Exporter Behavior, Country Size and Stage of Development*
Ana M. Fernandes
Caroline Freund
Martha Denisse Pierola
and Tolga Cebeci
November 2013
Abstract
We present new data on exporter characteristics for 45 countries and study how they vary with
country size and stage of development. The evidence implies that larger and more developed
countries have more exporters, which are on average larger and more diversified, and an
increasing share of exports is controlled by the top 5 percent of exporters. The channel through
which the number of exporters expands matures as countries develop—exporter turnover plays a
significantly more important role in countries at an early stage of development, while higher
survival rates of new exporters promote creative destruction in more developed countries, with
net entry being invariant to income. Overall, the evidence points toward allocative efficiency in
export markets as countries grow. We discuss the findings considering recent trade theories with
heterogeneous firms.
JEL Classification codes: C81, F14.
Keywords: exporter dynamics, exporter growth, firm-level data.
*
Ana Fernandes, Martha Denisse Pierola, and Tolga Cebeci are at the World Bank; Caroline Freund is at the Peterson Institute
for International Economics and CEPR. We are grateful for the support of the governments of Norway, Sweden and the United
Kingdom through the Multi-Donor Trust Fund for Trade and Development. We also acknowledge the generous financial support
from the World Bank research support budget and the Knowledge for Change Program (KCP), a trust funded partnership in
support of research and data collection on poverty reduction and sustainable development housed in the office of the Chief
Economist of the World Bank (www.worldbank.org/kcp). We thank Mario Gutierrez-Rocha and Jamal Haidar for their
contribution to this project and Erhan Artuc, Richard Baldwin, Peter Klenow, Aaditya Mattoo, Sam Kortum, and Andres
Rodriguez-Clare for many useful comments and discussions. We thank Alberto Behar and Madina Kukenova for collaboration
with the project early on and Francis Aidoo, Paul Brenton, Maurizio Bussolo, Huot Chea, Julian Clarke, Alain D’Hoore, Pablo
Fajnzylber, Elisa Gamberoni, Sidiki Guindo, Javier Illescas, Mehar Khan, Sanjay Kathuria, Josaphat Kweka, Erjon Luci, Mariem
Malouche, Eric Manes, Yira Mascaro, Mamadou Ndione, Zeinab Partow, Stefano Paternostro, Jorge Rachid, Richard Record,
Jose Guilherme Reis, Jose Daniel Reyes, Nadeem Rizwan, Liviu Stirbat, Dimitri Stoelinga, Daria Taglioni, Maria Nicolas, Erik
von Uexkull, Cristian Ugarte, and Ravindra Yatawara for facilitating the access to the data for developing countries. We thank
Lina Ahlin, Martin Andersson, Juan de Lucio, Emmanuel Dhyne, Luc Dresse, Cédric Duprez, Richard Fabling, Jaan Masso, Raul
Minguez, Asier Minondo, Andreas Moxnes, Francisco Requena, Lynda Sanderson, Joana Silva, Priit Vahter, and Hylke
Vandenbussche for providing us with measures for developed countries. We thank workshop participants at the Empirical
Investigations in International Trade at Santa Cruz, the Graduate Institute in Geneva, the World Bank, the World Trade
Organization, and the International Trade Center for useful suggestions. The findings expressed in this paper are those of the
authors and do not necessarily represent the views of the World Bank.
I. Introduction
We present new evidence on how the characteristics and dynamics of a country’s exporters vary
with country size and stage of development. A number of studies use data from individual
countries or a small group of countries in a region to demonstrate how firms export.1 Some
commonalities arise: exports are found to be concentrated in a small share of large firms that
export many products to many markets and high rates of entry and exit into exporting are
typically found. However, because the studies involve single countries or groups of similar
countries, it is not possible to determine how exporter behavior evolves with country
characteristics. In addition, while many of the conclusions appear to be comparable across
countries, it would require a larger sample to generalize the existing results across countries that
differ in income and population. This paper fills this gap in the literature by presenting a large
cross-country database of exporter behavior and exploring how exporter characteristics and
dynamics vary with country size and stage of development, and also what features are common
to all economies.
Specifically, we gather exporter-level customs information from 38 developing and 7
developed countries around the world to build the “Exporter Dynamics Database” (henceforth
referred to as the Database), containing comparable measures of exporter, product and market
dynamics. The Database contains measures at different levels of aggregation: a) country-year, b)
country-year-product (HS 2-digit, HS 4-digit, or HS 6-digit) and c) country-year-destination
mostly for the period between 2003 and 2010 (with longer time series for some countries). The
Database covers different aspects of firm dynamics, firm-product and firm-destination dynamics,
1
See Eaton, Kortum, and Kramarz (2008) for France; Eaton et. al. (2008) for Colombia; Amador and Opromolla (2013) for
Portugal; Iacovone and Javorcik (2010) for Mexico; Andersson, Lööf, and Johansson (2008) for Sweden, Albornoz et. al (2010)
for Argentina; Freund and Pierola (2010) for Peru; Manova and Zhang (2012) for China, Masso and Vahter (2011) for Estonia,
De Lucio et. al. (2011) for Spain, Ekholm, Moxnes, and Ulltveit-Moe (2012) for Norway, Fabling and Sanderson (2012) for New
Zealand, Mayer and Ottaviano (2008) for several EU countries, among others. See also Bernard et. al. (2007, 2011a) for US and a
review of the empirical literature, respectively.
2
as well as exporter growth patterns, concentration, and diversification in the non-oil exporting
sector.
This paper introduces the Database and establishes several novel stylized facts on how
exporter characteristics and dynamics vary with country characteristics. The distinguishing
feature of these stylized facts is that they could not have been uncovered using any other crosscountry source of trade data available to date. We also confirm or generalize facts found in the
recent empirical trade literature.
First, we examine the relationships between exporter characteristics and both country size
and stage of development. Controlling for the sectoral distribution of exports, we find that larger
and more developed economies have a greater number of exporters, larger average exporter size,
more diversified exporters, and that total exports are more concentrated in the top 5 percent of
exporting firms. Similar results are found when we control for export destination. These results
imply that as countries grow and exports expand, the export expansion happens through both the
intensive and extensive margins, and more resources flow to the larger and more diversified
firms.
Second, we study how exporter dynamics change with stage of development and country
size.
We find that rates of both entry and exit into exporting are decreasing in stage of
development, after controlling for sector fixed effects or destination effects, yielding similar rates
of net entry across rich and poor countries. There is also evidence that survival rates of entrants
into export markets increase as countries develop. Exporter dynamics are, however, not robustly
correlated with country size. This implies that exporter dynamics mature as countries grow, first
fueled by trial and error and later by more resilient exporters.
3
We also confirm for a large number of developing countries the well-known fact from the
literature focusing on individual countries that total exports are largely dominated by multiproduct multi-destination exporters, but these account for a small share of the number of
exporters. In particular, exporters selling more than four products to more than four markets
account for 60 percent of exports on average. In line with our earlier results on concentration
and diversification, we find that within broad sectors, these multi-product multi-destination
exporters make up a larger share of exporters and account for a larger share of total exports as
countries develop—again suggesting that a larger share of resources are directed to the most
diversified (and productive) firms as countries grow.
Overall the results imply that larger and more developed economies have more
competitive export sectors, with a greater share of larger and more diversified firms, as well as
more efficient creative destruction. While the standard heterogeneous firm model of trade does
not offer definitive predictions about country size or stage of development and exporter
characteristics, we discuss some adaptations that are consistent with these facts. In particular,
allowing variation among countries in the shape of the firm productivity distribution, such as
Spearot (2013), or allowing a wider support for the firm productivity distribution in bigger and
richer countries, in a model with a truncated distribution, such as Helpman Melitz and
Rubenstein (2008), can help to explain why average size and concentration vary with stage of
development. Trade models which highlight the greater productivity of multi-product and multidestination firms, such as Bernard, Redding and Schott (2011b), are consistent with our findings
on firm level diversification and stage of development. Trade models incorporating uncertainty
about a firm’s profits in foreign markets offer predictions similar to our results on entry and exit.
More broadly, the evidence supports allocative efficiency as countries grow in the sense that
4
resources are increasingly controlled by larger and more diversified exporters and dynamics
involve lower gross flows for the same amount of net entry.
The rest of the paper is structured as follows. Section II describes briefly the Database.
Section III presents three new stylized facts based on the Database. Section IV discusses models
that provide a theoretical framework to understand the facts presented. Finally, Section V
concludes.
II. The Exporter Dynamics Database
The Exporter Dynamics Database is constructed using customs data, covering the
universe of annual exporter transactions for 37 developing countries and 8 developed countries.2
The Database includes the basic characteristics of exporters in each country (e.g., number of
exporters, average exporter size and exporter growth rates), concentration/diversification (e.g.,
Herfindahl index, share of top exporters, number of products and destinations per exporter), firm,
product and market dynamics (e.g., entry, exit and one-year survival rates of entrants) and unit
prices (e.g., per exporter of a given product). The measures are available at different levels of
aggregation: a) country-year (98 measures), b) country-year-product (HS2, HS4, or HS6) (113
measures, 113 measures, and 89 measures, respectively) and c) country-year-destination (74
measures). The list of measures under each category and for each type of aggregation level is
presented in Appendix 1 along with the corresponding formula. The Database is publicly
available at http://data.worldbank.org/data-catalog/exporter-dynamics-database. Details on how
2
The group of countries is determined by the willingness to share data by local officials or other researchers who have access to
customs data. While the data underrepresent industrial countries and Asia, the patterns we find are robust and available statistics
from other studies, such as for the US and the EU, are consistent with the patterns observed here. We aim over time to expand
coverage of the Exporter Dynamics Database.
5
the countries’ customs data were cleaned and harmonized, and how the Database was
constructed are provided in Appendix 2.
Table 1 presents a summary of a representative set of measures in the Database, including
number of exporters, average and median exporter size, share of top 5 percent of exporting firms,
number of HS 6-digit products exported per firm, number of destinations served per firm, entry,
exit and entrant survival rates as well as total exports per country. We report averages per
country for the period 2006-2008, which is the period most commonly covered across countries
and captures trade performance before the global financial crisis that started at the end of 2008.3
Table 2 presents the same set of measures in the Database but for groups of HS 2-digit sectors,
where the measures for each sector group are obtained as averages across all countries that
export that particular group of HS 2-digit sectors, again focusing on the period 2006-2008.
Some interesting cross-country patterns emerge from Table 1. First, there is tremendous
variation across countries in the export base (number of exporters) and the average exports per
firm. Among developing countries, the largest numbers of exporters are found in Turkey and
Mexico followed by South Africa and Brazil (around 20,000) whereas the smallest pools of
exporters are found in Niger and Mali (around and below 300), followed by Burkina Faso, Laos,
Yemen, and Cambodia. This pattern seems to mirror the countries’ level of development, and
this link will be studied further in Section III. The number of exporters per capita is also largest
in the richer countries in the sample.
Average exports per firm are the highest for Belgium, followed by Brazil, Chile, and
Mexico (in the range of 7-8 million USD). Albania, Lebanon, Macedonia, Yemen, and Kenya
3
Note that the sharp drop in world trade ensuring from the global financial crisis occurs in 2009 not in 2008. For countries for
which data is available for only 1 or 2 years within the period 2006-2008 we compute the average based on those years. For
