian Sotelo for research
assistance. The paper has benefitted enormously from the reports of three anonymous referees
and from comments received at many venues, in particular the April 2007 conference to celebrate
the scientific contributions of Robert E. Lucas, Jr. Eaton and Kortum gratefully acknowledge the
support of the National Science Foundation under Grant SES-0339085.
2For example, Bernard and Jensen (1995), for the United States, and Aw, Chung, and Roberts
(2000), for Taiwan and Korea, documented the size and productivity advantage of exporters.
© 2011 The Econometric Society DOI: 10.3982/ECTA8318

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
1454 J. EATON, S. KORTUM, AND F. KRAMARZ
export decision of individual firms in partial equilibrium.3 However, the task of

building a structure that can simultaneously embed behavior at the firm level
into aggregate relationships and dissect aggregate shocks into their firm-level
components remains incomplete. This paper seeks to further this mission.
To this end, we examine the sales of French manufacturing firms in 113 des-
tinations, including France itself. Combining these data with observations on
aggregate trade and production reveals striking regularities in (i) patterns of
entry across markets, (ii) the distribution of sales across markets, (iii) how ex-
port participation connects with sales at home, and (iv) how sales abroad relate
to sales at home.
We adopt the approach of Melitz (2003), as augmented by Helpman, Melitz,
and Yeaple (2004) and Chaney (2008), as a basic framework for understanding
these relationships. Core elements of the model are that firms' efficiencies fol-
low a Pareto distribution, demand is Dixit-Stiglitz, and markets are separated
by iceberg trade barriers and require a fixed cost of entry. The model is the
simplest one we can think of that can square with the facts.
The basic model fails, however, to come to terms with some features of the
data: (i) Firms do not enter markets according to an exact hierarchy, (ii) Their
sales where they do enter deviate from the exact correlations the basic model
insists on. (iii) Firms that export sell too much in France, (iv) In the typical
destination, there are too many firms selling small amounts.
To reconcile the basic model with the first two failures, we introduce mar-
ket and firm-specific heterogeneity in entry costs and demand. We deal with
the last two by incorporating Arkolakis's (2010) formulation of market access
costs. The extended model, while remaining very parsimonious and transpar-
ent, is one that we can connect more formally to the data. We describe how the
model can be simulated and we estimate its main parameters using the method
of simulated moments.
Our parameter estimates imply that the forces underlying the basic model
remain powerful. Simply knowing a firm's efficiency improves our ability to ex-
plain the probability it sells in any market by 57 percent. Conditional on a firm
selling in a market, knowing its efficiency improves our ability to predict how
much it sells there, but by much less. While these results leave much to be ex-
plained by the idiosyncratic interaction between individual firms and markets,
they tell us that any theory that ignores features of the firm that are universal
across markets misses much.
We conclude by using our parameterized model to examine a world with
10 percent lower trade barriers. To do so, we embed our model, with our pa-
rameter estimates, into a general equilibrium framework, calibrating it to ag-
gregate data on production and bilateral trade. A striking finding is the extent
to which lower trade barriers, while raising welfare in every country, favor the
largest firms at the expense of others. Total output of French firms rises by
3A pioneering paper here is Roberts and Tybout (1997).

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
ANATOMY OF INTERNATIONAL TRADE 1455
3.8 percent, with all of the growth accounted for by firms in the top
in every other decile fall. Import competition leads to the exit of
of firms, 43 percent of which are in the smallest decile.
Section 2, which follows, explores four empirical regularities. Wi
mind, we turn, in Section 3, to a model of exporting by heterogen
Section 4 explains how we estimate the model and considers some
of the parameters. Section 5 explores the consequences of lower t
2. EMPIRICAL REGULARITIES
Our data are the sales, translated into U.S. dollars, of 230,423 French man-
ufacturing firms to 113 markets in 1986. (Table III in Section 5.3 lists the coun-
tries.) Among them, only 34,558 sell outside France. The firm that exports most
widely sells to 110 out of the 113 destinations. All but 523 of these firms indi-
cate positive sales in France. Since much of our analysis focuses on the rela-
tionship between exporting and activity in France, we exclude these firms from
much of our analysis, leaving 34,035 exporters also selling in France.4
We cut the data in four different ways, each revealing sharp regularities.
2.1. Market Entry
Figure 1 A plots the number of French manufacturing firms NnF selling to a

market against total manufacturing absorption Xn in that market across our
113 markets.5 While the number of firms selling to a market tends to increase
with the size of the market, the relationship is cloudy.
The relationship comes into focus, however, when the number of firms is
normalized by the share of France in a market. Figure IB continues to re-
port market size across the 113 destinations along the x axis. The y axis re-
places NnF, the number of French firms selling to a market, with NnF divided
by French market share, irnF, defined as
XnF
where XnF is total exports of our French firms to market n.
4Appendix A describes the data. Appendices, data, and programs are available in the Sup-
plemental Material (Eaton, Kortum, and Kramarz (2011)). Biscourp and Kramarz (2007) and
Eaton, Kortum, and Kramarz (2004; EKK) used the same sources. EKK (2004) partitioned firms
into 16 manufacturing sectors. While features vary across industries, enough similarity remains to
lead us to ignore the industry dimension here. If a firm's total exports declared to French customs
exceed its total sales from mandatory reports to the French fiscal administration, we treat the firm
as not selling in France. The 523 such firms represent 1.51 percent of all French exporters and
account for 1.23 percent of the total French exports to our 112 export destinations.
5 Manufacturing absorption is calculated as total production plus imports minus exports. See
EKK (2004) for details.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
Figure 1. - Entry and sales by market size.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
TABLE I
French Firms Exporting to the Seven Most Popular Destinations
Number of Fraction of
Export Destination Exporters Exporters
Belgium3 (BE) 17,699 0.520

Germany (DE) 14,579 0.428
Switzerland (CH) 14,173 0.416
Italy (IT) 10,643 0.313
United Kingdom (UK) 9752 0.287
Netherlands (NL) 8294 0.244
United States (US) 7608 0.224
Any destination (all French exporters) 34,035
a Belgium includes Luxembourg.
Note that the relationship is not only very tight, but linear in logs. Correcting
for market share pulls France from the position of a large positive outlier to a
slightly negative one. A regression line has a slope of 0.65.6
While the number of firms selling to a market rises with market size, so do
sales per firm. Figure 1С shows the 95th, 75th, 50th, and 25th percentile sales
in each market (on the y axis) against market size (on the x axis). The upward
drift is apparent across the board.
We now turn to firm entry into different sets of markets. As a starting point
for this examination, suppose firms obey a hierarchy in the sense that any firm
selling to the к + 1st most popular destination necessarily sells to the kth most
popular destination as well. Not surprisingly, firms are less orderly in their
choice of destinations. Consider exporters to the top seven foreign destina-
tions. Table I reports these destinations and the number of firms selling to
each, as well as the total number of exporters. The last column of the table re-
ports, for each top-seven destination, the unconditional probability of a French
exporter selling there.
Table II lists each of the strings of top-seven destinations that obey a hi-
erarchical structure, together with the number of firms selling to each string
(irrespective of their export activity outside the top 7). Overall, 27 percent of
exporters (9260/34,035) obey a hierarchy among these most popular destina-
tions. The next column of Table II uses the probabilities from Table I to predict
the number of firms selling to each hierarchical string, if selling in one market
is independent of selling in any other of the top seven. Under independence,
6 If we make the assumption that French firms do not vary systematically in size from other
(non-French) firms selling in a market, the measure on the у axis indicates the total number of
firms selling in a market. We can then interpret Figure IB as telling us how the number of sellers
varies with market size.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
TABLE II
French Firms Selling to Strings of Top-Seven Countries
Number of French Exporters
Export Stringa Data Under Independence Model
BEa 3988 1700 4417

BE-DE 863 1274 912
BE-DE-CH 579 909 402
BE-DE-CH-IT 330 414 275
BE-DE-CH-IT-UK 313 166 297
BE-DE-CH-IT-UK-NL 781 54 505
BE-DE-CH-IT-UK-NL-US 2406
Total 9260 4532 9648
aThe string BE means selling

selling to Belgium and Germany
the number of firms adh

see in the data.
2.2. Sales Distributions
Our second exercise expands on Figure 1С by looking at the entire distri-

bution of sales within individual markets. We plot the sales of each firm in a
particular market (relative to mean sales there) against the fraction of firms
selling in the market who sell at least that much.7 By doing so for all our 113
destinations, a remarkable similarity emerges. Figure 2 plots the results for
Belgium, France, Ireland, and the United States on common axes. The basic
shape is common for size distributions.8
7Following Gabaix and Ibragimov (2011), we construct the x axis as follows. Denote the rank
in terms of sales of French firm j in market n, among the NnF French firms selling there, as rn(j),
with the firm with the largest sales having rank 1. For each firm ;, we calculate (rn(;) - 0.5) /NnF.
To preserve confidentiality, our plots report the geometric mean of a group of adjacent sales. For
the top 1000 sales, a group contains four observations, with larger groups used for lower ranks.
8 Sales distributions are often associated with a Pareto distribution, at least in the upper tail
(Simon and Bonini (1958)). To interpret Figure 2 as distributions, let xqn be the ¿7th percentile
French sales in market n normalized by mean sales in that market. We can write
Pr[*„ <*«] = <?,
where xn is sales of a firm in market n relative to the mean. If the distribution is Pareto with
parameter a > 1 (so that the minimum sales relative to the mean is (a - l)/a), we have
1 ( ax" Va

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
Figure 2. - Sales distributions of French firm: Graphs by country.
2.3. Export Participation and Size in France
How does a firm's participation in export markets relate to i

France? We organize our firms in two different ways based on our
tion of their entry behavior above.
First, we group firms according to the minimum number of des
where they sell. All of our firms, of course, sell to at least one ma
none sells to all 113 destinations. Figure ЗА depicts average sale
on the у axis for the group of firms that sell to at least к market
the x axis. Note the near monotonicity with which sales in France ri
number of foreign markets served.
Figure 3B reports, on a log scale, average sales in France of firms
к or more markets against the number of firms selling to к or mor
The relationship is strikingly linear with a regression slope of -0.66
Second, we rank countries according to their popularity as destin
exports. The most popular destination is France itself, where all of
sell, followed by Belgium with 17,699 exporters. The least popular
where only 43 French firms sell. Figure 3C depicts average sales in
the у axis plotted against the number of firms selling to the A:th m
or
1п0ф =V1п(-
a ) a
)-iln(l-tf),
implying a straight line with slope -I/a. Considering only sales by
firms selling in the four destinations depicted in Figure 2, regress
gium), -0.87 (France), -0.69 (Ireland), and -0.82 (United States). N
tributions appear to deviate from a Pareto distribution, especially a

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
Figure 3. - Sales in France and market entry.
market on the x axis. The relationship is tight and linear in logs as in Figure 3B,
although slightly flatter, with a slope of -0.57. Firms selling to less popular
markets and to more markets systematically sell more in France.
Delving further into the French sales of exporters to markets of varying pop-
ularity, Figure 3D reports the 95th, 75th, 50th, and 25th percentile sales in
France (on the y axis) against the number of firms selling to each market. Note
the tendency of sales in France to rise with the unpopularity of a destination
across all percentiles (less systematically so for the 25th percentile).9
2.4. Export Intensity
Having looked separately at what exporters sell abroad and what they sell in
France, we now examine the ratio of the two. We introduce the concept of a
9We were able to observe the relationship between market popularity and sales in France for
the 1992 cross section as well. The analog (not shown) of Figure 3C is nearly identical. Further-
more, the changes between 1986 and 1992 in the number of French firms selling in a market
correlate negatively with changes in the mean sales in France of these firms. The only glaring
discrepancy is Iraq, where the number of French exporters plummeted between the two years,
while average sales in France did not skyrocket, as the relationship would dictate.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
Figure 4. - Distribution of export intensity, by market.
firm /'s normalized export intensity in market /г, which we define as
(XnF(j)/XnF)
(XFF{j)/XFF)
Here XnF(j) is French firm /'s sales in market n and XnF are aver
French firms in market n (XFF(j) and XFF are the corresponding
in France). Scaling by XnF normalizes by the effect of market n o
all French firms there. Scaling by XFF(j) normalizes by firm /'s si
Figure 4 plots the median and 95th percentile normalized export i
each foreign market n (on the y axis) against the number of firms sell
market (on the x axis) on log scales. Two aspects stand out: (i) As
becomes more popular, normalized export intensity rises. The slo
median is 0.39, but the relationship is noisy, (ii) Normalized "expor
for France itself is identically 1, while median export intensity i
usually 2 orders of magnitude or more below 1. Even among expo
sales abroad are small compared to sales at home.
3. THEORY
In seeking to interpret these relationships, we turn to a parsimonious model

which explains where firms sell and how much they sell there. The basic struc-
ture is monopolistic competition: Goods are differentiated, with each one cor-
responding to a firm. Selling in a market requires a fixed cost, while moving

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
goods from country to country incurs iceberg transport costs. Firms are het-
erogeneous in efficiency as well as in other characteristics, while countries vary
in size, location, and fixed cost of entry.10
3.1. Producer Heterogeneity
A potential producer of good j in country i has efficiency z¡(j). A bundle

of inputs there costs wi9 so that the unit cost of producing good j is Wi/z¿(j).
Countries are separated by iceberg trade costs, so that delivering 1 unit of a
good to country n from country i requires shipping dni > 1 units, where we
set da = 1 for all L Combining these terms, the unit cost to this producer of
delivering 1 unit of good j to country n from country i is
(1) c,,0, = - .
The measure of potential producers in country i who can prod

with efficiency at least z is
(2) ^{z) = TiZ-' z>0,

where в > 0 and 7) > 0 are parameters.11 Using (1), the measure of goods
can be delivered from country i to country n at unit cost below с is defin
(3) Мл,(с) = м-(^)=Ф«,св,

where
<Pni = TiiWidm)-6.
We now turn to demand and market structure in a typical destination.
10We go with Melitz's (2003) monopolistic competition approach rather than the Ric
framework with a fixed range of commodities used in BEJK (2003), since it more readily d
the feature that a larger market attracts more firms, as we see in our French data.
11 We follow Chaney (2008) in taking Г, as exogenous and Helpman, Melitz, and Yeaple (
and Chaney (2008) in treating the underlying heterogeneity in efficiency as Parete A
distribution of efficiencies can arise naturally from a dynamic process that is a history o
dependent shocks, as shown by Simon (1955), Gabaix (1999), and Luttmer (2007). The P
distribution is closely linked to the type II extreme value (Fréchet) distribution used in K
(1997), Eaton and Kortum (1999, 2002), and BEJK (2003). Say that the range of goods is
to the interval ; € [0, /] with the measure of goods produced with efficiency at least z gi
/^(z;/) =/{1 - txp[-(T/J)z~e]} (where / = 1 in these previous papers). This generaliz
allows us to stretch the range of goods while compressing the distribution of efficiencies f
given good. Taking the limit as / -> oo gives (2). (To take the limit, rewrite the expres
{1 - exp[-(r//)z-0]}//"1 and apply l'Hôpital's rule.)

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
3.2. Demand, Market Structure, and Entry
A market n contains a measure of potential buyers. To sell to a f

of them, a producer in country i selling good j in country n must in
cost:
(4) Eni(j) = sn(j)EniM(f).
Here en(j) is a fixed-cost shock specific to good j in market n and Eni is the
component of the cost shock faced by all sellers from country i in destination n.
The function M (/), the same across destinations, relates a seller's fixed cost of
entering a market to the share of consumers it reaches there. Any given buyer
in the market has a chance / of accessing the good, while / is the fraction of
buyers reached.
In what follows, we use the specification for M(f) derived by Arkolakis
(2010),
1_(1_ ni-i/A
where the parameter Л > 0 reflects the increasing cost of reaching a larger
fraction of potential buyers.12 This function has the desirable properties that
the cost of reaching zero buyers in a market is zero and that the total cost
increases continuously (and the marginal cost weakly increases) in the fraction
/ of buyers reached. This formulation can explain why some firms sell a very
small amount in a market while others stay out entirely.
Each potential buyer in market n has the same probability / of being reached
by a particular seller which is independent across sellers. Hence each buyer
can purchase the same measure of goods, although the particular goods in
question vary across buyers. Buyers combine goods according to a constant
elasticity of substitution (CES) aggregator with elasticity a, where we require
в + 1 > a > 1. Hence we can write the aggregate demand for good j, if it has
price p and reaches a fraction / of the buyers in market /t, as
(5) Xn{j) = an{j)fXn(^'
12By l'Hôpital's rule, at Л = 1, the function becomes M(f) = -ln(l - /). Arkolakis (2010)
provided an extensive discussion of this functional form, deriving it from a model of the micro-
foundations of marketing. He parameterized the function in terms of ß = l/À. The case ß > 0
corresponds to an increasing marginal cost of reaching additional buyers, which always results
in an outcome with 0 < / < 1. The case ß < 0 means a constant or decreasing marginal cost of
reaching additional buyers, in which case the firm would go to a corner (/ = 0 or / = 1) as in
Melitz (2003). Our restriction À > 0 thus covers all cases (including Melitz as Л -> oo) that we
could observe.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
where Xn is total spending there. The term an(j) reflects an exogenous demand
shock specific to good j in market n. The term Pn is the CES price index, which
we derive below.
The producer of good j from country i selling in market n with a unit cost
of cn(j), charging a price p and reaching a fraction / of buyers, earns a profit
(6) nni(p, /) = (l - ^)a»Wf(jr) X» - en(j)EniM(f).

To maximize (6), the producer sets the standard Dixit- Stiglitz (1977) markup
over unit cost,
(7) PnU) = mcn{j),

where
m =
a- 1
It seeks a fraction
(8, /.,0) = maxj1.[4,0)^.(^>)-<'-)]-*,0¡

of buyers in the market, where
is the entry shock in market n given by the ratio of the demand shock to the
fixed-cost shock. Note that it will not sell at all, hence avoiding any fixed cost
there, if
We can now describe a seller's behavior in market n in terms of its unit cost
cn(j) = c, demand shock an(j) = a, and entry shock r)n(j) = 17. From the con-
dition above, a firm enters market n if and only if
(9) с<сЛ1-(т»),
where
' l/(o--l) p
(Y ИН crEniJ ' =■m

crEniJ mp

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
We can use the expression for (10) to simplify expression (8) for
of buyers reached by a producer with unit cost с < сш(т/)
/ vA(cr-l)
(u> Л("'с)=1-Ы / ■
Substituting (7), (11), and (10) into (5) gives
r / '^h/ c '-(a~l)
(12) Хы(а,71,с) = - 1- =- - ^7"T °-£-
Conditioning on с„/(т}), а/ г] replaces a in (5) as the shock to sales.

Since it charges a markup m = a/ (a - 1) over unit cost, its total g
is simply Xm (a, tj, c)/a. Some of this profit is eaten up by its fixed cos
from (4) and (11), is
(13) Eni(a, г], c) = -Eni
7] 1 - I/A
To summarize, the relevant characteristics of market

ers are total purchases Xn, the price index Pn, and, fo
the common component of the fixed cost Eni. The pa
tential seller of product j in market n is captured b
unit cost cn(j), the demand shocks an(j), and the entr
an(j) and rin(j) as the realizations of producer-specifi
joint density g(a, rj) that is the same across destinat
CnO').13
Equations (9) and (10), which govern entry, and (12), which governs sales
conditional on entry, link our theory to the data on French firms' entry and
sales in different markets of the world described in Section 2. Before returning
to the data, however, we need to solve for the price index Pn in each mar-
ket.
3.3. The Price Index
As described above, each buyer in market n has access to the same measure
of goods (even though they are not necessarily the same goods). Every buyer
faces the same probability /„/(tj, c) of purchasing a good with cost с and entry
shock 7] for any value of a. Hence we can write the price index faced by a
13We require E[rje/i(T-l)] and E[ar)m<T-l)]-1] both to be finite. For any given g(a, rj), these
restrictions imply an upper bound on the parameter в.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
representative buyer in market n as
Г с С /А Гш(7?) '
Pn = 'J J ' ? J afni(7i' C)qX~G dlXni{C)
To solve, we use (3), (10), (11), and the laws of integ

Г N Г/ Y ' V(o-D p -|0-(<r-D
[Г -//■5>[('£) Г N Г/ Y ' V(o-D |] p
where
в в
K° ~ ö-(o-
Moving Pn to o
(14) P^miK^y^X^-^-»,
where
(15) Kl = ко ff ari^-W-Vgia, tfdaďq

and
(16) *в = J2^ni((TEnir[e-^-l^-l'
Note that the price index has an elasticity of(l/ö)-l/(o--l) with respect to
total expenditure (given the terms in Фп). Our restriction that в > cr - 1 makes
the effect negative: A larger market enjoys lower prices, for reasons similar
to the price index effect in Krugman (1980) and common across models of
monopolistic competition.14
14 In models of monopolistic competition with homogeneous firms, consumers anywhere con-

sume all varieties regardless of where they are made. Since more varieties are produced in a
larger market, local consumers benefit from their ability to buy these goods without incurring
trade costs. With heterogeneous firms and a fixed cost of market entry, an additional benefit to
consumers in a large market is greater variety.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
3.4. Entry, Sales, and Fixed Costs
An individual firm enters market n if its cost с is below the thr

by (10), which we can now rewrite using (14) as
(17) čl„-(4) = V/(<r-1)(^rJ {*Eniy^'