Kuwait and Portugal the data coverage does not include that period. Hence we use averages for the 2009-2010 period for Kuwait
and averages for the 2003-2005 period for Portugal in Table 1. However, these two countries are excluded from all summary
statistics and regressions shown from Table 4 to Table 7, though including them makes little difference.
6
exhibit the lowest average exports per firm (less than 800,000 USD). The tremendous difference
between mean and median exports per firm reflects the highly skewed exporter size distributions
in all countries with some very large exporters driving total exports.4 For Botswana, Chile, Mali,
Peru, South Africa, Mexico, and Malawi the skewness in the exporter size distribution is further
confirmed by the very high share of exports—more than 90 percent—accounted for by just the
top 5 percent of exporters. Interestingly, for most other developing countries those shares are
smaller than those for the developed countries in the sample, such as Norway, Sweden, and New
Zealand. Similarly, the U.S. and other European countries also exhibit high concentration, with
the top 5 percent of firms accounting for 80 percent or more of trade, as shown by Bernard,
Jensen, and Schott (2009) and Mayer and Ottaviano (2008) respectively.
The average number of products per exporter also exhibits a high degree of heterogeneity
across countries, ranging from 2 in Laos to 15 in South Africa. In terms of markets, the average
number of destinations per exporter is more similar across countries ranging from 1.4 in
Botswana to 4.8 in Cambodia and 6.8 in Belgium.
The measures of exporter dynamics also exhibit important variability across countries.
Entry rates range from 22 percent in Brazil to more than 50 percent in Malawi, Yemen, Laos and
Tanzania while exit rates range from 22 percent in Bangladesh to 61 percent in Malawi.5 Firstyear survival rates of new exporters vary between 23 percent in Cameroon and 61 percent in
Bangladesh. The magnitude of the survival rates in Table 1 suggests an extremely high attrition
rate of new entrants after just one year in export markets, particularly in Africa.
4
Using the same cross-country exporter-level dataset used for the construction of the Database, Freund and Pierola (2012) show
that exports are dominated by a small group of very large exporters (so-called ‘superstars’). These firms are remarkably larger
than the rest, they participate in many sectors and most importantly, they define sectoral exports and drive the export growth
observed in most countries.
5
These rates of entry into and exit from export markets are substantially larger than the rates of entry into and exit in production
reported by Bartelsman, Haltiwanger, and Scarpetta (2013). Focusing on census/registries of manufacturing firms (excluding 1employee businesses) they document turnover of 10-20 percent in industrial countries and somewhat higher in developing
countries.
7
The measures in Table 2 show a less pronounced variation in the number of exporters
across sectors. The number of exporters is higher for machinery (electrical and mechanical) and
plastics. Similar to the pattern within countries, the distribution of exporter size is also skewed
within sectors: the difference between average and median size is considerable. Larger exporter
size is more common in sectors that are highly resource-driven (mineral products and precious
metals) and where increasing returns to scale are important. As far as firm concentration is
concerned, the variation in shares of top exporters across sectors is not as wide as it is in the case
of countries. Again, precious metals and minerals are among the most concentrated sectors,
while firm concentration is lower in agricultural sectors. The average measures for product and
market diversification are more homogenous across sectors than across countries. Finally,
regarding exporter dynamics, entry and exit rates tend to be slightly higher in machinery and
transport, while exporter survival is slightly higher in agriculture.
In the analysis below, we also use data from the World Development Indicators (WDI) of
the World Bank on GDP and on GDP per capita in constant 2005 Purchasing Power Parity (PPP)
USD.6
III. Exporter Features and Country Characteristics
In this section, we evaluate how exporter characteristics and dynamics vary with country
size and stage of development. We first show the importance of both country and sector
characteristics in explaining exporter behavior. We then present three stylized facts on how the
structure of exporters changes as countries get larger or richer, controlling for sectoral variation
in exports and variation across destinations.
6
Note that all results presented from Table 4 to Table 7 are robust to the use of GDP and GDP per capita in current USD or in
constant USD.
8
We begin by evaluating the relative importance of country fixed effects and either HS2digit sector fixed effects or destination fixed effects in explaining exporter characteristics and
dynamics. Specifically, we perform an ANOVA with country and HS2-digit fixed effects, and
separately with country and destination fixed effects, on the main measures that we will explore
in detail below. 7
Panel A of Table 3 shows the results of the ANOVA with country and sector fixed
effects. On average, country and sector effects explain over 60 percent of the variation in
exporter characteristics, though somewhat less of the variation in exporter dynamics.8 With the
exception of exporter size, product diversification, and entrant survival, country effects explain a
greater share of the variation than sector effects. Still, the significance of sector effects implies
that it will also be important to control for them as we examine the relationships between country
characteristics and exporter characteristics.
We next repeat the ANOVA exercise using the country-destination level data available in
the database. The results are reported in Panel B of Table 3, and show that country and
destination fixed effects together explain somewhat less of the variation in the statistics than did
country and HS2 fixed effects. However, except for number of exporters and product
diversification, destination effects prevail over country effects. In that sense, it will be important
to control for destinations to ensure the results are not being driven by variation in trade barriers
or other destination-specific market access costs.
Next we present three stylized facts first focusing on basic characteristics of exporters,
then exporter diversification, and finally exporter dynamics.
7
Note that the sample used for concentration (fourth column) is relatively smaller because the share of exports accounted for by
the top 5 percent of exporting firms can be calculated only if a country-sector-year cell includes more than 20 exporters.
8
The percentages accounted for by country, sector, and residual effects in Panel A or country, destination, and residual effects in
Panel B do not add up to 100 percent due to the unbalanced nature of each of the cross-sectional datasets. Namely, in Panel A not
all sectors are exported by all countries while in Panel B not all destinations are served by all countries.
9
Stylized Fact 1: Larger and more developed economies have a larger export base (number of
exporters), larger average exporter size, and more concentrated export sectors among firms
(even after controlling for the sectoral distribution of exports and export destinations).
Figure 1 presents a scatter plot of export base, average exporter size, and concentration
and income per capita, using data averaged for the period 2006-2008.9 The figure shows that
more developed economies have a larger number of exporters, larger average size of exporters,
and more concentrated export sectors in terms of firms. Figure 2 uses income instead of income
per capita and shows that larger countries have more exporters, larger exporters, and higher
concentration. We next explore these relationships in more detail, controlling for sectoral
variation in exports, considering cross-section regressions based on the data averaged across the
period 2006-2008.
Table 4 reports the results from the cross-section regressions. Specifically, we regress
total exports, export base, exporter size, and firm concentration on GDP per capita and GDP.
Columns (1)-(4) report the results using the country data. All measures are positively associated
with country size and stage of development, though the correlation with average exporter size
appears to stem from country size and not from stage of development. Similarly, country size has
a greater impact on the number of exporters than the level of development. However, in terms of
firm concentration, stage of development seems to matter more.
A concern about the regressions in columns (1)-(4) is that larger or more developed
countries may be more likely to specialize in sectors with increasing returns to scale, thus the
results may be driven by the sectoral composition of trade and not by a change in exporter
9
The other scatter plots on diversification and exporter dynamics in Figures 1 and 2 are discussed as part of the stylized facts
below.
10
characteristics within sectors as countries develop. To account for this possibility, we next
consider the same measures at the country and HS 2-digit sector level averaged across the period
2006-2008 and control for HS 2-digit sectoral fixed effects.10 The results, shown in columns (5)(8) of Table 4, indicate that controlling for potential differences across sectors, larger and richer
countries have both more and larger exporters. In addition, within sectors, the correlations with
export concentration become highly significant for both stage of development and country size,
suggesting that within sectors exports are more concentrated among the top firms in both richer
and larger countries. The economic magnitude of these effects is such that if GDP per capita
increased from the value for the country in the 25th percentile (Laos) to the value for the country
in the 75th percentile (Mauritius) the number of exporters would increase by 92 percent, average
exports per firm would increase by 63 percent, and the share of the top 5% exporters would
increase by 8 percentage points.11
Columns (9)-(13) in Table 4 use the country-destination level data and results are
qualitatively similar. Bigger and richer countries have more exporters, larger exporters, and
greater export concentration in any given destination. These regressions show that even after
controlling for market-specific characteristics, such as trade costs, more developed and larger
countries have larger exporters that export larger volumes.
Note that when we control for either sector or destination, the results on average size and
concentration are both statistically and economically more significant than on aggregate. This
implies that within a destination or sector, exporters from smaller and poorer countries on
10
For these, and all subsequent regressions at the country-sector level, robust standard errors clustered at the country level are
reported.
11
These magnitudes are obtained as the product of the difference in logs GDP per capita, 1.744, and the corresponding
coefficient in columns (6)-(8) of Table 4: 0.527, 0.359, and 0.045.
11
average export relatively less; but that across all sectors or all destinations, they export relatively
more in sectors with large exporters or to high-demand destinations.
Stylized Fact 2:
Controlling for the sectoral distribution of exports, larger and richer
countries have more diverse exporters in terms of products and/or markets. Stage of
development matters more for exporter diversification than does country size.
Richer
countries also have a larger share of multi-product multi-destination firms and these firms
weigh more heavily in total exports.
We next examine the relationship between the average number of products and average
number of destinations per exporter and country characteristics. Figures 1 and 2 show that as
countries get richer and larger, exporters tend to be more diversified product- and destinationwise. In Table 5 we look at these relationships in more detail.
Columns (1) and (2) in Table 5 report the results from the cross-section regressions of the
average number of products and destinations per firm on GDP per capita and GDP using the
2006-2008 averaged country data. We find that firms in richer countries are more diversified
with respect to products and firms in larger countries serve a greater number of destinations. As
shown in the ANOVA (Table 3), sector effects explain a large share of the variation in the
average number of products per exporter. While this could suggest that multi-product firms are
relatively more common in some sectors, it could also reflect that the differences in product
diversification are endogenously determined by the structure of the HS classification, which
entails very different degrees of disaggregation by 2-digit sectors.
Next, we control for HS 2-digit fixed effects. The results, shown in columns (3) and (4),
provide strong evidence that within sectors exporters in richer countries export more products
12
and export to more destinations. However, the coefficients on per capita income are much larger
than those on income (and in the case of product diversification only the coefficient on per capita
income is significant), indicating that exporter diversification is associated to a greater extent
with stage of development than with country size. In terms of economic significance our results
imply that increasing GDP per capita from the 25th to the 75th percentile in our sample would
increase the number of products per exporter by 16 percent and the number of destinations per
exporter by 21 percent. In column (5), we examine how exporter diversification across products,