For firms from country i with a given value of 77, a measure /xm
the entry hurdle.
To obtain the total measure of firms from i that sell in n, den
integrate across the marginal density #2(17),
(18) Jni= />[^|.(č^4))]ft(4)dí? = -^^,

where
(19) K2 = Jrìe^-ì)g2(rì)drì
and
(20) TTni =
From (12), integra

ing (17), firms from
^ш№ V) - arì
Kl
in market n.
To obtain total sales by firms from country i in market n, Xni9 we integrate
across the joint density g(a, 17) to get
Xni - ъП1Хп.
Hence the trade share of source i in market n is just irniP
15 Of the total measure Jn of firms selling in n, the fraction from / is
У ^nk/Enk

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
To obtain an expression for the fixed costs incurred by firms from i in n,

given their values of a and 77, we integrate (13) over the measure with с below
čn/(r/)toget
Eni(a, v) = aV^™v™(^) VV/L[0/(<7-l)] '- + A-lJ ,

сгк' /L[0/(<7-l)]
Integrating across the joint density g(a, 17), total fixed costs incurred in market
n by firms from i are
Eni = - - [в- (a -I)].

crO
Summing across sources /, the total fixed costs incurred in market n are simply
(21)
(21) (21)
(21) EE En-
En- в-^-УХ»
в
Note that En (spending on fixed costs) does not depend on the Eni, the country-
pair components of the fixed cost per firm, just as in standard models of mo-
nopolistic competition. A drop in Eni leads to more entry and ultimately the
same total spending on fixed costs.
If total spending in a market is Xn, then gross profits earned by firms in that
market are Xn/cr. If firms were homogeneous, then fixed costs would fully dis-
sipate gross profits. With producer heterogeneity, firms with a unit cost below
the entry cutoff in a market retain a net profit there. Total entry costs are a
fraction [в - (a - ')]/ в of gross profits, and net profits are Хп/(тв).
3.5. A Streamlined Representation
We now employ a change of variables that simplifies the model in two re-
spects. First, it allows us to characterize unit cost heterogeneity in terms of a
uniform measure. Second, it allows us to consolidate parameters.
To isolate the heterogeneous component of unit costs, we transform the ef-
ficiency of any potential producer in France as
(22) u(j) = TFzF(jYe.
If Eni does not vary according to 1, then a country's share in the measure of firms selling in n and
its share in total sales there are both тгш, where
KW*)-9
Пп* = -ц

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
We refer to u(j) as firm /'s standardized unit cost. From (2), the
firms with standardized unit cost below и equals the measure wit
above (TF/u)1/e, which is simply fizF((TF/u)1/e) = u. Hence stand
have a uniform measure that does not depend on any parameter.
Substituting (22) into (1) and using (20), we can write unit cost
in terms of u{j) as
(23) СпЛ])=^п)=Ш ■
Associated with the entry hurdle cnF (17) is a standardized entry h
satisfying
ПАЛ (24) - < Л = í»nF(v)'1/e

ПАЛ (24) - Cnf(T,) < Л = í»nF(v)'1/e (- ] .
Firm / enters market n if its u(j) and r'n(j) satisfy
(25) u(j) <unF(Vn(D) = (-^^VnijY,

where
a -I
Conditional on firm /'s passing this hurdle, we can use (2

rewrite firm /'s sales in market /i, expression (12), in terms of
(26) Х.ЛП = «.(Ji - f ^^fl(^-V

Equations (25) and (26) reformulate the entry and sales equ
(12) in terms of u(j) rather than cn{j).
Since standardized unit cost u(j) applies across all markets
core of a firm's underlying efficiency as it applies to its entry
ferent markets. Notice that in reformulating the model as (25
parameters J?, a, and EnF enter only collectively through
parameter в translates unobserved heterogeneity in u(j) int
erogeneity in sales. A higher value of в implies less heterogeneit
while a higher value of a means that a given level of heterogene
translates into greater heterogeneity in sales. By observing just
we are able to identify only Ö.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
3.6. Connecting the Model to the Empirical Regularities
We now show how the model can deliver the features of the data about entry
and sales described in Section 2. We equate the measure of French firms JnF
selling in each destination with the actual (integer) number NnF and equate
XnF with their average sales there.
3.6.1. Entry
From (18), we get
/r*~' Nnf
' ')
7TnF K' (ThnF it
a relationship bet
to French market
Figure IB. The fac
less than 1 sugges
but not proportio
employ (27) to ca
(28) o-EnF =
Ki NnF ki
directly from the data.16

Using (28), we can write (25) as
(29) u(j) < unAVn(D) = -VnijY.

f<2
Without variation in the firm-specific and market specific entry shock r¡n(j),
(29) implies efficiency is all that matters for entry, dictating a deterministic
ranking of destinations with a less efficient firm (with a higher u(j)) selling to
a subset of the destinations served by any more efficient firm. Hence deviations
from market hierarchies, as we see in Table II, identify variation in rjn(j).
16We can use equation (28) to infer how fixed costs vary with country characteristics. Simply
regressing ~XnF against our market size measure Xn (both in logs) yields a coefficient of 0.35,
(1 minus the slope in Figure IB), in which NnF/7rnF rises with market size with an elasticity of
0.65. (The connection between the two regressions is a result of the accounting identity: XnF *
NnF/tTnF = Xn-) If gross domestic product (GDP) per capita is added to the regression, it has a
negative effect on entry costs. French data from 1992, and data from Denmark (from Pedersen
(2009)) and from Uruguay (compiled by Raul Sampognaro) show similar results for market size
but not a robust effect of GDP per capita. We also find that mean sales are higher in the home
country. Appendix В reports these results.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
3.6.2. Sales in a Market
Conditional on a firm's entry into market n, the term
(30) vnF(j) = _ u(j)

_ UnFÍVn(j))
is distributed uniformly on [0, 1]. Replacing u{j) with vnF(j) in expression (26),
and exploiting (28), we can write sales as
(31) XnF(j) = en(j)[l - vnA]Yre]vnF(jrire^XnF.

Not only does vnF have the same distribution in each market n, so does sn.11
Hence the distribution of sales in any market n is identical up to a scaling fac-
tor equal to XnF (reflecting variation in aEnF), consistent with the common
shapes of sales distributions exhibited in Figure 2. If the only source of vari-
ation in sales were the term VnFij)'1^, the sales distribution would be Pareto
with parameter 0. The term in square brackets, however, implies a downward
deviation from a Pareto distribution that is more pronounced as vnF(j) ap-
proaches 1, consistent with the curvature in the lower end of the sales distribu-
tions that we observe in Figure 2. The sales distribution also inherits properties
of the distribution of en(j).
3.6.3. Sales in France Conditional on Entry in a Foreign Market
We can also look at the sales in France of French firms selling to any mar-
ket n. To condition on these firms, we take (31) as it applies to France and use
(30) and (29) to replace vFF(j) with vnF(j):
(32) XpF(M _*£[,_ Vn(j)l

Vn(j)l „,И;у^у'7^)л1
'NFFJ'NFFJ 'Vf(j)JJ
'Vf(j)J J
x vnF(j) 1/u[ - - -Xpp.

'NFFJ кх
Since both vnF(j) and r]n(j) have the same distributions across destinations
the only systematic source of variation across n is NnF.
17To see that the distribution of en(j) is the same in any n, consider the joint density of a
7) conditional on entry into market n
unF{r)) rf
-
/ unF(vf)g2(ri')drif
which does not depend on n. Th

with typically better than avera

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
Consider first the presence of NnF in the term in square brackets, repre-
senting the fraction of buyers reached in France. Since NnF/NFF is near zero
everywhere but France, the term in square brackets is close to 1 for all пф¥.
Hence the relationship between NnF and XFF(j) is dominated by the appear-
ance of NnF outside the square bracket, implying that firms' sales in France
fall with NnF with an elasticity of -I/o. Interpreting Figure 3C in terms of
Equation (32), the slope of -0.57 implies a 0 = 1.75.
Expression (32) also suggests how we can identify other parameters of the
model. The gap between the percentiles in Figure 3D is governed by the vari-
ation in the demand shock aF(j) in France together with the variation in the
entry shock r¡n(j) in country n.
Together (31) and (32) reconcile the near log linearity of sales in France with
NnF and the extreme curvature at the lower end of the sales distribution in any
given market. An exporting firm may be close to the entry hurdle in the ex-
port market and hence sells to a small fraction of buyers there, while reaching
most consumers at home. Hence looking at the home sales of exporters iso-
lates firms that reach most of the French market. These equations also explain
why France itself is somewhat below the trend line in Figure ЗА and B: The
many nonexporting firms that reach just a small fraction of the French market
appear only in the data point for France.
3.6.4. Normalized Export Intensity
Finally, we can calculate firm ;'s normalized export intensity in market n:
{óá) XFF{j)/'xFF)
= OjjO) Г 1 - VnF(J)x/ì 1 (NnF_ '1/?
'NffJ 'Vf(j)J
Note first how the presence of the sales shock an(j) accommodates random
variation in sales in different markets conditional on entry.
As in (32), the only systematic source of cross-country variation on the right-
hand side of (33) is NnF. The relationship is consistent with three features of
Figure 4. First, trivially, the observation for France is identically 1. Second,
normalized export intensity is substantially below 1 for destinations served by
only a small fraction of French firms, as is the case for all foreign markets.
Third, normalized export intensity increases with the number of French firms
selling there. According to (33) the elasticity of normalized export intensity
with respect to NnF/NFF is I/o (ignoring NnF/NFF's role in the denominator of
the term in the square bracket, which is tiny since NnF/NFF is close to zero for

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
пф¥). The slope coefficient of 0.38 reported in Section 2.4 sug

ofÖ = 2.63.18
4. ESTIMATION
We estimate the parameters of the model by the method of simulated mo-

ments. We simulate firms that make it into at least one foreign market and into
France as well.19 Given parameter estimates, we later explore the implications
of the model for nonexporters as well.
We first complete our parameterization of the model. Second, we explain
how we simulate a set of artificial French exporters given a particular set of pa-
rameter values, with each firm assigned a cost draw u, and an a and 77 in each
market. Third, we describe how we calculate a set of moments from these arti-
ficial data to compare with moments from the actual data. Finally, we explain
our estimation procedure, report our results, and examine the model's fit.
4.1. Parameterization
To complete the specification, we assume that g(a, 17) is joint log normal.
Specifically, In a and In 17 are normally distributed with zero means and vari-
ances crl and tf, and correlation p 20 Under these assumptions, we can write
(15) and (19) as
(34) K1JJ- „ ? 1 exp! <га + 2рааач(в-1) + ач(в-т 1
18Equations (32) and (33) apply to the latent sales in France of firms that sell in n but do not
enter France. In Figure 3C and D we can only look at the firms that sell in both places, of course.
Since the French share in France is so much larger than the French share elsewhere, our theory
predicts that a French firm selling in another market but not in France is very unlikely. Indeed,
the number of such firms is small.
19The reason for the selling-in-France requirement is that key moments in our estimation pro-
cedure involve sales in France by exporters, which we can compute only for firms that enter the
home market. The reason for the foreign-market requirement is that firms selling only in France
are very numerous, so that capturing them would consume a large portion of simulation draws.
However, since their activity is so limited, they add little to parameter identification. We also
estimated the model matching moments of nonexporting firms. Coefficient estimates were simi-
lar to those we report below, but the estimation algorithm, given estimation time, was much less
precise.
20 Since the entry-cost shock is given by In e = In a - In 17, the implied variance of the entry-cost
shock is
which is decreasing in p.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
and
(35) fc2 = exp '-£- •

As above, in estimating the model, we equate the measure of French firms
JnF selling in each destination with the actual (integer) number NnF and equate
XnF with their average sales there.21 We are left with only five parameters to
estimate:
0 = {в,А,(та,(тгпр}.
For a given ©, we use (28) to back out the cluster of parameters aEnF using our
data on XnF = XnF/NnF, and the к' and к2 implied by (34) and (35). Similarly,
we use (29) to back out a firm's entry hurdle in each market wnF(r/n) given its
7]n and the к2 implied by (35).
4.2. Simulation Algorithm

To estimate parameters, to assess the implications of those estimates, and
to perform counterfactual experiments, we need to construct sets of artificial
French firms that operate as the model tells them, given some ©. We denote
an artificial French exporter by s and the number of such exporters by 5. The
number S does not bear any relationship to the number of actual French ex-
porters. A larger S implies less sampling variation in our simulations.
As we search over different parameters ®, we want to hold fixed the real-
izations of the stochastic components of the model. Hence, prior to running
any simulations, (i) we draw S realizations of v(s) independently from the uni-
form distribution t/[0, 1], putting them aside to construct standardized unit
cost u(s) below, and (ii) we draw S x 113 realizations of an(s) and hn(s) inde-
pendently from the standard normal distribution N(09 1), putting them aside
to construct the an(s) and r}n(s) below.
A given simulation of the model requires a set of parameters 0, data for
each destination n on total sales XnF by French exporters, and the number NnF
of French firms selling there. It involves the following eight steps:
Step 1. Using (34) and (35), we calculate ki and к2.
Step 2. Using (28), we calculate vEnF for each destination n.
Step 3. We use the an(sYs and hn(sYs to construct S x 113 realizations for
each of in an(s) and In r]n(s) as
rinan(s)l = [(Tay/'-p2 aapl [an(s)~'

|_1пть(5)_Г = L 0 (Tr, 'lhn(s) У
21 The model predicts that some French firms export but do not sell domestically. Consequently,
the data for NnF and XnF that we condition on in our estimation include the 523 French exporters
(mentioned in footnote 4) who do not enter the domestic market.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
Step 4. We construct the S x 113 entry hurdles
(36) un(s) = -Vn(s)'
where ~un(s) stands for unF(Vn(s))-

Step 5. We calculate
ux(s) = max{un(s)},
пф¥
the maximum и consistent with exporting somewhere, and
(37) w(s) = min{wF(s), ux(s)},

the maximum и consistent with selling in France and exporting somewhere.
Step 6. To simulate exporters that sell in France, u(s) should be a realiza-
tion from the uniform distribution over the interval [0, T¿(s)]. Therefore, we
construct
u(s) = v(s)u(s).
using the v(s)'s that were drawn prior to the simulation.

Step 7. In the model, a measure w of firms have standardized unit cost be-
low w. Our artificial French exporter s therefore gets an importance weight
w(s). This importance weight is used to construct statistics on artificial French
exporters that relate to statistics on actual French exporters.22
Step 8. We calculate 8nF(s), which indicates whether artificial exporter s en-
ters market n, as determined by the entry hurdles
8nF(s)-'0, otherwise.
Wherever 8nF(s) = 1, we calculate sales as
(38) w-^i-mfim)-"^. Vn(s)l 'un(s)J

Vn(s)l 'un(s)J ]'un(s)J]'un(s)J
This procedure gives us the behavior of 5 artificial French exporters. We
know three things about each one: where it sells, SnF(s), how much it sells
there, XnF(s), and its importance weight, ~u(s). From these terms, we can com-
pute any moment that could have been constructed from the actual French
data.23
22See Gouriéroux and Monfort (1994, Chap. 5) for a discussion of the use of importance
weights in simulation.
23 In principle, in any finite sample, the number of simulated firms that overcome the entry
hurdle for a destination n could be zero, even though the distribution of efficiencies is unbounded

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
4.3. Moments
For a candidate value ®, we use the algorithm above to simulate the sales of
500,000 artificial French exporting firms in 113 markets. From these artificial
data, we compute a vector of moments m(0 ) analogous to particular moments
m in the actual data.
Our moments are the number of firms that fall into sets of exhaustive and
mutually exclusive bins, where the number of firms in each bin is counted in the
data and is simulated from the model. Let NkJ>e the number of firms achieving
some outcome к in the actual data and let Nk be the corresponding number
in the simulated data. Using 8k(s)^as an indicator for when artificial firm s
achieves outcome k, we calculate Nk as
15
(39) Ñk = -J2»(s)8k(s).
We now describe the moments that we seek to match.24

We have chosen our moments to capture the four features of French firms'
behavior described in Section 2:
• The first set of moments relate to the entry strings discussed in Section 2.1.
We compute the proportion w*(l; 0) of simulated exporters selling to each
possible combination к of the seven most popular export destinations (listed in
Table I). One possibility is exporting yet selling to none of the top seven, giving
us 27 possible combinations (so that к - 1, . . . , 128). The corresponding mo-
ments from the actual data are simply the proportion mk{') of exporters selling
to combination k. Stacking these proportions gives us m(l; 0) and ra(l), each
with 128 elements (subject to 1 adding up constraint).
• The second set of moments relate to the sales distributions presented in
Section 2.2. For firms selling in each of the 112 export destinations л, we com-
pute the #th percentile sales sqn{2) in that market (i.e., the level of sales such
that a fraction q of firms selling in n sells less than sqn{2)) for q = 50, 75, 95.
Using these sq(2), we assign firms that sell in n into four mutually exclusive
and exhaustive bins determined by these three sales levels. We compute the
proportions mn(2; 0) of artificial firms falling into each bin analogous to the
actual proportion mn(2) = (0.5, 0.25, 0.2, 0.05 )'. Stacking across the 112 coun-
tries gives us m(2; 0) and m(2), each with 448 elements (subject to 112 adding-
up constraints).
• The third set of moments relate to the sales in France of exporting firms as
discussed in Section 2.3. For firms selling in each of the 112 export destinations
from above. Helpman, Melitz, and Rubinstein (2008) accounted for zeros by truncating the upper
tail of the Pareto distribution. Zero exports to a destination then arise simply because not even
the most efficient firm could surmount the entry barrier there.
^Notice, from (36), (37), and (19), that the average of the importance weights u(s) is the
simulated number of French firms that export and sell in France.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
n, we compute the qth percentile sales sqn0) in France for q = 50

ceeding as above, we get ra(3; 0) and ra(3), each with 448 elem
to 112 adding-up constraints).
• The fourth set of moments relate to normalized export intensi
as discussed in Section 2.4. For firms selling in each of the 112 ex
tions n, we compute the qth percentile ratio sqn(4) of sales in n to
for q = 50, 75. Proceeding as above, we get ra(4; 0) and ra(4), e
elements (subject to 112 adding-up constraints).
For the last three sets, we emphasize higher percentiles becaus
pear less noisy in the data and (ii) account for much more of tota
Stacking the four sets of moments gives us a 1360-element vec
tions between the moments of the actual and artificial data:
"m(l)-m(l,@)"
m(3)-m(3,0)
_m(4)-m(4,©)_
By inserting the actual data on XnF (to get aEnF) and NnF (to get un(s))
in our simulation routine, we are ignoring sampling error in these measures.
Inserting XnF has no effect on our estimate of 0. The reason is that a change
in crEnF shifts the sales in n of each artificial firm, the mean sales XnF, and
the percentiles of sales in that market, all by the same proportion, leaving our
moments unchanged. We only need XnF to get estimates of aEnF given 0.
Our estimate of 0 does, however, depend on the data for NnF. Appendix С
reports on Monte Carlo simulations examining the sensitivity of our parameter
estimates to this form of sampling error.25
4.4. Estimation Procedure
We base our estimation procedure on the moment condition
E[y(0o)] = O,
where 0O is the true value of 0. We thus seek a 0 that achieves
0 = argmin{y(@yWy(©)},
25The coefficient of variation of NnF is approximated by '/y/NnF, which is highest for Nepal,
with 43 sellers, at 0.152. The median number of sellers is 686 (Malaysia), implying a coefficient
of variation of 0.038.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
where W is a 1360 x 1360 weighting matrix.26 We search for 0 using the simu-
lated annealing algorithm.27 At each function evaluation involving a new value
of 0, we compute a set of 500,000 artificial firms and construct the moments
for them as described above. The simulated annealing algorithm converges in
1-3 days on a standard PC.
We calculate standard errors using a bootstrap technique, taking into ac-
count both sampling error and simulation error.28
26The weighting matrix is the generalized inverse of the estimated variance-covariance matrix
il of the 1360 moments calculated from the data m. We calculate il using the following bootstrap
procedure: (i) We resample, with replacement, 229,900 firms from our initial data set 2000 times,
(ii) For each resampling b, we calculate mb, the proportion of firms that fall into each of the
1360 bins, holding the destination strings fixed to calculate mb{') and holding the st(r) fixed to
calculate ть{т) for т = 2, 3, 4. (iii) We calculate
.. 2000
Because of the adding-up constraints, this matrix has rank 1023, forcing
inverse to compute W.
27Goffe, Ferrier, and Rogers (1994) described the algorithm. We u
specifically for GAUSS by Goffe (1996).
¿* To account for sampling error, each bootstrap b replaces m with a
for simulation error, each bootstrap b samples a new set of 500,000 vb% a
in Section 4.2, thus generating a new mb(0). (Just like mb, mb is calculat
defined from the actual data.) Defining yb{&) = mb - mb(&), for each b,
Bb = argmin{y46>)'W/(<9)}
using the same simulated annealing procedure. Doing this exercise 25 tim
1 25
b='
and take the square roots of the diagonal elements as the standard errors. Since we pursu
bootstrapping procedure only to calculate standard errors rather than to perform tests, we
recenter the moments to account for the initial misfit of our model. Recentering would
setting
yb(0) = mb - ть(в) - (га - т(в))
above. In fact, our experiments with recentered moments yielded similar estimates of the stan-
dard errors. See Horowitz (2001) for an authoritative explanation.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
4.5. Results
The best fit is achieved at the parameter values (with bootstrapped standard
errors in parentheses)
0 Л (Ta cr-q p
~2M Õ91 L69 Õ34 -0.65
(0.10) (0.12) (0.03) (0.01) (0.03)
As a check on our procedure, given our sample size, we conduct a Monte