within a destination, changes as countries grow. Again, we find a strong positive correlation
with per capita income.
Another way of looking at the importance of diversification is to examine the share of
multi-product multi-destination firms in total trade. The recent trade literature shows a dominant
role for firms that export many products to many destinations in explaining trade flows (see, for
example, Bernard, Jensen and Schott 2009 on the U.S.; Eaton, Kortum and Kramarz, 2008 on
France; and Amador and Opromolla, 2013 on Portugal). Using the firm-level information from
developing countries over the 2006-2008 period, we calculate for each country the average share
of exporters and of total exports accounted for by single-product single-destination firms and by
firms exporting to more than four countries and to more than four destinations and record them
in Table 6.12 Unfortunately, we do not have access to the firm-level raw data for the developed
countries, used in other parts of this paper, to calculate these indices for them. The results show
that single-product, single-destination firms represent more than a third of exporters on average
across developing countries. They represent an even higher share over 40 percent in most
12
This analysis is based on an additional component of the Database which is a set of matrix tables showing for each developing
country and year the joint distribution of exporters and of total exports by the number of HS 6-digit products exported and the
number of destination markets served. This information can provide insights onto the importance of multi-product multidestination exporters across countries and over time.
13
African countries as well as in Albania and Mexico. The percentages shown in Table 6 are quite
close to the 40 percent reported by Bernard, Jensen and Schott (2009) for the U.S.13 However,
these single-product single-destination firms account for a minimal fraction of total exports, on
average less than 3 percent across countries. That fraction is particularly low in Botswana, Costa
Rica, and Niger at 0.5 percent or less.
Table 6 also shows that firms exporting more than four products to more than four
destinations represent a relatively small proportion of exporters, 12 percent on average across
countries. But those percentages exhibit a substantial degree of heterogeneity across countries
ranging from about 3.5 percent in Albania and Botswana to 29.4 percent in Cambodia. These
multi-product multi-destination firms account for a large share of total exports in all countries,
more than 60 percent on average. But again large differences are observed across countries: in
Cambodia, Costa Rica, and South Africa that share is actually close to 80 percent whereas in
Albania and Niger it is less than 30 percent.
We next examine how the share of multi-product and multi-destination firms varies with
stage of development and country size. As noted above, we do not have access to this data for
the more developed countries in our sample, giving us less variation in level of development. As
a result, we focus on country-sector data. Results are reported in columns (6) and (7) in Table 5.
Again we find that as countries develop, more resources are allocated to the more diversified
exporters within sectors.
Stylized Fact 3: Exporter turnover is significantly lower in more developed countries, while net
entry is invariant with respect to income per capita.
13
Note, however, that the U.S. percentage is based on products defined at the HS 10-digit level.
14
Figures 1 and 2 show exporter turnover and stage of development and turnover and
country size, respectively. In both cases, there is a negative relationship indicating that richer
and larger countries have lower entry and exit rates. We next explore these relationships in more
detail.
Columns (1)-(5) in Table 7 show the results from cross-section regressions of exporter
entry, exit, survival of new exporters, net entry, and turnover on income and income per capita
based on the 2006-2008 averaged data. The estimates show that entry and turnover rates are
lower in larger countries, but point to no relationship between net entry or entrant survival rates
and stage of development.
In columns (6)-(10) in Table 7 we report results using country-sector level data,
controlling for sector fixed effects. This is crucial here as some sectors have much lower
turnover than others, as highlighted by the relative importance of sector effects in the ANOVA
(Table 3). The results show that controlling for the sectoral composition of trade, stage of
development has important effects on exporter dynamics. In particular, richer countries have
lower turnover and higher rates of survival of new exporters, leaving net entry roughly
unchanged. The economic magnitudes of the results are sizable—increasing GDP per capita
from the 25th to the 75th percentile in the sample would lead to a decrease in entry rates by 11.5
percentage points, would lead to a decline in exit rates by 12.4 percentage points, and to an
increase in the one-year survival rate of new exporters by 7.3 percentage points.
In columns (11)-(15) in Table 7 we report results using country-destination level data and
controlling for destination fixed effects. The results are qualitatively similar to the results using
country-HS2 data, though the magnitudes are smaller. In this case, there is also some evidence
that size matters, with larger countries having lower entry and exit and higher survival rates.
15
The coefficients on entry, exit, survival, and turnover are both economically and
statistically more significant when either sectors or destinations are controlled for than when
aggregate data are used, implying that poorer countries tend to have exporters that enter and exit
less in industries and destinations where entry and exit is rare. These results show that exporter
dynamics change as countries get richer. In poor countries turnover is largely a process of trial
and error where many firms enter into export markets and exit almost immediately. In rich
countries, fewer but more resilient exporters enter in any given year.
Robustness of Stylized Facts
In Appendix 3, we consider a number of alternative specifications to test the robustness
of these results. (i) We re-estimate the regressions using country-destination data but focusing
only on manufacturing industries (available only for the 32 developing countries for which we
have access to the raw exporter-level data). (ii) We include distance in the cross-section
regressions that use country-destination data to control for trade costs. (iii) We re-estimate the
specifications for average exporter size, using data on incumbent exporters only, in order to
exclude possible biases due to “partial year” data. 14 Finally, (iv) for the results on average
exporter size and number of exporters, we re-estimate the regressions using a short panel from
2004 to 2008 with country-sector and country-destination fixed effects to see how exports adjust
over time as income grows. 15 For this specification, we include GDP and GDP per capita
14
Bernard et al. (2013) show that the entry and exit of firms into and from export markets at different points in a calendar year
means that the use of calendar annual data on export values to measure the average size of exporter entrants or exiters may
introduce a bias - mainly their size being under-estimated during the first year - which they refer to as “partial year” effects. Our
consideration of the average size of incumbent exporters addresses this concern to the extent that bias should not permeate the
calculation of average size for exporters that have not interrupted their exporting activity.
15
We also try panel regressions with the other dependent variables: concentration, diversification, and dynamics. All results are
robust in sign and significance when country-sector or country-destination fixed effects are included, however when we include
year fixed effects, many of the results become insignificant—though the signs remain robust. This suggests that the change over
the period in these variables is coming largely from global demand growth, making it difficult to identify correlation with home
16
separately, as over a short period they are highly correlated because population is relatively
constant. We consider exports to the top 10 destinations only so as to avoid noise due to small
destinations or changes in destinations. This final specification shows that in the short run as
countries grow, much of the corresponding export growth is driven by larger exporter size as
opposed to more exporters, especially when per capita income is considered. The other results
remain qualitatively similar for all alternative specifications.
IV. Theoretical Interpretation
In this section, we review the evidence presented above on exporter characteristics and
stage of development in light of recent trade models with heterogeneous firms.
The basic theory of trade and heterogeneous firms, following Melitz (2003), focuses on
the reallocation of production across firms within an industry in response to trade liberalization.
More recently, efforts have been made to extend the model to make it consistent with other
observed features of exporter behavior that have become known via the growing empirical work,
such as, the numerous zeros in international trade (Helpman, Melitz and Rubinstein 2008 and
Eaton, Kortum and Sotelo 2012), or the large numbers of very small exporters (Arkolakis 2010).
While the theory was not developed to explain how economies at various stages of development
or scales differ in terms of exporter behavior, there are some features of these models we can use
to address these relationships.
The standard heterogeneous firm model assumes firms are differentiated by the variety
they produce and by their productivity, and that there is a fixed entry cost into exporting. A firm
will enter the export market if its net export profits cover the fixed exporting cost. Because of
country characteristics. Note that this is a fairly strong test of the results, as the time period is only 5 years and the fixed effects
swamp up much of the variation (see the high R squares in Table iv in Appendix 3).
17
the entry costs, there is a cutoff productivity level, such that firms at or above it will export while
those below it produce only for the domestic market.
Thus, firms that export are more
productive than firms that produce only for the domestic market, and export quantity is
increasing in productivity. This is consistent with our findings to the extent that developed
countries are richer because their firms are more highly productive or they have a greater number
of highly productive firms. Thus, if we allow for variation in the productivity distribution of
firms across countries, the model would be consistent with our findings on stage of development
although it would not help to explain them. However, it would be harder to justify our findings
on country size.
This is in part because constant elasticity of substitution (CES) preferences assumed in
the standard model imply that the cutoff productivity level for an exporter to enter a market does
not change with domestic market size, which implies that differences in the number or size of
exporters must stem from trade costs. Moreover, in these models, if variation stems from fixed
trade costs, exporter size and the number of exporters in a country would be inversely related
since a higher productivity threshold would mean that there are fewer but bigger exporters.
While if variation stems from variable trade costs, then all of the adjustment should be in the
extensive margin. Thus, our findings of a positive correlation of country size and stage of
development with number of exporters and exporter size could not be generated.
The higher concentration of exports in the top 5 percent of exporting firms in larger
countries and at higher levels of income that we show is also not consistent with a standard
heterogeneous firm model. With firm productivity following a Pareto distribution, the larger
average exporter size and the increasing concentration at the top of the size distribution as
countries grow could come from either more skewed distribution in the Pareto shape parameter
18
(a lower shape parameter) or a larger support for the firm productivity distribution in bigger and
richer countries.
A number of models are consistent with larger markets having more and larger firms, but
not necessarily with exporters from larger markets exporting larger volumes to any given
destination. For example, Melitz and Ottaviano’s (2008) model with variable markups allows
aggregate demand to matter for entry and firm size. In their model, larger markets demand more
of each variety, generating more entry. However, when prices rise, demand falls off more
sharply for each specific variety, inducing high-cost firms to exit. As a result, larger markets are
characterized by tougher competition, with both more firms and bigger firms.16 This model can
explain why firms export larger quantities to larger destination markets, but not why firms from
larger markets export more than firms from smaller markets to a particular destination market.
In recent work, Spearot (2013) adapts the Melitz-Ottaviano model to address this issue by
allowing the Pareto shape parameter of the firm productivity distribution to vary across
countries. He shows that such an accommodation can explain variation in the average size of
exporters. He further shows that the response of export flows across countries of varying sizes to
a trade shock can be used to estimate the shape parameters across countries. Using this
methodology, he finds that larger and richer countries have more skewed distributions, a result
that could explain our findings on the increasing share of the top 5 percent and large average size
in large and rich countries. His model, however, does not explain why the shape parameter is
more skewed in larger or richer countries.17
16
Chaney and Ossa (2012) also address the relationship between market size and firm size. They develop a model with labor
division, where the production of a final good requires the performance of a number of tasks, each of them carried out by a
specialized team. The larger the output, the larger the number of teams and the lower the average costs–each team performs more
efficiently lowering marginal costs. In this context, an increase in market size –represented by an increase in the number of
consumers–increases demand elasticity and is associated with more firms and larger firms.
17
One potential explanation for why this might be the case in richer countries derives from the work of Hsieh and Klenow
(2012), who find that weak firm dynamics prevent developing countries from reaping the benefits of large firms. They
19
Alternatively, Helpman, Melitz and Rubinstein (2008) employ a heterogeneous firm
model with a productivity distribution truncated from above to explain why some countries have
zero trade. Such a model may also be able to explain why average exports are greater in larger
and richer countries if the upper limit is increasing in size and stage of development. For
example, a larger pool of potential technologies or managers in bigger and richer countries may
yield a wider support. In this case, large and rich countries would have some very large
exporters, raising the average size and the share of exports in the top 5 percent. Overall, the
results on size and concentration are consistent with variation in the firm productivity
distribution across countries, and more specifically with either a lower shape parameter or a
wider support in bigger and richer countries. This is an important area for future research.
Turning next to our results on diversification, a number of trade models with
heterogeneous firms have also incorporated multi-product multi-destination firms. For example,
in the model of Bernard, Redding and Schott (2011), firms are heterogeneous in terms of overall
ability and expertise in each product, in addition they face fixed costs in terms of labor to export
each product and serve each market. Only higher ability firms are able to generate variable
profits to cover those fixed costs and thus supply a wider range of products to each market. This
model helps to explain why the largest firms are multiproduct, as these are the most able firms,
and also why they account for such a high share of exports. It also allows for across and within
firm reallocation of resources in response to trade liberalization.
While, like the standard
heterogeneous firm model, the model does not directly address stage of development, assuming
richer countries have more able firms or managers, our results on increased diversification and
their increased share of exports as countries grow would hold. This helps to explain why richer
hypothesize that this could be the result of constraints to growth in developing countries, via for example taxes or weak
investment incentives, which constrain aggregate TFP because of a misallocation of resources away from the high-productivity
firms. Over time this would lead to too many small firms and not enough large firms.
20
countries have more diversified firms, and can also explain why these firms account for a larger
share of total exports as countries develop since there are more highly productive firms that are
able to export widely.18
What is harder to explain with existing trade models with heterogeneous firms are our
results on exporter turnover because those models have little to say about dynamics (Melitz and
Redding 2012). Some recent work, however, has used uncertainty about export productivity to
explain the high rates of entry into and immediate exit from foreign markets observed in the data
(Albornoz et. al. 2010, Segura-Cayuela and Vilarrubio 2008 and Freund and Pierola 2008). In
these models, there is high entry because there is effectively an option value of remaining in the
export sector if the firm is successful in foreign markets.
To illustrate how this works, consider the following simple case: each firm has
probability p of getting high positive export profits, H, and probability (1-p) of getting negative
export profits, L. Exporting requires fixed cost F. In a repeated game, with discount factor δ
(smaller than 1), a firm will enter the export sector if
(1)
(
)
The important point is that there is an asymmetry because getting a good profit draw yields
positive profits forever in this repeated game (
), while export failure yields only one period of
negative profits (L) followed by exit. This implies that a firm may enter the export sector even if
its expected one-period profits are negative, pH+(1-p)L<0. This can explain why there is so
much entry into exporting with high rates of immediate exit corresponding also to low one-year
survival rates of new exporters.
18
Note that the firm-level results are distinct from evidence on production and industrial exports on diversification, where
empirical evidence shows that countries first diversify and only later specialize (Imbs and Wacziarg 2003 , Klinger and
Lederman 2006). The results are not contradictory, as it is possible that firms on average become more diversified as countries
grow, but if the products are ultimately similar across firms at high levels of income, this is consistent with increased
specialization by industry or product at the country level.
21
For both exporter entry and exit to be negatively correlated with the level of
development, while net entry is roughly invariant, it would be necessary in the model above that
fixed costs of exporting are lower, the discount factor is higher, or that the gap between H and L
is greater in developing countries.19 The first, lower fixed costs of exporting, seems unlikely
given that existing evidence points to higher costs of entering export markets in poorer
countries.20 The second, higher discount factor also seems unlikely as developing countries have
higher interest rates/discount rates for the future, indicating that money now is worth more than
money later in poor countries.
Finally, firms in developing countries may have greater
variability in profit draws (outcomes). Even if one-period expected profits are the same across
countries (pH + (1-p) L= k), a greater H and lower L will lead to more entry because H weighs
more heavily on the decision to enter. In this case, the first term in Eq. (1) would be relatively
larger for developing countries, implying that more firms would enter the export sector in poorer
countries, while a similar or even a lower fraction of entrants would survive (depending on p),
yielding higher overall entry and exit. Exit goes up precisely because there are more entrants.
Indeed, this is consistent with high correlations between entry and exit at the country level and
country-sector level. For example, using the 2004-2008 panel data, the simple within-country
correlation between entry and exit is 0.77 and the within country-sector the correlation is 0.76.
Search and matching models of trade partnerships can also generate similar exporter
dynamics. While these models were designed to understand the importance of networks in
international trade, they also offer conditions under which there will be more or less search. For
example, Rauch and Watson (2003) show that with imperfect information about the ability of a
19
Note that a higher probability of a good outcome in developing countries would create more entry but is not consistent with the
lower rates of survival that are observed in developing countries.
20
A simple correlation between various indicators on trade costs from the Doing Business database and GDP per capita using all
available countries as well as the countries in the Exporter Dynamics Database sample is clearly negative and significant.
22
developing country exporter to supply a large order, developed country buyers may have more
small trials before investing in a developing country exporter for a long-term trade relationship.
This would lead to more exporter entry and exit from developing countries. In a different type of
model, Araujo, Mion, and Ornelas (2012) integrate work on institutions and heterogeneous firm
models of trade and show that there is less entry, less exit, less turnover and greater survival rates
in countries with better institutions, resembling closely our findings with per capita income.
In sum, the results on exporter characteristics are consistent with various strands of the
recent theoretical work, with some accommodations. The results on average exporter size and
increasing concentration are consistent with a heterogeneous firm trade model with variation
across countries in the Pareto shape parameter or the support of the firm productivity
distribution. The results on diversification and stage of development follow from theories with
multi-product and multi-destination firms.
The results on exporter dynamics and stage of
development are consistent with models with uncertainty about export profitability or the
viability of a particular importer-exporter match.
Overall, the systematic variation in exporter behavior across countries by size and stage
of development observed in the data is suggestive of more efficient resource allocation as
countries get richer and larger. While tweaking existing models can explain specific features of
the data, the results point to the need for a dynamic model to develop a better understanding of
how exporter behavior changes with country size and stage of development. For example, why
would the Pareto shape parameter be more skewed or the support wider in more developed or
larger countries? Such a model may also help explain whether the changing firm characteristics
and dynamics help drive income growth or whether demand growth drives firm behavior.
23
V. Conclusion
While the literature on exporter dynamics is rapidly growing, there has not been a
comparison across a large number of countries of different sizes and at differing stages of
development. The Exporter Dynamics Database presented in this paper addresses that limitation
by providing the research community and policymakers with a set of measures of exporter
characteristics and dynamics that allow for cross-country comparisons, that expand our
understanding of how exporting happens and show novel results on how these are related to
stage of development. The Database compiles measures for 45 countries (37 of which are
developing countries) covering the universe of firm-level export transactions primarily for the
period between 2003 and 2010.
The data point to systematic ways in which the micro structure of exports changes as
countries develop. As countries grow, they have more and larger (likely better) exporters and
also relatively more concentrated export sectors. Product and market diversification of exporters
is also positively correlated with stage of development. Together with results from previous
studies showing that exporters are more productive than non-exporters and larger exporters are
more productive than smaller exporters, our results point towards allocative efficiency in export
markets as countries develop.
Of interest is how reallocation happens. In particular, exporter dynamics are also closely
linked with stage of development, with richer countries experiencing lower entry and exit rates,
higher survival of new firms, and net entry invariant to per capita income. This implies that as
countries develop, there is less entry of firms that immediately exit to generate the same rates of
net entry. We highlight that with uncertainty about export profitability at the firm level, this
could reflect lower variability in export potential at higher stages of development.
24
The measures in the Database will allow the examination of several interesting crosscountry questions, cross-country cross-sector questions, and within-country questions. The
measures can also be used as controls in estimations that require exporter characteristics at the
country-industry level. In particular, as the measures in the Database offer the first opportunity to
study exporter characteristics and dynamics on a global basis, some of the facts described above
open the door to questions such as: Does trade promote growth via firm size or firm count? How
can countries attract more large multi-product firms? What determines exporter survival? How
comparative advantage is related to the typical exporter characteristics in an industry? And many
others that will hopefully be addressed in future research to be conducted using the Database.
25
References
Amador, J. and L. Opromolla (2013). “Product and Destination Mix in Export Markets,” Review
of World Economics 149: 23-53.
Albornoz, F., H. Calvo Pardo, G. Corcos and E. Ornelas (2010) “Sequential Exporting,” Journal
of International Economics 88: 17-31.
Andersson, M., H. Lööf and S. Johansson (2008). “Productivity and International Trade: Firm
Level Evidence from a Small Open Economy,” Review of World Economics 144: 774-801.
Araujo, L., G. Mion and E. Ornelas (2012). "Institutions and Export Dynamics," CEPR
Discussion Papers 8809, C.E.P.R. Discussion Papers.
Arkolakis, C. (2010). “Market Penetration Costs and the New Consumers Margin in
International Trade,” Journal of Political Economy 118(6): 1151–99.
Bartelsman, E., J. Haltiwanger and S. Scarpetta (2013). "Cross-Country Differences in
Productivity: The Role of Allocation and Selection," American Economic Review, American
Economic Association, vol. 103(1), pages 305-34, February.