analysis, which is described in Appendix С A basic finding is that the stan
errors above are good indicators of the ability of our procedure to rec
parameters. We also analyze the sensitivity of our results, as described in
pendix D, to different moments. A basic finding is that the results are lar
insensitive to the alternatives we explore.
We turn to some implications of our parameter estimates.
Our discussion in Section 3.6 foreshadowed our estimate of 0, which
between the values implied by the slopes in Figure 3C and Figure 4. F
equations (31), (29), and (30), the characteristic of a firm determining
entry and sales conditional on entry is i;"1^, where v ~ f/[0, 1]. Our estim
of в implies that the ratio of the 75th to the 25th percentile of this term is 1.5
Another way to assess the magnitude of в is by its implication for aggre
fixed costs of entry. Using expression (21), our estimate of 2.46 implies t
fixed costs dissipate 59 percent of gross profit in any destination.
Our estimate of cra implies enormous idiosyncratic variation in a firm's
across destinations. In particular, the ratio of the 75th to the 25th perce
of the sales shock a is 9.78. In contrast, our estimate of av means much l
idiosyncratic variation in the entry shock 17, with the ratio of the 75th to
percentile equal to 1.58. Given aa and av, the variance of sales within a mar
decreases in p, as can be seen from equation (38). Hence the negative estim
of p reflects high variation of sales in a market.
A feature of the data is the entry of firms into markets where they sell
little, as seen in Figure 1С. Two features of our estimates reconcile these s
sales with a fixed cost of entry. First, our estimate of Л, which is close t
means that a firm that is close to the entry cutoff incurs a very small entry co
Second, the negative covariance between the sales and entry shocks explai
why a firm with a given и might enter a market and sell relatively little.
first feature applies to firms that differ systematically in their efficiency, wh
the second applies to the luck of the draw in individual markets.
29Arkolakis (2010) found a value around 1, consistent with various observations from sev
countries.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
4.6. Model Fit
We can evaluate the model by seeing how well it replicates features of the
data described in Section 2. To glean a set of predictions from our model, we
use our parameter estimates 0 to simulate a set of artificial firms including
nonexporters.30 We then compare four features of these artificial firms with
corresponding features of the actual firms.
Entry
Since our estimation routine conditions entry hurdles on the actual number
of French firms selling in each market, our simulation would hit these numbers
were it not for simulation error. The total number of exporters is a different
matter since the model determines the extent to which the same firms are sell-
ing to multiple countries. We simulate 31,852 exporters, somewhat below the
actual number of 34,035. Table II displays all the export strings that obey a hi-
erarchy out of the 128 subsets of the seven most popular export destinations.
The Data column is the actual number of French firms selling to that string
of countries, while the last column displays the simulated number. In the ac-
tual data, 27.2 percent of exporters adhere to hierarchies compared with 30.3
percent in the model simulation. In addition the model captures very closely
the number of firms selling to each of the seven different strings that obey a
hierarchy.
Sales in a Market
Equation (31) motivates Figure 5 A, which plots the simulated (x's) and the
actual (circles) values of the median and the 95th percentile sales to each mar-
ket against actual mean French sales in that market. The model captures very
well both the distance between the two percentiles in any given market and the
variation of each percentile across markets. The model also nearly matches the
amount of noise in these percentiles, especially in markets where mean sales
are small.
Sales in France Conditional on Entry in a Foreign Market
Equation (32) motivates Figure 5B, which plots the median and the 95th
percentile sales in France of firms selling to each market against the actual
number of firms selling there. Again, the model picks up the spread in the
distribution as well as the slope. It also captures the fact that the datum point
for France is below the line, reflecting the role of Л. The model understates
noise in these percentiles in markets served by a small number of French firms.
30Here we simulate the behavior of S = 230,000 artificial firms, both nonexporters and ex-
porters that sell in France, to mimic more closely features of the raw data behind our analysis.
Thus in Step 5 in the simulation algorithm, we reset u(s) = TiF(s).

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
Figure 5.- Model versus data.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
Export Intensity
Equation (33) motivates Figure 5C, which plots median normalized export
intensity in each market against the actual number of French firms selling
there. The model picks up the low magnitude of normalized export intensity
and how it varies with the number of firms selling in a market. Despite our high
estimate of aa, however, the model understates the noisiness of the relation-
ship.
4.7. Sources of Variation
In our model, variation across firms in entry and sales reflects both differ-
ences in their underlying efficiency, which applies across all markets, and id-
iosyncratic entry and sales shocks in individual markets. We ask how much of
the variation in entry and in sales can be explained by the universal rather than
the idiosyncratic components.
4.7.1. Variation in Entry
We first calculate the fraction of the variance of entry in each market that
can be explained by the cost draw и alone. A natural measure (similar to R2 in
a regression) of the explanatory power of the firm's cost draw for market entry
is
RE E[Vnc(u)]
VП
where Vf is the variance of entry and E[Vf(u)] is the expected variance of

entry given the firm's standardized unit cost u?x
We simulated the term E[Vf(u)] using the techniques employed in our es-
timation routine with 230,000 simulated firms and obtained a value of Щ for
each of our 112 export markets. Looking across markets, the results indicate
that we can attribute on average 57 percent (with a standard deviation of 0.01)
of the variation in entry in a market to the core efficiency of the firm rather
than to its draw of tj in that market.
31 By the law of large numbers, the fraction of French firms selling in n approximates the proba-
bility that a French firm will sell in n. Writing this probability as qn = NnF/NFF, the unconditional
variance of entry for a randomly chosen French firm is Vj? = qn{' - qn). Conditional on its stan-
dardized unit cost m, a firm enters market n if its entry shock r¡n satisfies т]п > (uK2/NnF)l/e. Since
In rjn is distributed N(0, cr^), the probability that this condition is satisfied is
where Ф is the standard normal cumulative density. The variance conditional on и is therefore
Vnc(u) = qn(u)[l-qn(u)].

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
4.7.2. Variation in Sales
Looking at the firms that enter a particular market, how much does variation
in и explain sales variation? Consider firm j selling in market n. Inserting (29)
into (26), the log of sales is
lnXnF(j) = lng„(fl +ln[l - ( У?* g) 1

2
- llnu(j) + ln((NnF/K2)ire<jEnF),
U
1 3
where we have divided s

all firms selling in marke
The first component in
while component 3 invo
Complicating matters is
entry shock in market
with this issue by first
variation in componen
together.
We simulate sales of 230,000 firms across our 113 markets and divide the
contribution of each component to its sales in each market where it sells. Look-
ing across markets, we find that component 3 itself contributes only 4.8 per-
cent of the variation in lnXnF(j), and components 2 and 3 together contribute
around 39 percent (with very small standard deviations).32
The dominant role of a is consistent with the shapes of the sales distribu-
tions in Figure 2. Even though the Pareto term и must eventually dominate
the log-normal a at the very upper tail, such large observations are unlikely in
a reasonably sized sample. With our parameter values, the log-normal term is
evident in both the curvature and the slope, even among the biggest sellers.33
Together these results indicate that the general efficiency of a firm is very im-
portant in explaining its entry into different markets, but makes a much smaller
contribution to the variation in the sales of firms actually selling in a market.34
32 In comparison, Munch and Nguyen (2009) found that firm effects drive around 40 percent
of the sales variation of Danish sellers across markets.
33 Our estimate of в implies that the sales distribution asymptotes to a slope with absolute value
of around 0.41, much lower than the values reported in footnote 7 among the top 1 percent of
firms selling in a market. Sornette (2000, pp. 80-82) discussed how a log normal with a large vari-
ance can mimic a Pareto distribution over a very wide range of the upper tail. Hence a continues
to play a role, even among the largest sellers in our samóle.
34 Our finding that и plays a larger role in entry than in sales conditional on entry is consistent
with our higher estimate of aa relative to arì. A lower value of 6 (implying more sales hetero-

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
4.8. Productivity
Our methodology so far has allowed us to estimate 0, which incorporates

both underlying heterogeneity in efficiency, as reflected in 0, and how this het-
erogeneity in efficiency gets translated into sales, through o' To separate 0 into
these components, we turn to data on firm productivity.
A common observation is that exporters are more productive (according to
various measures) than the average firm.35 The same is true of our French
exporters: The average value added per worker of exporters is 1.22 times the
average for all firms. Moreover, value added per worker, like sales in France,
tends to rise with the number of markets served, but not with nearly as much
regularity.
A reason for this relationship in our model is that a more efficient firm typ-
ically both enters more markets and sells more widely in any given market. As
its fixed costs are not proportionately higher, larger sales get translated into
higher value added relative to inputs used, including those used in fixed costs.
An offsetting factor is that more efficient firms enter tougher markets where
sales are lower relative to fixed costs. Determining the net effect of efficiency
on productivity requires a quantitative assessment.
We calculate productivity among our simulated firms as value added per unit
of factor cost, proceeding as follows.36 The value added Vi(j) of firm j from
country i is its gross production Yi(j) = Yln^niU), less spending on interme-
diates /,•(/):
Vi(j) = Yi(j)-Ii(j).
We calculate intermediate spending as
/•(/) = (1 - ß)m-1YiU) + EiUh
where ß is the share of factor costs in variable costs and Et(j) = ^2nEni(j).37
Value added per unit of factor cost is
m qM=wTwr
geneity attributable to
both in entry and in sal
35See, for example, Be
(2003).
36 In our model, value added per worker and value added per unit of factor cost are propor-
tional since all producers in a country face the same wage and input costs, with labor having the
same share.
37We treat all fixed costs as purchased services. See EKK (2008) for a more general treatment.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
The only potential source of heterogeneity is in the ratio E¿(j)/ Y/

The expression for firm productivity (40) depends onjhe elastici
stitution a (through m = a/ (a - 1)) independently of 0. We can
BEJK (2003) and find the a that makes the productivity advantage
in our simulated data match their productivity advantage in the
(1.22). To find the productivity of a firm in the simulated data, we
Xni(j) and fixed cost Eni(j), using (13), in each market where it se
sum to get Yi(j) and Ei(j).39 We then calculate the average q¡(j) for
relative to that for all firms. The a that delivers the same advantag
ulated data as the actual is a = 2.98, implying ß = 0.34. Using our
в = в/ (a - 1) = 2.46, the implied value of в is 4.87.40
4.9. The Price Index Effect
With values for the individual parameters в and a, we can retur

tion (14) and ask how much better off buyers are in a larger mar
38 An implication of (40) is that the distribution of productivity of sellers in suppl

market is invariant to any market-specific feature such as size, location, or trade
this result, replace £,</)/ Y/0') in (40) with Eni(j)/Xni(j). Using (13) for the nume
for the denominator, exploiting (30), we can write this ratio in terms of vni(j) as
A VniUY1-»'*-!
Eni(j) = o-(A-l) VniU)-x/i-l ' '
^niO") -lnVni(j) _
a0(vniU)-l/i-iy
Since vni(j) is distributed uniformly on [0, 1], in any market /
fixed costs to sales revenue depends only on À, a, and 0, and n
39We calibrate ß from data on the share of manufacturing
Denoting the value-added share as sv , averaging across data fr
Development Organization (UNIDO (1999)) gives us sv = 0.36.
fixed costs, we calculate
ß = msv -1/0,
so that ß is determined from sv simultaneously with our estim

40While the model here is different, footnote 11 suggests tha
role here to its role in Eaton and Kortum (2002) and in BEJK
falls between the estimates of 8.28 and 3.60, respectively, fr
a = 2.98 is not far below the estimate of 3.79 from BEJK (20
framework, Crozet and Koenig (2010) estimated 0 and a (along
the effect of distance D on trade costs) using French firm-l
contiguous to France. They employed a three-step procedure
for firm entry, retrieving 86 from the coefficient Ini), (ii) an
retrieving 8((r - 1) from the coefficient lnD, and (iii) a relat
sales, retrieving [в - (a - 1)]. Performing this procedure for 34
of в ranging from 1.34 to 4.43 (mean 2.59) and estimates of a
1.72).