Bernard, A., J. Jensen, and P. Schott (2009). “Importers, Exporters and Multinationals: A Portrait
of Firms in the U.S. that Trade Goods,” in T. Dunne, J. Jensen and M. Roberts (eds.), Producer
Dynamics: New Evidence from Micro Data. University of Chicago Press.
Bernard, A., J. Jensen, S. Redding, and P. Schott (2007). “Firms in International Trade,” Journal
of Economic Perspectives 21: 105-130.
Bernard, A., J. Jensen, S. Redding, and P. Schott (2011a). “The Empirics of Firm Heterogeneity
and International Trade,” NBER Working paper No. 17627.
Bernard, A., S. Redding, and P. Schott (2011b). “Multi-Product Firms and Trade Liberalization,”
Quarterly Journal of Economics 126: 1271-1318.
Bernard, A., R. Massari, J. Reyes, and D. Taglioni (2013). “Exporter Dynamics, Firm Size and
Growth, and Partial Year Effects,” World Bank mimeo.
Cebeci, T. (2012). “A Concordance Among Harmonized System 1996, 2002 and 2007
Classifications,” World Bank mimeo available at http://econ.worldbank.org/exporter-dynamicsdatabase.
Cebeci, T., A. Fernandes, C. Freund, and M. Pierola (2012). “Exporter Dynamics Database,”
Policy Research Working Paper No. 6229, The World Bank.
Chaney, T. and R. Ossa (2012). "Market Size, Division of Labor, and Firm Productivity," CEPR
Discussion Papers 9069, C.E.P.R. Discussion Papers.
26
De Lucio, J., R. Mínguez-Fuentes, A. Minondo, and F. Requena-Silvente (2011). “The Extensive
and Intensive Margins of Spanish Trade,” International Review of Applied Economics 25: 615631.
Eaton, J., M. Eslava, M. Kugler, and J. Tybout (2008). “Export Dynamics in Colombia:
Transactions Level Evidence,” Borradores de Economia No. 522, Banco de la Republica,
Colombia.
Eaton, J., S. Kortum, and F. Kramarz. (2008). “An Anatomy of International Trade: Evidence
from French Firms,” Econometrica 79: 1453-1498.
Eaton, J., S. Kortum, and F. Sotelo. (2012). “International Trade: Linking Micro and Macro,”
NBER Working Paper 17864.
Ekholm, K., A. Moxnes, and K. Ulltveit-Moe (2012). “Manufacturing Restructuring and the
Role of Real Exchange Rate Shocks: A Firm-Level Analysis,” Journal of International
Economics 86: 101-117.
Fabling, R. and L. Sanderson (2012). “Whatever Next? Export Market Choices of New Zealand
Firms,” Papers in Regional Science 91: 137-159.
Freund, C. and M. Pierola (2010). “Export Entrepreneurs Evidence from Peru,” Policy Research
Working Paper No. 5407, The World Bank.
Freund, C. and M. Pierola (2012). “Export Superstars,” Policy Research Working Paper No.
6222, The World Bank.
Helpman, E, M. Melitz, and Y. Rubinstein (2008). “Estimating Trade Flows: Trading Partners
and Trading Volumes,” The Quarterly Journal of Economics, 123: 441-487.
Hsieh, C. and P. Klenow (2012). “The Life-cycle of Plants in India and Mexico,” NBER
Working Paper 18133, National Bureau of Economic Research, Inc.
Iacovone, L. and B. Javorcik (2010). “Multi-Product Exporters: Product Churning, Uncertainty,
and Export Discoveries,” Economic Journal 120: 481-499.
Imbs, J. and R. Wacziarg (2003). "Stages of Diversification," American Economic Review,
American Economic Association, vol. 93(1), pages 63-86, March.
Klinger, B. and D. Lederman (2006). "Diversification, innovation, and imitation inside the
Global Technological Frontier," Policy Research Working Paper Series 3872, The World Bank.
Manova, K. and Z. Zhang (2012). “Export Prices across Firms and Destinations,” Quarterly
Journal of Economics 127: 379-436.
27
Masso, J. and P. Vahter (2011). “Exporting and Productivity: the Effects of Multi-Market and
Multi-Product Export Entry,” University of Tartu, Faculty of Economics and Business
Administration Working Paper No. 83.
Mayer, T. and G. Ottaviano (2008). “The Happy Few: The Internationalisation of European
Firms,” Intereconomics: Review of European Economic Policy 43: 135-148.
Melitz, M. (2003). “The Impact of Trade on Intra-Industry Reallocations and Aggregate Industry
Productivity,” Econometrica 71: 1695-1725.
Melitz, M. and S. Redding (2012). “Heterogeneous Firms and Trade,” NBER Working Paper
18652.
Segura-Cayuela, R. and J. Vilarrubia (2008) “Uncertainty and Entry into Export Markets,”
Banco de España Working Paper No. 0811.
Rauch, J. and J. Watson (2003). "Starting Small in an Unfamiliar Environment," International
Journal of Industrial Organization, Elsevier, vol. 21(7), pages 1021-1042, September.
Schott, P. and J. Pierce (2011). “Concording U.S. Harmonized System Categories over Time,”
Journal of Official Measures forthcoming.
Spearot, A. (2013). “Tariffs, Competition, and the Long of Firm Heterogeneity Models,”.
University of California – Santa Cruz. Mimeo
Wagner, R. and A. Zahler (2011). “New Exports from Emerging Markets: Do Followers Benefit
from Pioneers?,” MPRA Paper No. 30312, University Library of Munich, Germany.
United Nations (1998). International Merchandise Trade Statistics: Concepts and Definitions.
United Nations (2008). International Merchandise Trade Statistics: Supplement to the Compilers
Manual.
28
Figure 1: Correlations of Selected Database Measures with GDP per Capita
(averages over 2006-2008)
Number of Exporters - GDPpc
17
12
Mean Exports per Exporter - GDPpc
ESP
16
Ln Mean Exports per Exporter
BRA
15
NER
MWI
13
4
7
8
9
10
CRI
ZAF BWA
TUR
BGD
COL
CMR
ECU
JOR
EGYSLVDOM
GTM
UGA
LAO
BFA
TZA SEN PAK
MUS
NIC
IRN
BGR
KEN
YEM
MKD
LBN
ALB
11
6
7
8
Ln GDPpc
9
ESP
NOR
EST NZL
10
11
Ln GDPpc
R-squared=0.58
R-squared=0.13
Number of Products per Exporter - GDPpc
3
1
Share of Top 5% Exporters - GDPpc
CHL
PERZAF MEX
MUS
DOM
MKD
JOR
BGR
CMR
BRA
CRI
SLV COL
KEN
TUR
EGY ECU
LBN
GTM
UGA
NIC
MAR
PAK
IRN
SEN
BGD
SWE
NOR
ZAF
ESP
BEL
TUR
MUS
KHM
KEN
EST
GTM
MAR
NIC
SEN
SLVPER
PAK
1.5
ALB
YEM
CMR
MWI
BGD
TZA
NER MLI
BFA
UGA
LBN
MEX
BWA
BGR
IRN
CRI
BEL
EST NZL
SWE
NOR
COL
DOM
ECUMKD
ESP
CHL
ALB
JOR
1
.8
.6
YEM
NZL
2.5
LAO
TZA
BFA
2
MLI
Ln Number of Products per Exporter
BWA
MWI
NER
Share of Top 5% Exporters
SWE
PER
MAR
MLI
NER
6
CHL
MEX
KHM
14
10
6
8
Ln Number of Exporters
BEL
TUR
MEX
SWE
BEL
ZAF
BRA
NOR
PAK
BGR
IRN
NZL
COL
EGY
CHL
PER
BGD
MAR
LBN
KEN
EST
GTM
ECUMKD
CRI
SLVDOM
MUS
TZA
JOR ALB
BWA
NIC
UGA
CMR
SEN
MWI
KHM
YEM
LAO
BFA
MLI
.4
KHM
LAO
6
7
8
9
10
11
6
7
8
Ln GDPpc
9
10
11
Ln GDPpc
R-squared=0.08
R-squared=0.20
Entry Rate - GDPpc
2
Number of Destinations per Exporter - GDPpc
MWI
YEM
LAO
TZA
.5
BEL
CMR
UGA
IRN
.5
NOR
DOM
ALB
ESP
ALB
PER
BGR CHL
JOR
MKD
NIC
KHM
BGD
LAO
NER
EST
BWA
ECU
SEN
KEN
NOR
MEX
MAR
TUR
GTM SLV COL
MUS
CRI
ZAF
.3
1
SEN PAK
.4
TUR
ZAF
CHL
CRI
JOR
LBN
NZL
CMR
COL
EGY PER MUS
EST
KEN
MAR
GTM SLV
BGR
TZA
UGA
YEM
BFA
ECU
DOM
MKD MEX
MLI
NIC
IRN
BGD
MWI
MLIBFA
SWE
ESP
Entry Rate
1.5
KHM
PAK
BWA
BEL
NZL SWE
EGY
0
.2
BRA
6
7
8
9
10
11
6
7
8
Ln GDPpc
.6
BGD
KHMPAK
BRA TUR
Survival Rate of Entrants
YEM
.3
.4
BFA
MLI
UGA
DOM
LAO
BWA
BGR
.5
ESP
SEN
NOR
NZL BEL
SWE
NZL
BEL
CHL
KEN
TZA
EST
UGA
ESP
MWI
CMR
BRA
.2
BGD
7
BFA
EST
ECU
MEX
PER
CHL
MKD
NICMAR
ALB
JOR
COL MUS
KHM
SLV
GTM
TUR
EGY
PAK
CRI
ZAF
SEN
ZAF
CRI
ALB
MKD
SLVPER
MAR
MUS
GTM
COL
ECU IRN
DOM
BWA
MEX
NIC
MLI
.4
CMR
TZA
KEN
EGY
JOR
LAO
.3
.5
IRN
.2
Exit Rate
11
Survival Rate of Entrants - GDPpc
Exit Rate - GDPpc
6
10
R-squared=0.18
MWI
.6
9
Ln GDPpc
R-squared=0.14
8
9
10
6
11
7
8
9
Ln GDPpc
Ln GDPpc
R-squared=0.00
R-squared=0.12
29
10
11
Turnover Rate - GDPpc
1.2
Net Entry Rate - GDPpc
.1
LAO
MWI
UGA
YEM
1
JOR
ALB
ECU
CHL
PER
MKD TUR
KHM
SEN
EST
CRI
BEL
GTM
ZAF BWA
NIC
CMR
SLV
PAK
ESPSWE NOR
COL
DOM
NZL
MAR
BRA
MUS
EGY
MEX
BGR
YEM
KEN
IRN
TZA
CMR
LAO
DOM
BFA
UGA
KEN
MLI
.8
Turnover Rate
BWA
ECU BGR
PER
CHL
MKD
ALB
MEX
JOR
NIC
MAR
COL
MUS
TUR
GTM SLV
SEN
IRN
KHM
.6
.05
0
TZA
MLI
BFA
-.05
Net Entry Rate
BGD
PAK
BGD
EST
ESP
BEL
NZL SWE
CRI
ZAF
EGY
NOR
BRA
.4
-.1
MWI
6
7
8
9
10
11
6
Ln GDPpc
7
8
9
10
11
Ln GDPpc
R-squared=0.01
R-squared=0.16
Notes: The measures plotted are based on measures in the country-year level dataset averaged across the 2006-2008 period for
each country. GDP per capita is measured in 2005 PPP U.S. dollars per inhabitant. Kuwait and Portugal are excluded from these
plots due to lack of data for the 2006-2008 period.
30
Figure 2: Correlations of Selected Database Measures with GDP
(averages over 2006-2008)
17
Mean Exports per Exporter - GDP
12
Number of Exporters - GDP
ESP
MEX
BRA
BRA
CHL
IRN
MEX
KHM
15
PER
MLI
CRI
BWA
NER
NER
ZAF
NOR
BGD
NZL
DOM ECU
GTM
UGA
TZA
BGR
KEN
YEM
LBN
MKD
SWE
MAR
EST
JOR CMR
SLV
LAO
BFASEN
MWI MUSNIC
14
8
6
MAR
LBN
KEN
EST
GTM
MKD
DOM ECU
SLVCRI
MUS
ALB
TZA
JOR
BWA
NIC
UGA
CMR
SEN
MWI
KHM
YEM
LAO BFA
MLI
CHL
PER
BGD
SWE
BEL ZAF
PAK
COL
EGY
16
Ln Mean Exports per Exporter
10
Ln Number of Exporters
BGR NZL
BEL
TUR
NOR
ESP
TUR
COL
EGY
PAK
IRN
4
13
ALB
23
24
25
26
27
28
23
24
25
Ln GDP
27
28
R-squared=0.24
Number of Products per Exporter - GDP
3
1
Share of Top 5% Exporters - GDP
CHL
NOR
PER
SWE ZAF
SEN
EST
YEM
.6
ALB
BEL
COL
EGY
PAK
BRA
TUR
KHM
EST
LBNGTM
SLV KEN
BWA
SEN
NIC
MKD
MWI
NER MLI BFA
BGD
TUR
BEL
MUS
IRN
1.5
NIC
ZAF
MEX
ESP
NZL
BGR
CRI
DOM
YEM
TZA
CMR
UGA
PER
MAR
MEX
SWE
NOR
CHL
BGD
ECU
IRN
PAK
COL
ESP
ALB
JOR
1
.8
DOM
JOR CMR
BGR
SLVCRIKEN
ECU
LBNGTM
UGA
MAR
2.5
NZL
TZA
2
Ln Number of Products per Exporter
BWA
MLI
MWI
NER
LAO
MUS
BFA
MKD
Share of Top 5% Exporters
26
Ln GDP
R-squared=0.78
.4
KHM
LAO
23
24
25
26
27
28
23
24
25
Ln GDP
26
27
28
Ln GDP
R-squared=0.01
R-squared=0.15
Entry Rate - GDP
2
Number of Destinations per Exporter - GDP
MWILAO
YEM
TZA
.5
BEL
CMR
UGA
IRN
1.5
KHM
MLI BFA
SWE
.5
PER
TUR
ZAF
PAK
ESP
COL
EGY
IRN
.4
BGD
CHL
NOR
LAO
NER
ECU
SEN
ALB
MKD JOR
KEN
ESP
PER
CHL
NOR
BGR
NIC
MEX
KHM
MEX
MAR
SLV
MUS
.3
MWI
MUS
BFA
MLI MKD
NIC
CRI
LBN
NZL
CMR
EST
KEN
GTM BGR MAR
SLVTZA
UGA
YEM
ECU
DOM
Entry Rate
1
SEN JOR
DOM
EST
BWA
TUR
COL
BEL
GTM
CRI
NZL
BGD
ALB
BWA
SWE ZAF
PAK
EGY
0
.2
BRA
23
24
25
26
27
23
28
24
25
Exit Rate - GDP
Survival Rate of Entrants - GDPpc
.6
BGD
KHMPAK
BRA TUR
.4
EST
BWA
UGA
SEN
MKD
NICALB
JOR
MUS
KHM SLV
.3
BGR
ECU
NZL
CRI
.5
BFA
25
26
SEN
ESP
MEX
TUR
NZL
BEL
CHL
KEN
COL
BEL
SWE
EGY
PAK
ZAF
TZA
EST
UGA
ESP
MWI
CMR
BRA
.2
BGD
24
ZAF
CRI
ALB
MKD
SLVPER
MAR
MUS
GTM
COL
ECU IRN
DOM
BWA
MEX
NIC
MLI
NOR
PER
CHL
MAR
GTM
EGY
JOR
LAO
.4
BFA
IRN
.3
CMR
TZA
KEN DOM
.2
Exit Rate
.5
Survival Rate of Entrants
YEM
23
28
R-squared=0.25
MWI
LAO
MLI
27
Ln GDP
Ln GDP
R-squared=0.25
.6
26
27
28
6
Ln GDP
7
8
9
Ln GDPpc
R-squared=0.14
R-squared=0.00
31
10
11
Turnover Rate - GDP
1.2
Net Entry Rate - GDP
.1
LAO
MWI
UGA
YEM
CHL
PER
BEL
ZAF
PAK
NOR SWE
COL
GTM
SLV
CMR
0
DOM
MUS
YEM
NZL
MAR
EGY
BGR
KEN
IRN
TZA
CMR
LAO
TUR
ESP
MLI
MEXBRA
IRN
BFA
DOM
EST UGA
BWA
KEN
ECU
BGR
SEN
MKD
ALB
JOR
NIC
MUS
.6
CRI
ECU
.8
TZA
Turnover Rate
.05
MLI MKD
EST
BFASENKHM
NIC BWA
-.05
Net Entry Rate
1
BGD
JOR
ALB
KHM
ESP
NOR
PER
CHL
MEX
MAR
SLV
GTM
CRI
NZL
BGD
COL
BEL
SWE ZAF
PAK
EGY
TUR
BRA
.4
-.1
MWI
23
24
25
26
27
28
23
Ln GDP
24
25
26
27
28
Ln GDP
R-squared=0.06
R-squared=0.20
Notes: The measures plotted are based on measures in the country-year level dataset averaged across the 2006-2008 period for
each country. GDP is measured in 2005 PPP U.S. dollars. Kuwait and Portugal are excluded from these plots due to lack of data
for the 2006-2008 period.
32
Table 1: Measures by Country (2006-2008 Averages)
Total Exports Number of
(bn USD)
Exporters
ALB
BEL
BFA
BGD
BGR
BRA
BWA
CHL
CMR
COL
CRI
DOM
ECU
EGY
ESP
EST
GTM
IRN
JOR
KEN
KHM
KWT
LAO
LBN
MAR
MEX
MKD
MLI
MUS
MWI
NER
NIC
NOR
NZL
PAK
PER
PRT
SEN
SLV
SWE
TUR
TZA
UGA
YEM
ZAF
Albania
Belgium
Burkina Faso
Bangladesh
Bulgaria
Brazil
Botswana
Chile
Cameroon
Colombia
Costa Rica
Dominican Republic
Ecuador
Egypt
Spain
Estonia
Guatemala
Iran
Jordan
Kenya
Cambodia
Kuwait *
Laos
Lebanon
Morocco
Mexico
Macedonia
Mali
Mauritius
Malawi
Niger
Nicaragua
Norway
New Zealand
Pakistan
Peru
Portugal *
Senegal
El Salvador
Sweden
Turkey
Tanzania
Uganda
Yemen
South Africa
Average - developing countries
Median - developing countries
Average - all countries
Median - all countries
Number of
Exporters
per Capita
Mean
Exports per
Exporter
('000s USD)
Median
Number of Number of
Exports per Share of Top
Products per Destinations
Exporter 5% Exporters
Exporter per Exporter
('000s USD)
1.1
309.1
0.5
12.4
12.9
165.4
4.6
60.9
1.7
19.1
8.7
4.5
5.7
14.3
229.9
9.3
6.3
12.8
3.4
4.0
3.4
3.0
0.6
3.4
15.3
226.3
2.2
0.8
2.6
0.6
0.3
1.3
39.1
24.6
16.8
25.2
33.5
0.9
4.2
129.5
98.7
2.3
1.2
0.4
58.8
1,895
23,204
425
6,356
13,804
19,375
1,715
7,314
938
9,768
2,931
2,709
3,110
8,370
89,798
4,915
4,420
13,770
1,869
5,057
595
3,315
462
5,177
5,429
34,382
2,926
305
2,251
631
160
1,236
18,309
13,276
15,023
6,732
16,217
727
2,554
30,126
44,570
1,899
938
492
21,721
0.60
2.18
0.03
0.05
1.79
0.10
0.89
0.44
0.05
0.22
0.66
0.28
0.22
0.11
2.00
3.66
0.33
0.19
0.33
0.14
0.04
1.23
0.08
1.24
0.18
0.31
1.43
0.02
1.79
0.05
0.01
0.22
3.93
3.14
0.09
0.24
1.44
0.06
0.42
3.32
0.64
0.05
0.03
0.02
0.45
550
13,312
1,177
1,946
934
8,539
2,666
8,317
1,879
1,957
2,970
1,708
1,830
1,717
2,559
1,885
1,421
940
1,804
796
5,706
915
1,284
659
2,811
6,588
751
2,729
1,138
1,077
2,160
1,031
2,137
1,853
1,116
3,740
2,064
1,228
1,648
4,299
2,204
1,180
1,289
779
2,699
35
64
37
277
22
233
2
49
19
58
54
26
25
65
21
109
38
88
57
18
546
27
42
38
90
44
24
48
17
8
18
27
14
24
62
37
68
73
30
17
105
17
15
49
29
63%
84%
85%
50%
83%
82%
99%
94%
82%
81%
82%
85%
80%
79%
86%
69%
78%
72%
83%
81%
44%
86%
88%
78%
74%
91%
83%
93%
87%
91%
89%
76%
93%
90%
73%
92%
77%
71%
82%
92%
80%
86%
77%
64%
92%
21.7
4.2
35.1
5.7
7,017
2,931
10,027
4,420
0.49
0.22
0.77
0.28
2,206
1,708
2,489
1,830
63
37
61
37
81%
82%
81%
82%
3.0
9.3
3.8
4.2
6.2
1.5
6.8
2.4
3.8
2.4
6.6
4.5
4.0
4.9
5.6
4.7
4.4
4.7
7.8
7.8
6.0
2.7
7.2
8.3
4.4
2.3
7.7
6.4
6.7
4.5
3.8
9.0
4.2
3.9
5.9
5.2
7.5
5.5
7.2
8.5
6.2
6.9
6.5
9.6
4.1
3.8
4.4
15.0
1.4
3.4
2.8
2.8
3.2
2.3
2.4
2.7
4.0
2.7
2.5
2.1
3.1
2.5
4.8
2.0
1.6
3.1
2.5
2.1
2.2
2.2
2.6
1.9
1.6
2.1
3.4
3.1
3.3
2.6
3.5
3.2
2.4
4.3
3.9
2.4
2.4
2.4
3.6
5.7
5.2
5.9
5.6
2.6
2.5
2.8
2.6
Entrant
Survival
Rate
Entry
Rate
Exit
Rate
39%
31%
44%
28%
38%
22%
42%
38%
48%
32%
29%
44%
41%
25%
39%
44%
31%
47%
38%
40%
33%
53%
52%
33%
28%
41%
22%
40%
23%
40%
35%
46%
31%
26%
43%
37%
27%
38%
41%
29%
51%
32%
44%
30%
53%
40%
47%
40%
42%
61%
33%
35%
38%
43%
30%
52%
34%
36%
35%
39%
31%
61%
43%
39%
45%
45%
43%
25%
36%
38%
29%
28%
39%
30%
40%
31%
29%
32%
51%
47%
52%
28%
34%
37%
29%
27%
35%
29%
37%
30%
28%
29%
46%
38%
54%
26%
47%
38%
38%
38%
38%
37%
35%
36%
35%
43%
43%
43%
43%
54%
39%
35%
23%
42%
48%
40%
41%
51%
30%
30%
42%
41%
49%
35%
57%
50%
42%
56%
44%
45%
40%
44%
55%
32%
29%
49%
Notes: The figures shown are based on measures in the country-year level dataset averaged across the 2006-2008 period for each
country. * indicates exceptions to the sample period: for Kuwait averages are taken across the 2009-2010 period and for Portugal
averages are taken across the 2003-2005 period. Total exports are obtained as the number of exporters multiplied by the average
exports per exporter.
33
Table 2: Measures by Sector (2006-2008 Averages)
Number of
Exporters
Mean
Exports per
Exporter
('000s USD)
01-05
108
1,109
107
53%
1.9
06-15
168
631
40
58%
1.6
147
199
201
989
1,427
9,730
1,205
408
147
294
50
7
63%
76%
73%
77%
476
1,221
55
153
413
30
HS 2-Digit Codes HS Section Description
Live Animals and Animal Products
Vegetable Products (including Animal and
Vegetable Fats)
16-24
Foodstuff (Beverages, Spirits, Vinegar, Tobacco etc.)
25-26
Mineral Products (except hydrocarbons)
28-38
Chemicals and Parachemical Products
39-40
Plastics and Articles Thereof
Wood and Articles Thereof (including Paper &
44-46, 47-49, 94
Articles, Furniture)
50-59, 41
Textiles (Including Raw Skins and Leather)
Apparel (Including Footwear, Headgear, Art. of
60-63, 64-67, 42-43
Feathers, Fur, Leather Products)
68-70
Glass, Ceramics and Articles of Stone, Cement, etc.
Precious Metals (Pearls, Jewellery, Coin, Precious
71
Stones etc.)
72-83
Base Metal and Articles Thereof
Mechanical Machinery (including Clocks and Music
84, 91-92
Instruments)
Electrical Machinery (including Optical, Medical,
85, 90
Photographic Instruments)
86-89
Transportation Vehicles
93
Arms and Ammunitions
Median
Number of Number of
Exports per Share of Top
Products per Destinations
Exporter 5% Exporters
Exporter per Exporter
('000s USD)
Entry
Rate
Exit
Rate
Entrant
Survival
Rate
1.9
49%
48%
38%
1.8
50%
48%
36%
1.7
1.3
1.6
1.9
2.3
1.6
2.0
1.8
46%
54%
53%
54%
43%
49%
51%
51%
38%
36%
31%
33%
73%
1.5
1.7
57%
56%
28%
66%
1.5
1.8
58%
57%
27%
353
402
35
69%
1.9
1.7
58%
57%
28%
429
212
4
76%
1.5
1.7
60%
58%
26%
227
7,503
473
78%
1.4
1.6
50%
47%
33%
327
2,010
26
75%
1.5
1.6
60%
58%
26%
870
252
8
72%
1.9
1.6
62%
61%
24%
1,183
737
8
76%
2.4
1.8
57%
54%
30%
326
25
1,275
523
41
39
71%
69%
1.4
1.4
1.6
1.7
65%
59%
63%
61%
24%
19%
Notes: The figures shown in the table for each group of sectors are based on measures at the country-year-HS2-digit level
averaged across HS2-digit codes within the groups listed in the first column, countries (Brazil, Kuwait, New Zealand, and
Portugal excluded), and the 2006-2008 period. Brazil, Kuwait, New Zealand, and Portugal are excluded from these averages due
to lack of measures at the country-year-HS 2-digit level (Brazil and New Zealand) or to lack of data for the 2006-2008 period
(Kuwait and Portugal).
34
Table 3: ANOVA Decompositions at Country and HS 2-digit Sector and
Country and Destination Level
Panel A: Country-HS2 Data
Ln Total
Exports
Ln Mean
Ln Number
Ln Number
Ln Number Exports per Share of Top
of
of Products
of Exporters Exporter 5% Exporters
Destinations
per Exporter
(mn USD)
per Exporter
Entry
Rate
Exit
Rate
Entrant
Survival
Rate
Net Entry
Rate
Turnover
Rate
Country
47%
56%
23%
26%
17%
42%
28%
27%
14%
11%
27%
HS 2-digit Sector
29%
32%
30%
22%
52%
13%
17%
18%
16%
3%
18%
Residual
29%
15%
48%
57%
34%
46%
56%
55%
70%
85%
56%
Observations
3,621
3,750
3,621
2,584
3,530
3,621
3,571
3,584
3,167
3,504
3,504
Panel B: Country-Destination Data
Ln Total
Exports
Ln Mean
Ln Number
Ln Number Exports per Share of Top
of Products
of Exporters Exporter 5% Exporters
per Exporter
(mn USD)
Entry
Rate
Exit
Rate
Entrant
Survival
Rate
Net Entry
Rate
Turnover
Rate
Country
29%
38%
14%
21%
23%
17%
16%
8%
4%
11%
Destination
40%
36%
30%
21%
10%
20%
18%
13%
8%
13%
Residual
43%
36%
61%
65%
68%
66%
68%
80%
89%
77%
Observations
5,558
6,636
5,558
2,887
5,395
6,289
6,236
5,406
5,497
5,497
Notes: The variance decompositions shown are based on measures at the country-year-HS2-digit level and country-yeardestination level data averaged across the 2006-2008 period for each country. The upper panel uses the country sector level data
and the lower panel uses the country destination level data. Brazil, Kuwait, New Zealand, and Portugal are excluded from these
ANOVA decompositions due to lack of measures at the country-year-HS2-digit level (Brazil and New Zealand) or to lack of data
for the 2006-2008 period (Kuwait and Portugal). The sums add up to slightly more than 100 percent in some cases because the
data form an unbalanced panel. Results are similar if we use a balanced panel.
35
Table 4: Export Base, Exporter Size, Concentration and Country Characteristics
Cross-Section, Averaged Data
Ln Total
Exports
Ln GDPpc
Ln GDP
Sector Fixed
Effects
Destination Fixed
Effects
Observations
R-squared
Basic Regressions
Ln Mean
Share of Top
Ln Number Exports per
5%
of Exporters Exporter
Exporters
(mn USD)
Ln Total
Exports
HS 2-digit Regressions
Ln Mean
Share of Top
Ln Number Exports per
5%
of Exporters Exporter
Exporters
(mn USD)
Ln Total
Exports
Destination Regressions
Ln Mean
Share of Top
Ln Number Exports per
5%
of Exporters Exporter
Exporters
(mn USD)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
0.598***
(0.093)
0.868***
(0.068)
0.516***
(0.086)
0.668***
(0.061)
0.077
(0.097)
0.200**
(0.077)
0.034*
(0.020)
-0.008
(0.013)
0.845***
(0.166)
1.115***
(0.103)
0.527***
(0.112)
0.709***
(0.073)
0.359***
(0.111)
0.436***
(0.083)
0.045***
(0.012)
0.037***
(0.008)
0.516***
(0.126)
0.902***
(0.116)
0.408***
(0.110)
0.719***
(0.111)
0.136**
(0.063)
0.274***
(0.046)
0.056***
(0.013)
0.024***
(0.007)
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
5,558
0.515
6,636
0.56
5,558
0.331
2,887
0.297
43
0.889
43
0.889
43
0.249
43
0.093
3,621
0.647
3,750
0.754
3,621
0.449
2,584
0.345
Notes: Robust standard errors in parentheses, clustered at the country level for all HS2-digit and destination regressions. ***, **, and * indicate significance at 1%, 5%, and 10%
confidence levels, respectively. The regressions in columns (1)-(4) are based on measures at the country-year level averaged across the 2006-2008 period for each country; those in
columns (5)-(8) are based on measures at the country-HS2-digit sector-year level averaged across the 2006-2008 period for each country-sector (Brazil and New Zealand are
excluded due to the lack of measures at the country-year-HS2-digit level); and those in columns (9)-(12) are based on measures at the country-destination-year level averaged
across the 2006-2008 period for each country-destination (Brazil and New Zealand are excluded due to the lack of measures at the country-destination level). Kuwait and Portugal
are excluded from all regressions due to lack of data for the 2006-2008 period.
36
Table 5: Exporter Diversification and Country Characteristics
Cross-Section, Averaged Data
Ln GDPpc
Ln GDP
Basic Regressions
HS 2-digit Regressions
Ln Number
Ln Number of
of
Destinations
Products per
per Exporter
Exporter
Ln Number
Ln Number of
of
Destinations
Products per
per Exporter
Exporter
HS 2-digit Regressions
Share of
Share of Total
Exporters
Exports
Ln Number
Accounted for Accounted for
of
by Multiby MultiProducts per
Product Multi- Product MultiExporter
Destination
Destination
Firms
Firms
(1)
(2)
(3)
(4)
(5)
(6)
(7)
0.103**
(0.044)
0.055
(0.048)
0.038
(0.043)
0.094**
(0.040)
0.091***
(0.018)
0.008
(0.013)
0.122***
(0.035)
0.046*
(0.026)
0.116***
(0.040)
0.020
(0.034)
0.006**
(0.002)
0.000
(0.001)
0.040***
(0.010)
0.019*
(0.010)
Yes
Yes
Yes
Yes
3,530
0.587
3,621
0.314
3,018
0.191
3,018
0.263
Sector Fixed Effects
Destination Fixed
Effects
Observations
R-squared
Destination
Yes
41
0.225
42
0.267
5,395
0.155
Notes: Robust standard errors in parentheses, clustered at the country level for all HS2-digit and destination regressions. ***, **,
and * indicate significance at 1%, 5%, and 10% confidence levels, respectively. Products are defined at the HS6-digit level.
Egypt is excluded from all the number of products per exporter regressions since its exporter-level raw data included products at
the HS4-digit rather than the HS6-digit level. The regressions in columns (1)-(2) are based on measures at the country-year-level
averaged across the 2006-2008 period for each country; those in columns (3)-(4) are based on measures at the country-HS2-digit
sector-year level averaged across the 2006-2008 period for each country-sector (Brazil and New Zealand are excluded due to the
lack of measures at the country-year-HS2-digit level); and that in column (5) is based on measures at the country-destination-year
level averaged across the 2006-2008 period for each country-destination (Brazil and New Zealand are excluded due to the lack of
measures at the country-destination level). Kuwait and Portugal are excluded from all regressions due to lack of data for the
2006-2008 period. Only the developing countries for which we have access to the raw exporter-level data (the list is provided in
Table 6) are included in the regressions in columns (6) and (7).
37
Table 6: Share of Single-Product Single-Destination and Multi-Product Multi-Destination
Exporters
Share of Exporters Accounted for by:
Single-Product
Single-Destination
Firms
ALB
BFA
BGD
BGR
BWA
CHL
CMR
COL
CRI
DOM
ECU
GTM
IRN
JOR
KEN
KHM
LBN
MAR
MEX
MKD
MLI
MUS
MWI
NER
NIC
PAK
PER
SEN
SLV
TZA
UGA
YEM
ZAF
Albania
Burkina Faso
Bangladesh
Bulgaria
Botswana
Chile
Cameroon
Colombia
Costa Rica
Dominican Republic
Ecuador
Guatemala
Iran
Jordan
Kenya
Cambodia
Lebanon
Morocco
Mexico
Macedonia
Mali
Mauritius
Malawi
Niger
Nicaragua
Pakistan
Peru
Senegal
El Salvador
Tanzania
Uganda
Yemen
South Africa
Average
Median
45.4%
41.2%
26.5%
37.2%
38.9%
38.5%
39.1%
33.1%
27.5%
37.9%
37.5%
27.7%
34.3%
39.1%
35.3%
25.5%
31.3%
28.2%
40.1%
35.2%
35.8%
26.6%
42.5%
41.3%
34.2%
25.2%
29.8%
35.5%
30.1%
42.8%
42.9%
38.6%
25.0%
34.8%
35.5%
Share of Total Exports Accounted for by:
Firms Exporting
Firms Exporting
Single-Product
More than 4
More than 4
Single-Destination
Products to More
Products to More
Firms
than 4 Destinations
than 4 Destinations
3.4%
8.4%
13.4%
13.4%
2.8%
66.5%
22.4%
2.0%
63.9%
11.2%
1.4%
74.8%
3.5%
0.4%
53.0%
14.4%
0.8%
75.1%
9.3%
3.6%
55.2%
12.9%
3.0%
60.1%
18.3%
0.7%
79.4%
10.4%
1.6%
60.3%
8.8%
4.6%
54.6%
12.8%
1.5%
56.7%
6.9%
5.7%
47.5%
13.8%
2.4%
52.7%
12.8%
2.6%
58.7%
29.4%
1.1%
78.7%
19.9%
2.6%
70.8%
12.6%
3.0%
51.9%
9.1%
1.2%
63.4%
11.1%
1.7%
76.2%
11.3%
2.4%
60.9%
15.1%
3.4%
62.7%
5.6%
1.6%
57.9%
8.1%
0.5%
29.9%
8.6%
4.2%
53.4%
18.6%
2.2%
68.9%
11.9%
3.8%
73.0%
19.2%
3.6%
64.0%
14.3%
3.6%
72.5%
10.2%
3.1%
75.1%
9.3%
3.7%
38.5%
11.1%
6.8%
54.9%
21.5%
1.3%
82.3%
12.8%
11.9%
2.8%
2.6%
60.8%
60.9%
Notes: The shares are based on several cells in matrix tables provided along with the Exporter Dynamics Database for all
countries averaged across the 2006-2008 period. The information is available only for the developing countries for which we
have access to the raw exporter-level data.
38
Table 7: Exporter Dynamics and Country Characteristics
Cross-Section, Averaged Data
Entry
Rate
Ln GDPpc
Ln GDP
Exit
Rate
Basic Regressions
HS2-digit Regressions
Entrant
Survival
Rate
Entrant
Survival
Rate
Net Entry
Rate
Turnover
Rate
Exit
Rate
Net Entry
Rate
Turnover
Rate
Entry
Rate
Exit
Rate
Entrant
Survival
Rate
Net Entry
Rate
Turnover
Rate
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)
-0.016
(0.011)
-0.020**
(0.010)
-0.016
(0.014)
-0.014
(0.010)
-0.010
(0.018)
0.016
(0.011)
0.000
(0.007)
-0.006
(0.004)
-0.032
(0.025)
-0.035*
(0.020)
-0.065***
(0.012)
-0.007
(0.011)
-0.070***
(0.013)
0.001
(0.011)
0.042***
(0.010)
-0.007
(0.007)
0.002
(0.004)
-0.006*
(0.003)
-0.125***
(0.025)
0.004
(0.022)
-0.031***
(0.010)
-0.038***
(0.010)
-0.039***
(0.012)
-0.035***
(0.011)
0.017**
(0.007)
0.020***
(0.007)
0.007
(0.006)
-0.001
(0.003)
-0.061***
(0.018)
-0.027
(0.016)
Yes
Yes
Yes
Yes
Yes
3,571
0.314
3,584
0.312
3,167
0.21
3,504
0.043
3,504
0.296
Sector Fixed Effects
Destination Fixed
Effects
Observations
R-squared
Entry
Rate
Destination Regressions
41
0.29
41
0.17
37
0.057
41
0.058
41
0.235
Yes
Yes
Yes
Yes
Yes
6,289
0.26
6,236
0.247
5,406
0.152
5,497
0.08
5,497
0.162
Notes: Robust standard errors in parentheses, clustered at the country level for all HS2-digit and destination regressions. ***, **, and * indicate significance at 1%, 5%, and 10%
confidence levels, respectively. The regressions in columns (1)-(5) are based on measures at the country-year level averaged across the 2006-2008 period for each country; those in
columns (6)-(10) are based on measures at the country-HS2-digit sector-year level averaged across the 2006-2008 period for each country-sector (Brazil and New Zealand are
excluded due to lack of measures at the country-year-HS2-digit level); and those in columns (11)-(15) are based on measures at the country-destination-year level averaged across
the 2006-2008 period for each country-destination (Brazil and New Zealand are excluded due to the lack of measures at the country-destination level). Bulgaria, Norway, Sweden,
and Yemen are excluded from the regressions in columns (3), (8), and (13) since their shorter sample period does not allow for the calculation of any entrant survival rates between
2006 and 2008 while Kuwait and Portugal are excluded from regressions due to lack of data for the 2006-2008 period.
39
APPENDIX FOR ONLINE PUBLICATION
Appendix 1: Measures in the Database
LEVEL
CODE
MEASURES
A
Country – Year –
Product
Country
- Year
Country
- Year Destinati
on
HS2 &
HS4
HS6
BASIC CHARACTERISTICS
1,2,3,4,5
N (Exporters, Entrants, Exiters, Survivors, Incumbents)