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
Em's as given, the elasticity of Pn with respect to Xn is 1/0 - 1 /O - 1) = -0.30

Doubling market size leads to a 30 percent lower manufacturing price index.41
5. GENERAL EQUILIBRIUM AND COUNTERFACTUALS
We now use our framework to examine how counterfactual aggregate shocks

in policy and the environment affect individual firms. To do so, we need to con-
sider how such changes affect wages and prices. So far we have conditioned on
a given equilibrium outcome. We now have to ask how the world rééquilibrâtes.
5.1. Embedding the Model in a General Equilibrium Framework
Embedding our analysis in general equilibrium requires additional assump-

tions:
Al. Factors are as in Ricardo (1821). Each country is endowed with an
amount Li of labor (or a composite factor), which is freely mobile across activ-
ities within a country but does not migrate. Its wage in country i is Wi.
A2. Intermediates are as in Eaton and Kortum (2002). Consistent with
Section 4.8, manufacturing inputs are a Cobb-Douglas combination of labor
and intermediates, where intermediates have the price index Pt given in (14).
Hence an input bundle costs
w^WfP]-*,
where ß is the labor share.

A3. Nonmanufacturing is as in Alvarez and Lucas (2007). Final output,
which is nontraded, is a Cobb-Douglas combination of manufactures and non-
manufactures, with manufactures having a share y. Labor is the only input to
nonmanufactures. Hence the price of final output in country i is proportional
to P]Wl~'
A4. Fixed costs pay for labor in the destination. We thus decompose the
country-specific component of the entry cost Eni = WnFni, where Fni reflects
the labor required for entry for a seller from i in n.42
A5. Each country /'s manufacturing deficit Д and overall trade deficit Df
are held at their 1986 values (relative to world GDP).
41 Recall that our analysis in footnote 16 suggested that entry costs for French firms rise with
market size with an elasticity of 0.35, attenuating this effect. Assuming that this elasticity applies
to the entry costs Eni for all sources i, a calculation using (14) and (16) still leaves us with an
elasticity of (1 - 0.35) * (-0.30) = -0.20.
42 Combined with our treatment of fixed costs as intermediates in our analysis of firm-level
productivity, A4 implies that these workers are outsourced manufacturing labor.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
Equilibrium in the world market for manufactures requires th

across countries of absorption Xn of manufactures from each co
its gross output Y/, or
(41) Y, = £>„,*„.
tl='
We can turn these equilibrium conditions for manufacturing output into con-
ditions for labor market equilibrium that determine wages Wt in each country
as follows:
As shown in Appendix E, we can solve for Y¡ in terms of wages W¿ and ex-
ogenous terms as
(42) K } Y_ycr(WiLi + Df)-aDi

(42) K } ' l + (o--l)(/3-y/0) "
Since Xn = Yn + Dn,
, ,- Y _ J<r(WnLn + DAn) - (a - 1)(1 - ß + у/в)Рп

, { } Y n~ _ ' + {a-'){ß-y/e)
From expression (20), we can write
fAAs UWfpy-'drtyiaFrt)-^'-™-"
fAAs (44) тг„, = -
k=ì
¿гпж/^-Ч
Substituting (42), (43), and (44) into (41) gives us a set of equations that dete
mine wages Wi given prices P¿. From expression (14), we have
Yl/e
Ç TdWfP
(X ' (1/0)-i
*'w.)
giving us prices Pi given wages Wt. Together the two sets of equations deliver
Wi and Pi around the world in terms of exogenous variables.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
5.2. Calculating Counterfactual Outcomes
We apply the method used in Dekle, Eaton, and Kortum (2008; henceforth
DEK) to equations (41) and (45) to calculate counterfactuals.43 Denote the
counterfactual value of any variable x as x' and define jc = x'/x as its change.
Equilibrium in world manufactures in the counterfactual requires
(46) ¥! = ^<К-
We can write each of the components in terms of each country's baseline labor
income У 'f = W¡Li, baseline trade shares тг„„ baseline deficits, and the change
in wages W¡ and prices /*, using (42), (43), and (44) as
yaíYtWi + W-aDi
1 l + (<r-l)(/3-y/0) '
Y,^ya(YnLWn + Dt)-((r-l)(l-ß + y/e)Dn

1 + (о- -1)(/3-у/0)
_ w-0eíj-(l-/3)fl;j-ej7-[0-(o-l)]/(o-l)
, Tini" i ri "-ni rm
7Tni = ^
k='
where sticking these three equations into (46) yields a set of equations inv
ing Wi's for given Pi's. From (45), we can get a set of equations that involve
for given Wi's:
~T1/Ö /Y ' d/ö)-l/(o-l)
We implement counterfactual simulations for

gregating the rest of the world (ROW) into a
TTni with data on trade shares. We calibrate ß
tries) as described in Section 4.8, and take YtL
data on GDP, manufacturing production, and
sources of data and our procedures for assemb
terfactual.
43 DEK (2008) calculated counterfactual equilibria in a perfectly competitive 44-country world.

Here we adopt their procedure to accommodate the complications posed by monopolistic com-
petition and firm heterogeneity, expanding coverage to 114 countries (our 113 plus the rest of the
world).

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
A simple^ iterative algorithm solves jointly for changes in wa

giving us Wn, Pn, 7Tni, and Xn. From these values, we calculat
change in French exports in each market n, using the French
manufactures as the numéraire, as XnF = rrnFXn/PF and (ii) th
number of French firms selling there, using (27), as NnF = (7rn
We then calculate the implications of this change for individ
hold fixed all of the firm-specific shocks that underlie firm h
as to isolate the microeconomic changes brought about by gen
forces. We produce a data set, recording both baseline and coun
level behavior, as follows: ^ ^
• We apply the changes XnF and NnF derived above to our or
to get counterfactual values of total French sales in each mark
number of French sellers there N%F.
• We run the simulation algorithm described in Section 4.2 w
and the same stochastic draws applying both to the baseline a
terfactual. We set 0 to the parameter estimates reported in
accommodate counterfactuals, we tweak our algorithm in four
(a) In Step 2, we use our baseline values XnF and NnF to calcu
(тЕп/s, and use our counterfactual values X%F and N%F to calc
factual o-E^/s in each destination.
(b) In Step 4, we use our baseline values XnF and NnF to calc
Hn(s)% and use our counterfactual values X„F and N%F to calcu
factual u„(sYs for each destination and firm, using (28) and (3
(c) In Step 5, we set
u(s) = max{ïïn(s), ucn(s)}.

n
A firm for which u(s) < ~un(s) sells in market n in the baselin
for which u(s) < u^(s) sells there in the counterfactual. Hence
allows for entry, exit, and survival.
(d) In Step 8, we calculate entry and sales in each of the 113
baseline and in the counterfactual.
5.3. Counterfactual: Implications of Globalization
We consider a 10 percent drop in trade barriers, that is, dni = 1/(1.1) for
/ Ф n and dnn = Fni = 1. This change roughly replicates the increase in French
import share over the decade following 1986.44 Table III shows the aggregate
44Using time-series data from the Organization for Economic Cooperation and Development's
(OECD's) STAN data base, we calculated the ratio of manufacturing imports to manufacturing
absorption (gross production + imports - exports) for the 16 OECD countries with uninter-
rupted annual data from 1986 through 2000. By 1997, this share had risen for all 16 countries,

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
TABLE III
Aggregate Outcomes of Counterfactual Experiment
Counterfactual Changes
(ratio of counterfactual to baseline)
_ . „ . . French Firms
_ Real . „ Relative . .
Country Code Wage Wage Sales Number
Afghanistan AFG 1.01 0.92 1.22 1.23

Albania ALB 1.01 0.94 1.35 1.34
Algeria ALG 1.00 0.90 1.09 1.12
Angola ANG 1.00 0.90 1.08 1.10
Argentina ARG 1.01 0.96 1.57 1.52
Australia AUL 1.02 0.96 1.35 1.29
Austria AUT 1.04 1.04 1.49 1.32
Bangladesh BAN 1.01 0.95 1.37 1.33
Belgium3 BEL 1.09 1.11 1.44 1.19
Benin BEN 1.02 0.94 1.12 1.09
Bolivia BOL 1.02 0.94 1.21 1.18
Brazil BRA 1.01 0.96 1.64 1.57
Bulgaria BUL 1.02 0.95 1.38 1.34
Burkina Faso BUK 1.01 0.93 1.17 1.17
Burundi BUR 1.01 0.92 1.21 1.21
Cameroon CAM 1.01 0.92 1.19 1.19
Canada CAN 1.04 1.05 1.43 1.26
Central African Republic CEN 1.02 1.03 1.33 1.20
Chad CHA 1.01 0.90 1.07 1.10
Chile CHI 1.03 1.02 1.53 1.38
China CHN 1.01 0.94 1.38 1.36
Colombia COL 1.01 0.92 1.22 1.23
Costa Rica COS 1.02 0.94 1.22 1.20
Cote d'Ivoire COT 1.03 0.98 1.36 1.28
Cuba CUB 1.01 0.93 1.26 1.24
Czechoslovakia CZE 1.03 1.01 1.52 1.38
Denmark DEN 1.04 1.06 1.46 1.27
Dominican Republic DOM 1.04 0.99 1.36 1.28
Ecuador ECU 1.02 0.96 1.33 1.28
Egypt EGY 1.02 0.92 1.12 1.12
El Salvador ELS 1.02 0.93 1.10 1.10
Ethiopia ETH 1.01 0.92 1.08 1.09
Finland FIN 1.03 1.02 1.53 1.38
France FRA 1.02 1.00 0.95 0.88
Germany, East GEE 1.01 0.96 1.58 1.52
Germany, West GER 1.03 1.02 1.60 1.45
Ghana GHA 1.02 0.99 1.38 1.28
(Continues)
with a minimum increase of 2.4 for Norway and a maximum of 21.1 percentage points for Bel-
gium. France, with a 10.0, and Greece, with an 11.0 percentage point increase, straddled the
median.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
TABLE III- Continued
„ . ™ , • French Firms
„ Real . ™ Relative , •
Greece GRE 1.02 0.97 1.37 1.30