6,7,8,9,10
MMS & Q1Q3 (TEV per Exporter, Entrant, Exiter, Survivor, Incumbent)




11
MMS & Q1Q3 (Growth of Incumbents)t= MMS (ln(TEV of incumbentt in t) –
ln(TEV of incumbentt in t-1))




12
MMS & Q1Q3 (Growth of Survivors)t = MMS (ln(TEV of survivort in t+1) ln(TEV of survivort in t))




B
CONCENTRATION/DIVERSIFICATION
1
Herfindahl Index




2
Share of top 1%, 5%, 25% Exporters in TEV




3
MMS (N. HS6 Products per Exporter)



NA
4
MMS (N. Destinations per Exporter)

NA


5
MMS (N. Exporters per HS6 Product)



NA
6
MMS (N. Exporters per Destination)

NA


C
FIRM DYNAMICS
1
Firm Entry Ratet = N. Entrantst / N. Exporterst




2
Firm Exit Ratet = N. Exiterst/ N. Exporterst-1




3
Firm Survival Ratet = N. Survivorst / N. Entrantst




4
Share of Entrantst= TEV of Entrantst in t / TEV in t




5
Firm 2-year Survival Ratet = N. 2-year Survivorst / N. Entrantst




6
Firm 3-year Survival Ratet = N. 3-year Survivorst / N. Entrantst




D
PRODUCT DYNAMICS
1
MMS (Product Entry Rate of Incumbents)t = MMS (N. products not exported
in t-1 but exported in t by incumbentt / N. all products exported by
incumbentt in t)



NA
2
MMS (Product Entry Rate of Survivors)t = MMS (N. products not exported in
t-1 but exported in t by survivort-1 / N. all products exported by survivort-1 in t)



NA
3
MMS (Share of New Products in TEV of Incumbents)t= MMS (EV of
incumbentt from products not exported in t-1 but exported in t / TEV of
incumbentt in t)



NA
40
4
MMS (Share of New Products in TEV of Survivors)t= MMS (EV of survivort-1
from products not exported in t-1 but exported in t / TEV of survivort-1 in t)



NA
5
MMS (Product Exit Rate of Incumbents)t = MMS (N. products exported by
incumbentt in t-1 but not in t/ N. all products exported by incumbentt in t-1)



NA
6
MMS (Product Survival Rate of 2-year Incumbents)t = MMS (N. products not
exported in t-1 but exported in both t and t+1 by 2-year incumbentt /N. all
products not exported in t-1 but exported in t by 2-year incumbentt)



NA
E
DESTINATION DYNAMICS
1
MMS (Destination Entry Rate of Incumbents)t = MMS (N. destinations not
exported in t-1 but exported in t by Incumbentt / N. all destinations exported
by Incumbentt in t)

NA


2
MMS (Destination Entry Rate of Survivors)t = MMS (N. destinations not
exported in t-1 but exported in t by Survivort-1 / N. all destinations exported
by Survivort-1 in t)

NA


3
MMS (Share of New Destinations in TEV of Incumbents)t= MMS (EV of
Incumbentt from destinations not exported in t-1 but exported in t / TEV of
Incumbentt in t)

NA


4
MMS (Share of New Destinations in TEV of Survivors)t= MMS (EV of survivort-1
from destinations not exported in t-1 but exported in t / TEV of survivort-1 in
t)

NA


5
MMS (Destination Exit Rate of Incumbents)t = MMS (N. destinations exported
by Incumbentt in t-1 but not in t/ N. all destinations exported by Incumbentt
in t-1)

NA


6
MMS (Destination Survival Rate of 2-year Incumbents)t = MMS (N.
destinations not exported in t-1 but exported in both t and t+1 by 2-year
Incumbentt /N. all destinations not exported in t-1 but exported in t by 2-year
Incumbentt)