Guatemala GUA 1.01 0.92 1.16 1.17
Honduras HON 1.02 0.95 1.19 1.16
Hong Kong HOK 1.14 1.20 1.33 1.02
Hungary HUN 1.05 1.04 1.41 1.25
India IND 1.01 0.95 1.40 1.37
Honduras HON 1.02 0.95 1.19 1.16
Hong Kong HOK 1.14 1.20 1.33 1.02
Hungary HUN 1.05 1.04 1.41 1.25
India IND 1.01 0.95 1.40 1.37
Indonesia INO 1.02 0.96 1.44 1.38
Iran IRN 1.01 0.93 1.16 1.16
Iraq IRQ 1.04 0.94 1.08 1.06
Ireland IRE 1.07 1.09 1.43 1.21
Israel ISR 1.04 1.01 1.44 1.31
Italy ITA 1.02 0.99 1.57 1.46
Jamaica JAM 1.05 1.02 1.35 1.22
Japan JAP 1.01 0.98 1.80 1.69
Jordan JOR 1.03 0.95 1.16 1.13
Kenya KEN 1.01 0.93 1.18 1.18
Korea, South KOR 1.04 1.04 1.58 1.40
Kuwait KUW 1.02 0.94 1.14 1.12
Liberia LIB 1.49 1.03 1.27 1.14
Libya LI Y 1.02 0.95 1.10 1.07
Madagascar MAD 1.01 0.94 1.22 1.20
Malawi MAW 1.01 0.92 1.17 1.17
Malaysia MAY 1.07 1.08 1.45 1.23
Mali MAL 1.02 0.95 1.15 1.12
Mauritania MAU 1.08 1.19 1.36 1.05
Mauritius MAS 1.07 1.05 1.47 1.29
Mexico МЕХ 1.01 0.94 1.32 1.30
Morocco MOR 1.02 0.98 1.37 1.30
Mozambique MOZ 1.01 0.94 1.29 1.26
Nepal NEP 1.01 1.01 1.35 1.24
Netherlands NET 1.06 1.14 1.41 1.14
New Zealand NZE 1.03 1.00 1.45 1.33
Nicaragua NIC 1.01 0.89 1.06 1.09
Niger NIG 1.02 1.09 1.47 1.25
Nigeria NIA 1.00 0.89 1.07 1.12
Norway NOR 1.04 1.03 1.38 1.23
Oman ОМА 1.04 0.99 1.11 1.03
Pakistan РАК 1.02 0.97 1.41 1.34
Panama PAN 1.09 0.96 1.15 1.10
(Continues)

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
TABLE III- Continued
„ , тл , • French Firms
„ Real , тл Relative , •
Papua New Guinea PAP 1.07 1.09 1.33 1.13

Paraguay PAR 1.01 0.93 1.21 1.20
Peru PER 1.02 0.97 1.39 1.32
Philippines PHI 1.02 0.97 1.50 1.43
Portugal POR 1.03 1.03 1.47 1.32
Romania ROM 1.01 0.97 1.68 1.61
Rwanda RWA 1.00 0.90 1.14 1.17
Saudi Arabia SAU 1.02 0.95 1.15 1.11
Senegal SEN 1.03 1.01 1.36 1.24
Sierra Leone SIE 1.03 1.17 1.36 1.08
Singapore SIN 1.24 1.15 1.37 1.10
Somalia SOM 1.03 0.96 1.09 1.05
South Africa SOU 1.03 1.01 1.56 1.43
Spain SPA 1.02 0.97 1.49 1.42
Sri Lanka SRI 1.03 0.99 1.34 1.24
Sudan SUD 1.00 0.91 1.13 1.15
Sweden SWE 1.04 1.05 1.51 1.33
Switzerland SWI 1.05 1.05 1.48 1.31
Syria SYR 1.02 0.96 1.20 1.15
Taiwan TAI 1.04 1.05 1.64 1.44
Tanzania TAN 1.01 0.94 1.15 1.13
Thailand THA 1.03 0.99 1.50 1.40
Togo TOG 1.03 0.96 1.11 1.07
Trinidad and Tobogo TRI 1.04 1.01 1.22 1.12
Tunisia TUN 1.04 1.00 1.36 1.26
Turkey TUR 1.01 0.95 1.37 1.33
Uganda UGA 1.00 0.90 1.06 1.08
United Kingdom UNK 1.03 1.00 1.46 1.35
United States USA 1.01 0.96 1.45 1.40
Uruguay URU 1.02 1.00 1.65 1.53
USSR USR 1.00 0.92 1.32 1.33
Venezuela VEN 1.01 0.91 1.18 1.20
Vietnam VIE 1.01 0.95 1.37 1.33
Yugoslavia YUG 1.02 0.97 1.48 1.41
Zaire ZAI 1.06 1.21 1.37 1.04
Zambia ZAM 1.03 1.12 1.49 1.22
Zimbabwe ZIM 1.02 0.97 1.43 1.36
a Belgium includes Luxembourg.
general equilibrium conseque

wage wn = (Wn/Pn)y% (ii) the
change in the sales of French

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
TABLE IV
COUNTERFACTUALS: FIRM TOTALS3
Counterfactual
Change From Percentage

Baseline Baseline Change
Number
All firms 231,402 -26,589 -11.5
Exporting 32,969 10,716 32.5
Values ($ millions)
Total sales 436,144 16,442 3.8
Domestic sales 362,386 -18,093 -5.0
Exports 73,758 34,534 46.8
a Counterfactual simulation of a 10% decline in trade costs.
in the number of French firms selling to each market NnF. Lower trade bar-
riers raise the real wage in every country, typically by less than 5 percent.45
Relative wages move quite a bit more, capturing terms of trade effects from
globalization.46 The results that matter at the firm level are French sales and
the number of French firms active in each market. While French sales decline
by 5 percent in the home market, exports increase substantially, with a maxi-
mum 80 percent increase in Japan.47 The number of French exporters increases
roughly in parallel with French exports.48
Table IV summarizes the results, which are dramatic. Total sales by French
firms rise by $16.4 million, the net effect of a $34.5 million increase in exports
and a $18.1 million decline in domestic sales. Despite this rise in total sales,
45There are a several outliers on the upper end, with Belgium experiencing a 9 percent gain,
Singapore a 24 percent gain, and Liberia a 49 percent gain. These results are associated with
anomalies in the trade data due to entrepôt trade or (for Liberia) ships. These anomalies have
little consequence for our overall results.
46 A good predictor of the change in country «'s relative wage is its baseline share of exports
in manufacturing production, as the terms of trade favor small open economies as trade barriers
decline. A regression in logs across the 112 export destinations yields an R2 of 0.72.
47 A good predictor of the change in French exports to n comes from log-linearizing тгпР, noting
that 'nXnF = 1п7гп/г + 'nXn. The variable that capturesjthe change in French cost advantage
relative to domestic producers in n is x' = 7rnn[ß'n(Wn/WF) - 'ndnF] with a predicted elasticity
of 0 (we ignore changes in the relative value of the manufacturing price index, sincejhey are
small). The variable that captures the percentage change in «'s absorption is X2 = 'nWn with a
predicted elasticity of 1. A regression across the 112 export destinations of 'nXnF on x' and x2
yields an R2 of 0.88 with a coefficient on xi of 5.66 and on x2 of 1.30.
48 We can compare these counterfactual outcomes with the actual change in the number of
French sellers in each market between 1986 and 1992. We tend to overstate the increase and to
understate heterogeneity across markets. The correlation between the counterfactual change and
the change in the data is nevertheless 0.40. Our counterfactual, treating all parameters as given
except for the iceberg trade costs, obviously misses much of what happened in those 6 years.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
TABLE V
COUNTERFACTUALS: FIRM ENTRY AND EXIT BY INITIAL SIZE
All Firms Exporters

Counterfactual Counterfactual
Initial Size Interval Baseline No. Change From Change Baseline No. Change From Change
(percentile) of Firms Baseline (%) of Firms Baseline (%)
Not active 0 1118 - 0 1118 -

Oto 10 23,140 -11,551 -49.9 767 15 2.0
10 to 20 23,140 -5702 -24.6 141 78 55.1
20 to 30 23,140 -3759 -16.2 181 192 106.1
30 to 40 23,140 -2486 -10.7 357 357 100.0
40 to 50 23,140 -1704 -7.4 742 614 82.8
50 to 60 23,138 -1141 -4.9 1392 904 65.0
60 to 70 23,142 -726 -3.1 2450 1343 54.8
70 to 80 23,140 -405 -1.8 4286 1829 42.7
80 to 90 23,140 -195 -0.8 7677 2290 29.8
90 to 99 20,826 -38 -0.2 12,807 1915 15.0
99 to 100 2314 0 0.0 2169 62 2.8
Total 231,402 -26,589 32,969 10,716
competition from imports drives almost 27

almost 11,000 firms start exporting.
Tables V and VI decompose these changes
of different baseline size, with Table V con
half the firms in the bottom decile are wip
avoids any attrition. Because so many firm
the greatest number of new exporters eme
The biggest percentage increase in number of
from the bottom decile.
Table VI decomposes sales revenues. All of the increase is in the top decile,
and most of that in the top percentile. For every other decile, sales decline.
Almost two-thirds of the increase in export revenue is from the top percentile,
although lower deciles experience much higher percentage increases in their
export revenues.
Comparing the numbers in Tables V and VI reveals that, even among sur-
vivors, revenue per firm falls in every decile except the top. In summary, the
decline in trade barriers improves the performance of the very top firms at the
expense of the rest.49
In results not shown, we decompose the findings according to the number of
markets where firms initially sold. Most of the increase in export revenues is
49The first row of the tables pertains to firms that entered only to export. There are only 1108
of them selling a total of $4 million.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
TABLE VI
COUNTERFACTUALS: FIRM GROWTH BY INITIAL SIZE
Total Sales Exports

Counterfactual Counterfactual
Initial Size Interval Baseline Change From Change Baseline Change From Change
(percentile) ($ millions) Baseline (%) ($ millions) Baseline (%)
Not active 0 3 - 0 3 -
Oto 10 41 -24 -58.0 1 2 345.4
10 to 20 190 -91 -47.7 1 2 260.3
20 to 30 469 -183 -39.0 1 3 266.7
30 to 40 953 -308 -32.3 2 7 391.9
40 to 50 1793 -476 -26.6 6 18 307.8
50 to 60 3299 -712 -21.6 18 48 269.7
60 to 70 6188 -1043 -16.9 58 130 223.0
70 to 80 12,548 -1506 -12.0 206 391 189.5
80 to 90 31,268 -1951 -6.2 1085 1501 138.4
90 to 99 148,676 4029 2.7 16,080 11,943 74.3
99 to 100 230,718 18,703 8.1 56,301 20,486 36.4
Total 436,144 16,442 73,758 34,534
among the firms that were already expor

increase falls with the initial number of m
export to few markets, a substantial shar
ing new markets.
Finally, we can also look at what happen
of the increase in exports to each market
We can also look at growth in sales by in
would predict, sales by firms with an ini
tially more than those at the top.50
6. conclusion
A literature that documents the superior performance of exporter