NA


NA
NA


F
UNIT PRICES
1,2,3,4,5
MMS (Unit Price (TEV/Quantity) per Exporter, Entrant, Exiter, Incumbent,
Survivor)
Notes: MMS indicate Mean, Median, Standard Deviation; Q1Q3 indicate the 25th and the 75th percentiles; N indicates Number
of; EV indicates Export Value; and TEV indicates Total Export Value.
41
Appendix 2: Constructing the Exporter Dynamics Database Using Customs Data at the
Exporter-Level
The measures included in the Database are computed using customs data from 45 countries at the
exporter-product-destination-year level covering the universe of export transactions. 21 For 11
countries (Brazil, Egypt, Estonia, Laos, New Zealand, Norway, Portugal, Spain, Sweden, and
Turkey), we do not have access to the raw data and the statistics were calculated for us by
researchers or institutions with access to the raw data. Pooling across the raw datasets for the
remaining 34 countries, we obtain a panel with 15 million unique observations at the countryfirm-product-destination-year level which is the raw dataset used to construct the Database.
Some of the variables in this cross-country raw dataset were subjected to a series of cleaning
procedures or had peculiarities that we briefly describe below: firm codes, country of destination,
product and export values.22
Regarding firm codes, exporters are uniquely identified by their actual names, their tax
identification number or artificial unique codes randomly created by our data providers. In the
case of Albania, Burkina Faso, Cambodia, Cameroon, Mexico, and Uganda their firm coding
systems underwent changes in 2007 and in the case of Yemen in 2008. Hence, it is not possible
to calculate exporter dynamics measures in the years when those changes occur.
Regarding the country of destination variable, two cleaning operations are applied. One
relates to the use of special trade regimes by Bulgaria, Costa Rica, Colombia, Ecuador, Egypt,
Jordan, Kuwait, Lebanon, Morocco, Peru, Turkey, Yemen, which implies that customs agencies
record the sales from inland to their own free zones/customs warehouses as exports resulting in a
larger set of potential destinations.23 Since there is a lack of uniformity across countries in the
definition of the special trade regime we drop the observations related to sales to free zones from
our dataset. 24 The other relates to the changes in names that some statistical territories have
undergone over time due to spatial divisions. When territories split, we keep them as a single
destination since disregarding these changes in names would bias some calculations, especially
those related to destination diversification and dynamics. For example, destination entry and exit
rates in a given year would be overestimated as the appearance of a new destination in the
dataset would not represent the true expansion of exporters into a different territory but would
instead be due to the change in territory classification. Likewise, destination measures that are
calculated using only one year of data (e.g., number of destinations per exporter etc.) would not
be comparable to each other. In the end, each country has a potential set of 246 destinations as of
end-2011.
Regarding products, although most customs agencies record export transactions at 8-digit
of the Harmonized System (HS) or higher, we aggregate the information to the HS 6-digit, since
21
The providers of the raw datasets for each country were mostly governmental agencies, mainly customs offices. Appendix 1 in
Cebeci et. al (2012) provides a complete list of the countries included in the Database, the periods for which data is available, and
the data sources.
22
Cebeci et. al. (2012) describe the cleaning procedures in further detail. Note that the cross-country raw dataset at country-firmproduct-destination-year level includes also quantities exported.
23
The “special trade regime” considers transactions where the goods are sold from the domestic territory only to both third
countries and free zones/customs warehouses of the origin country as exports. In contrast, the “General trade regime” considers
transactions where goods are sold from any national territory (including free zones) to third countries only as exports (see p. 32 of
United Nations, 2008).
24
See p. 34 of United Nations (1998). This operation had a minor effect on total export volumes as sales to free zones/customs
warehouses are negligible in most cases. The export volumes of Turkey, the country most affected by this operation, drop by 2-3
percent.
42
this is the most detailed level comparable internationally. That is, a specific HS code at 6 digits
represents the same product in all countries in a given year. Several cleaning operations are
applied to the product variable given that our sample period spans three different HS
classifications (HS1996, HS2002 and HS2007). First, we combine all the codes existing under
the three HS classifications into an aggregated list of 6065 unique HS 6-digit codes and eliminate
the observations with a code that is not in that aggregated list.25 Second, through a process of
“consolidation” among the three HS classifications, we account for the modifications introduced
in each of the HS revisions: a) two different codes with low trade volume were converted into a
single code and b) an existing code with an increasing trade volume was split into various codes.
Ignoring these modifications would create problems for the tracking of trade volumes for certain
products over time.26 The basic principle of consolidation is to identify the HS codes related to
each other (e.g., codes that were split or merged with the modifications introduced by the
HS2002 or the HS2007) and to replace them with a single code for the entire period. As a result
of this consolidation 1104 codes are replaced by 402 codes that already exist in the HS lists but
whose contents are altered, as discussed in Cebeci (2012). Consequently, the 6065 unique
potential HS 6-digit codes in the list that aggregates across 1996, 2002, and 2007 classifications
decline to a final number of 4961 unique potential HS 6-digit codes in the consolidated
classification. 27 In our cross-country raw dataset 4767 of those 4961 HS 6-digit codes are
present. In terms of the array of products included in the cross-country raw dataset, we eliminate
observations in HS Chapter 27 (hydrocarbons such as oil, petroleum, natural gas, coal etc.) given
that we do not have exporter-level data on that chapter for important oil exporting countries such
as Burkina Faso, Cameroon, Iran, Kuwait, and Yemen.28
Regarding export values in the cross-country raw dataset, they are measured in US
Dollars (USD) and they are Free on Board (FOB) figures, except for El Salvador and Senegal,
whose export values represent Cost, Insurance and Freight (CIF) figures. 29 This difference
should be taken into account for cross-country comparisons of measures related for example to
the size of exporters but is not expected to affect other measures related to concentration,
diversification and firm, product and market dynamics.
In order to have a sense of the reliability of the raw data in each of the countries included
in the Database, we apply two filters. For the first filter we compare the total values exported
(excluding HS 27) calculated from the cross-country raw dataset with the total values exported
from the United Nations’ COMTRADE database (excluding HS 27) for every country and
25
The number of HS categories included in the original classifications HS 1996, HS 2002, and HS 2007 are 5209, 5224, and
5053, respectively. Most of this elimination results from eliminating observations with a product code belonging to HS Chapter
99 as this is reserved for national use and the HS 6-digit codes under this chapter differ across countries. This elimination
accounts for about 0.5 percent of total exports in our cross-country raw dataset.
26
See Cebeci (2012) for the methodology used in the consolidation. The paper along with a list of consolidated codes and
concordances are available at http://econ.worldbank.org/exporter-dynamics-database. A similar process was used by Schott and
Pierce (2011) to concord 10-digit United States Harmonized System codes between 1989 and 2007 and by Wagner and Zahler
(2011) to homologate among 6-digit HS1992, HS1996, and HS2002 classifications.
27
The number of HS 6-digit codes not affected by the consolidation process is 4559, obtained as the total number of codes 6065
minus the 1104 codes that disappeared and minus the 402 codes whose content changed.
28
Possible reasons for this lack of data are confidentiality reasons or the fact that goods exported through pipeline are not
recorded at customs but instead are recorded by other government/private institutions in the countries.
29
For countries for which export values were provided in local currency, they were converted to USD using the exchange rate
series PA.NUS.FCRF taken from the World Development Indicators (whose original source is the IMF’s International Financial
Statistics) which is based on an annual average of daily exchange rates determined by national authorities or in the legally
sanctioned exchange market.
43
year.30 This comparison yields quite different results for different countries. On the one hand, for
Albania, Brazil, Cameroon, Chile, Colombia, Costa Rica, Ecuador, Guatemala, Kenya, Lebanon
Mexico, Morocco, Pakistan, Peru, South Africa, Tanzania and Turkey, the ratio of total values
exported in the raw dataset to total values exported in COMTRADE is about 100 percent. While
for countries such as Mali and Yemen, the ratios indicate that total values exported in the raw
dataset are as low as half of the total values exported in COMTRADE, for others such as
Mauritius total values exported in the raw dataset are on average 30 percent above total values
exported in COMTRADE.31 For the second filter, we focus on the countries and years that would
have been left out because of a unfavorable match with COMTRADE (below 60 percent) and we
keep those countries (and years) where we observe internal consistency within the export totals
calculated from the corresponding exporter-level raw dataset over time. 32 As a result of this
second filter, we keep in the raw dataset the information for Macedonia, Jordan, Mali, Mauritius,
and several years of El Salvador’s data which would be left out due to an unfavorable match with
COMTRADE data.33
To protect the confidentiality of the firms in the raw datasets at the exporter-productdestination-year level, some of the measures in the files with product or destination
disaggregation are missing when the underlying country-product-year or country-destinationyear cell includes a single firm whose individual information cannot be revealed.34
For the measures of firm dynamics in the country-year dataset, of firm-product dynamics in
the country-year-product datasets, and of firm-destination market dynamics in the country-yeardestination datasest, we use the following definitions:
 Exportert: any firm that exports in year t;
 Entrantt: a firm that does not export in year t-1 but exports in year t;
 Exitert: a firm that exports in year t-1 but does not export in year t;
 Incumbentt: a firm that exports in both years t-1 and t;
 2-Year Incumbentt: a firm that exports in years t-1, t, and t+1;
 Survivort: a firm that does not export in year t-1 but exports in both years t and t+1;
 2-Year Survivort: a firm that does not export in year t-1 but exports in years t, t+1 and t+2;
 3-Year Survivort: a firm that does not export in year t-1 but exports in years t, t+1, t+2 and
t+3.
30
Given our understanding about which transactions are included in the raw files, for this comparison we consider gross export
figures in COMTRADE (which include re-exports) for Albania, Kenya, Mauritius, Tanzania and Senegal and net export figures
(gross exports minus re-exports) for Jordan, Uganda and Yemen. For all other countries, it does not matter which COMTRADE
figure we choose as they report either insignificant or no re-exports.
31
The Appendix in Cebeci et. al. (2012) provides detailed results on these comparisons.
32
Although both sources of trade data, COMTRADE as well as the exporter-level raw datasets that we use, originate from
customs authorities, we are aware of potential difficulties in the processing of the exporter-level raw datasets that may justify the
differences sometimes observed between the export totals obtained from the exporter-level raw datasets and the export totals
available in COMTRADE. One of the potential reasons for these difficulties relates to the manual registration of export
transactions that still takes place in some countries and may result in under-recording of export transactions for some countries
and years.
33
For countries such as Macedonia for which the match with export totals in COMTRADE in the first few sample years is
relatively poor, the match improves over time and this is likely due to a shift from manual to digital registration of export
transactions.
34
For such cells the number of exporters shows a value of 1 and most measures except those based on exiter firms in Appendix
Table 1 are missing. The developed countries in our sample applied their own confidentiality rules (often stricter) as detailed in
Cebeci et. al (2012). A file with the percentage of total exports corresponding to the hidden values by country and year, for the
developing countries for which we have the raw exporter-level data is available from the authors upon request.
44
Appendix 3: Robustness Specifications
i. Estimations using country-destination data – Only manufacturing
Cross-Section, 2006-2008 Averaged Data
Ln Mean
Ln
Ln Number
Exports Share of
Ln Total Number
of
per
Top 5%
Exports
of
Products per
Exporter Exporters
Exporters
Exporter
(mn USD)
Ln GDPpc
Ln GDP
Destination Fixed
Effects
Observations
R-squared
(1)
(2)
(3)
(4)
0.372**
0.198
0.194** 0.060***
(0.179) (0.141) (0.087) (0.017)
1.041*** 0.611*** 0.467*** 0.021**
(0.163) (0.116) (0.075) (0.009)
Entry
Rate
Exit
Rate
Entrant
Survival
Rate
(5)
0.125**
(0.053)
0.026
(0.037)
(6)
-0.033**
(0.015)
-0.034**
(0.013)
(7)
-0.043**
(0.017)
-0.036**
(0.014)
(8)
0.018
(0.013)
0.019**
(0.008)
(9)
0.006
(0.007)
0.001
(0.004)
(10)
-0.084***
(0.026)
-0.015
(0.023)
Net Entry Turnover
Rate
Rate
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
4,838
0.425
4,838
0.432
4,838
0.289
1,600
0.328
4,838
0.134
4,503
0.204
4,498
0.206
4,090
0.128
3,776
0.089
3,776
0.151
Notes: Robust standard errors in parentheses, clustered at the country level for all destination regressions. ***, **, and * indicate
significance at 1%, 5%, and 10% confidence levels, respectively. The regressions are based on measures at the countrydestination-year level averaged across the 2006-2008 period for each country-destination (Brazil and New Zealand are excluded
due to the lack of measures at the country-destination level). Kuwait and Portugal are excluded from all regressions due to lack of
data for the 2006-2008 period.
45
ii.
Estimations using country-destination data – Controlling for trade costs (distance)
Cross-Section, 2006-2008 Averaged Data
Ln Mean
Ln Number
Ln Number Exports Share of
Ln Total
of
of
per
Top 5%
Exports
Products per
Exporters Exporter Exporters
Exporter
(mn USD)
Ln GDPpc
Ln GDP
Distance
Destination Fixed
Effects
Observations
R-squared
(1)
(2)
(3)
(4)
0.434*** 0.344***
0.113* 0.054***
(0.131)
(0.112)
(0.063)
(0.013)
1.066*** 0.811*** 0.317*** 0.037***
(0.123)
(0.111)
(0.051)
(0.009)
-1.606*** -1.310*** -0.401*** -0.064***
(0.100)
(0.077)
(0.059)
(0.010)
Entry
Rate
Exit
Rate
Entrant
Survival
Rate
Net
Turnover
Entry
Rate
Rate
(5)
0.105**
(0.040)
0.041
(0.037)
-0.201***
(0.025)
(6)
-0.027**
(0.011)
-0.042***
(0.011)
0.070***
(0.007)
(7)
-0.034***
(0.012)
-0.038***
(0.011)
0.066***
(0.008)
(8)
0.015*
(0.008)
0.022***
(0.007)
-0.043***
(0.006)
(9)
(10)
0.007 -0.057***
(0.005) (0.019)
-0.001 -0.033*
(0.003) (0.018)
0.003 0.076***
(0.003) (0.016)
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
5,346
0.655
6,344
0.73
5,346
0.361
2,806
0.363
5,188
0.223
6,012
0.292
5,957
0.274
5,163
0.16
5,295
0.07
5,295
0.17
Notes: Robust standard errors in parentheses, clustered at the country level for all destination regressions. ***, **, and * indicate
significance at 1%, 5%, and 10% confidence levels, respectively. The regressions are based on measures at the countrydestination-year level averaged across the 2006-2008 period for each country-destination (Brazil and New Zealand are excluded
due to the lack of measures at the country-destination level). Kuwait and Portugal are excluded from all regressions due to lack of
data for the 2006-2008 period. Distance is taken from the CEPII database.
46
iii.
Estimations on average exporter size using sample of incumbent exporters
Cross-Section, 2006-2008 Averaged Data
Ln GDPpc
Ln GDP
Basic Regressions
HS 2-digit Regressions
Destination Regressions
Ln Mean Exports per
Continuing Exporter
(mn USD)
Ln Mean Exports per
Continuing Exporter
(mn USD)
Ln Mean Exports per
Continuing Exporter
(mn USD)
(1)
(2)
(3)
0.10
(0.097)
0.162**
(0.076)
0.327***
(0.109)
0.466***
(0.083)
0.121
(0.079)
0.250***
(0.053)
Sector Fixed Effects
Destination Fixed
Effects
Observations
R-squared
Yes
Yes
41
0.221
3316
0.404
4,699
0.276
Notes: Robust standard errors in parentheses, clustered at the country level for all HS2-digit and destination regressions. ***, **,
and * indicate significance at 1%, 5%, and 10% confidence levels, respectively. The regression in column (1) is based on
measures at the country-year-level averaged across the 2006-2008 period for each country; that in column (2) is based on
measures at the country-HS2-digit sector-year level averaged across the 2006-2008 period for each country-sector (Brazil and
New Zealand are excluded due to the lack of measures at the country-year-HS2-digit level); and that in column (3) is based on
measures at the country-destination-year level averaged across the 2006-2008 period for each country-destination (Brazil and
New Zealand are excluded due to the lack of measures at the country-destination level). Kuwait and Portugal are excluded from
regressions due to lack of data for the 2006-2008 period.
47
iv.
Estimations using panel data, 2004-2008.
HS 2-digit Regressions
Panel
Destination Regressions
Panel
HS 2-digit Regressions
Panel
Destination Regressions
Panel
Ln Mean
Ln Mean
Ln Mean
Ln Mean
Ln
Ln
Ln
Ln
Exports
Exports
Exports
Exports
Ln Total Number
Ln Total Number
Ln Total Number
Ln Total Number
per
per
per
per
Exports
of
Exports
of
Exports
of
Exports
of
Exporter
Exporter
Exporter
Exporter
Exporters
Exporters
Exporters
Exporters
(mn USD)
(mn USD)
(mn USD)
(mn USD)
Ln GDPpc
(1)
1.549**
(0.684)
(2)
0.390
(0.534)
(3)
1.170*
(0.644)
(4)
1.262*
(0.728)
(5)
0.207
(0.472)
(6)
1.055*
(0.602)
Ln GDP
Year Fixed Effects
Country Sector
Fixed Effects
CountryDestination Fixed
Effects
Observations
R-squared
Yes
Yes
Yes
Yes
Yes
Yes
14335
0.962
14939
0.984
14335
0.922
Yes
Yes
Yes
Yes
Yes
Yes
1660
0.983
1660
0.994
1660
0.954
(7)
(8)
(9)
(10)
(11)
(12)
2.149***
(0.790)
Yes
0.867
(0.626)
Yes
1.357**
(0.553)
Yes
1.160
(0.749)
Yes
0.366
(0.478)
Yes
0.793
(0.589)
Yes
Yes
Yes
Yes
Yes
Yes
Yes
1660
0.983
1660
0.994
1660
0.954
14335
0.962
14939
0.984
14335
0.922
Note: Robust standard errors in parentheses, clustered at the country level. ***, **, and * indicate significance at 1%, 5%, and 10% confidence levels, respectively. All regressions
cover the 2004-2008 period, those in columns (1)-(3) and (7)-(9) are based on measures at the at the country-HS 2-digit sector-year level; and those in columns (4)-(6) and (10)912) are based on measures at the country-destination-year level. Destination regressions include the top 10 destinations only for each exporting country and year.
48