Bernard and Jensen (1995), inspired a new generation of trade theor
corporate producer heterogeneity. These theories, in turn, deliver
beyond the facts that motivated them. To account for the data prese
50We can examine another counterfactual analytically: a uniform change F in labo

for entry. Proportional changes in F cancel out in the expression for тг'п1, so that th
is in equation (47) for the manufacturing price index:
With F = 1/1.1, the number of French firms selling in every market rises by 10 per
sales increase by the factor 1/0.919 = 1.088 in each market.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
which break down firms' foreign sales into individual export destinations, these
theories are confronted with a formidable challenge.
We find that the Melitz (2003) model, with parsimonious modification, suc-
ceeds in explaining the basic features of such data along a large number of
dimensions: entry into markets, sales distributions conditional on entry, and
the link between entry and sales in individual foreign destinations and sales at
home. Not only does the framework explain the facts, it provides a link between
firm-level and aggregate observations that allows for a general-equilibrium ex-
amination of the effect of aggregate shocks on individual firms.
Our framework does not, however, constitute a reductionist explanation of
what firms do in different markets around the world. In particular, it leaves
the vastly different performance of the same firm in different markets as a
residual. Our analysis points to the need for further research into accounting
for this variation.
REFERENCES
Alvarez, F, and R. E. Lucas (2007): "A General Equilibrium Analysis of the Eaton
Model of International Trade," Journal of Monetary Economics, 54, 1726-1768. [I486!
Arkolakis, K. (2010): "Market Penetration Costs and the New Consumers Margin in
tional Trade," Journal of Political Economy, 118, 1151-1199. [1454,1463,1479,14951
Aw, B. Y., S. Chung, and M. J. Roberts (2000): "Productivity and Turnover in t
Market: Micro-Level Evidence From the Republic of Korea and Taiwan (China)," W
Economic Review, 14, 65-90. [14531
Bernard, А. В., and J. B. Jensen (1995): "Exporters, Jobs, and Wages in U.S. Manu
1976-1987 '," Brookings Papers on Economic Activity: Microeconomics, 1995, 67-119. [1
14951
Bernard, А. В., J. Eaton, J. B. Jensen, and S. Kortum (2003): "Plants and Productivity in
International Trade," American Economic Review, 93, 1268-1290. [1453,1462,1484,1485]
BlSCOURP, P., AND F. Kramarz (2007): "Employment, Skill Structure, and International Trade:
Firm-Level Evidence for France ," Journal of International Economics, 72, 22-51. [14551
CHANEY, T. (2008): "Distorted Gravity: Heterogeneous Firms, Market Structure, and the Geog-
raphy of International Trade ," American Economic Review, 98, 1707-1721. [1454,1462,1485]
Clerides, S., S. Lach, and J. R. Tybout (1998): "Is Learning by Exporting Important? Micro-
Dynamic Evidence From Colombia, Mexico, and Morocco," Quarterly Journal of Economics,
113, 903-947. [14841
Crozet, M., and P. KOENIG (2010): "Structural Gravity Equations With Intensive and Extensive
Margins," Canadian Journal of Economics, 43, 41-62. [1485]
Dekle, R., J. Eaton, and S. Kortum (2008): "Global Rebalancing With Gravity," IMF Staff
Papers, 55, 511-540. [1488]
DIXIT, A., AND J. E. STIGLITZ (1977): "Monopolistic Competition and Optimum Product Diver-
sity, "American Economic Review, 67, 297-308. [1464]
Eaton, J., and S. Kortum (1999): "International Technology Diffusion: Theory and Measure-
ment," International Economic Review, 40, 537-570. [1462]
1485,1486]
Eaton, J., S. Kortum, and F. Kramarz (2004): "Dissecting Trade: Firms, Industries, and Ex-
port Destinations," American Economic Review, Papers and Proceedings, 94, 150-154. (Data
sources and tables available in NBER Working Paper 10344.) [1455]

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
Paper 14610, NBER. [14841
French Firms'," Econométrica Supplemental Materia

ecta/Supmat/83 18_miscellaneous.pdf; http://www.
data and programs-l.zip; http://www.econometri
programs-2.zip. [14551
Gabaix, X. (1999): "Zipf's Law for Cities: An Explana
739-767. [14621
Gabaix, X., and R. Ibragimov (2011): "Rank-1/2: A
tion of Tail Exponents ," Journal of Business & Econ
Goffe, W L. (1996): "SIMANN: A Global Optimiza
ing," Studies in Nonlinear Dynamics and Economet
www.bepress.com/snde/voll/iss3/algorithml. [1478]
Goffe, W. L., G. D. Ferrier, and J. Rogers (1994):
tions With Simulated Annealing," Journal of Econom
GOURIÉROUX, С, AND A. MONFORT (1994): Simu
Lecture Series. Louvain-la-Neuve: Université catho
Helpman, E., M. J. Melitz, and Y. Rubinstein (2
Partners and Trading Volumes," Quarterly Journal o
Helpman, E., M. J. Melitz, and S. R. Yeaple (2004
Firms," American Economic Review, 94, 300-316. [14
Horowitz, J. (2001): "The Bootstrap," in Handbook
man and E. E. Learner. Amsterdam: Elsevier Scien
Kortum, S. (1997): "Research, Patenting, and T
1389-1419. [14621
Krugman, P. R. (1980): "Scale Economies, Product Differentiation, and the Patterns of Trade,"
American Economic Review, 70, 950-959. [1466]
Luttmer, E. (2007): "Selection, Growth, and the Size Distribution of Firms," Quarterly Journal
of Economics, 122, 1103-1144. [1462]
Melitz, M. J. (2003): "The Impact of Trade on Intra-Industry Reallocations and Aggregate In-
dustry Productivity," Econometrica, 71, 1695-1725. [1453,1454,1462,1463,14961
Munch, J. R., and D. X. Nguyen (2009): "Decomposing Firm-Level Sales Variation," Working
Paoer 2009-05. EPRU. П4831
OECD (1995): The OECD STAN Database. [1489]
PEDERSEN, N. K. (2009): "Essays in International Trade," Ph.D. Thesis, Northwestern University,
Evanston, IL. [1470]
RICARDO, D. (1821): On The Principles of Political Economy and Taxation. London: Murray.
[14861
Roberts, M., and J. R. Tybout (1997): "The Decision to Export in Colombia: An Empirical
Model of Entry With Sunk Costs," American Economic Review, 87, 545-564. [14541
SIMON, H. A. (1955): "On a Class of Skew Distribution Functions," Biometrika, 42, 425-440.
[14621
SIMON, H. A., AND C. P. BONINI (1958): "The Size Distribution of Business Firms," American
Economic Review, 98, 607-617. [14581
SORNETTE, D. (2000): Critical Phenomena in Natural Sciences. Berlin: Springer. [14831
UNIDO (1999): Industrial Statistics Database. [1485]
Dept. of Economics, Pennsylvania State University, 608 Kern Gr

ing, University Park, PA 16802, USA.; [email protected],
Dept. of Economics, University of Chicago, 1126 East 59th Stre
IL 60637, U.S.A.; [email protected],

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC
and
CREST (ENSAE), 15 Boulevard Gabriel Péri, 92245, Malakoff, France;
[email protected].
Manuscript received December, 2008; final revision received December, 2010.

168.176.5.118 on Fri, 31 Jul 2020 17:50:09 UTC

Eaton 2011

Uploaded by

Copyright:

Available Formats

Eaton 2011

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Eaton 2011

Uploaded by

Copyright:

Available Formats

AN ANATOMY OF INTERNATIONAL TRADE: EVIDENCE FROM FRENCH FIRMS

Author(s): Jonathan Eaton, Samuel Kortum and Francis Kramarz

Stable URL: http://www.jstor.com/stable/41237782

This content downloaded from

AN ANATOMY OF INTERNATIONAL TRADE: EVIDENCE FROM

By Jonathan Eaton, Samuel Kortum, and Francis Kramarz1

Keywords: Export destinations, firms, efficiency, sales distribution, simulated m

We exploit detailed data on the exports of French firms to confront a

© 2011 The Econometric Society DOI: 10.3982/ECTA8318

This content downloaded from

export decision of individual firms in partial equilibrium.3 However, the task of

3A pioneering paper here is Roberts and Tybout (1997).

This content downloaded from

2.1. Market Entry

Figure 1 A plots the number of French manufacturing firms NnF selling to a

where XnF is total exports of our French firms to market n.

This content downloaded from

Figure 1. - Entry and sales by market size.

This content downloaded from

French Firms Exporting to the Seven Most Popular Destinations

Export Destination Exporters Exporters

Belgium3 (BE) 17,699 0.520

a Belgium includes Luxembourg.

This content downloaded from

French Firms Selling to Strings of Top-Seven Countries

Number of French Exporters

Export Stringa Data Under Independence Model

BEa 3988 1700 4417

aThe string BE means selling

the number of firms adh

2.2. Sales Distributions

Our second exercise expands on Figure 1С by looking at the entire distri-

Pr[*„ <*«] = <?,

This content downloaded from

Figure 2. - Sales distributions of French firm: Graphs by country.

2.3. Export Participation and Size in France

How does a firm's participation in export markets relate to i

This content downloaded from

Figure 3. - Sales in France and market entry.

2.4. Export Intensity

This content downloaded from

Figure 4. - Distribution of export intensity, by market.

firm /'s normalized export intensity in market /г, which we define as

In seeking to interpret these relationships, we turn to a parsimonious model

This content downloaded from

3.1. Producer Heterogeneity

A potential producer of good j in country i has efficiency z¡(j). A bundle

The measure of potential producers in country i who can prod

(2) ^{z) = TiZ-' z>0,

(3) Мл,(с) = м-(^)=Ф«,св,

We now turn to demand and market structure in a typical destination.

This content downloaded from

3.2. Demand, Market Structure, and Entry

A market n contains a measure of potential buyers. To sell to a f

(4) Eni(j) = sn(j)EniM(f).

(5) Xn{j) = an{j)fXn(^'

This content downloaded from

(6) nni(p, /) = (l - ^)a»Wf(jr) X» - en(j)EniM(f).

(7) PnU) = mcn{j),

(8, /.,0) = maxj1.[4,0)^.(^>)-<'-)]-*,0¡

Pr[„ <«] = <?,