Mohler, Lukas und Seitz, Michael:
The Gains from Variety in the European Union
Munich Discussion Paper No. 2010-24
Department of Economics
University of Munich
Online at http://epub.ub.uni-muenchen.de/11477/
The Gains from Variety in the European Union
Faculty of Business and Economics, University of Basel, Peter Merian-Weg 6, 4002 Basel, Switzerland
Department of Economics, Ludwig-Maximilians-University of Munich, Ludwigsstrasse 28, 80539 Munich, Germany
Over the last decade, European Union members have experienced a dramatic increase in im-
ports. This increase was accompanied by a strong growth in the number of imported goods and
trading partners, indicating positive welfare gains for consumers via an extended set of consumption
possibilities, as pointed out in the "New Trade Theory". In this paper, we apply the methodology
developed by Feenstra (1994) and Broda and Weinstein (2006) to estimate structurally the gains
from imported variety for the 27 countries of the European Union using highly disaggregated trade
data at the HTS-8 level from Eurostat for the period of 1999 to 2008. Our results show that, within
the European Union, especially “newer” and smaller member states exhibit high gains from newly
imported varieties. Furthermore, we find that the vast majority of the gains from variety for con-
sumers stem from intra-European Union trade.
JEL classification: F12, F14;
Keywords: European Union, Welfare Gains from Trade, Trade in Variety;
∗Tel.: +41 61 267 0770; fax +41 61 267 1316; Email address: Lukas.Mohler@unibas.ch
†Corresponding author. Tel.: +49 89 2180 6286; fax +49 89 2180 6227; Email address: Michael.Seitz@lrz.uni-
The European Union with its 27 member states today constitutes the largest single market in the
world. Over the past decade, several historical events have deepened the economic integration of the
economies within the European Union but also of the member states into the world economy, resulting
in a strong increase in trade flows. First, the euro was introduced as book money in 1999 and today
is the official currency of 16 European Union member states. Second, the transition of the Eastern
European economies from planned economies to market economies after the fall of the Iron Curtain was
accompanied by a surge and redirection of trade flows towards the "old" member states as well as a surge
of trade between Eastern European countries. This transition finally led to the eastern enlargement in
2004, when ten new member states joined the European Union, followed by Romania and Bulgaria
in 2007. Finally, the European Union and its member states were confronted with the integration of
fast-growing emerging markets into the world trading system over the last decade, with China and other
East Asian economies at the forefront. This dynamic process of economic integration on the one hand
caused more competition and new challenges for the European member states and European firms. On
the other hand the integration process was accompanied by a dramatic increase in internal and external
trade flows for all the member states. One important aspect of this phenomenon is the positive effect on
consumer welfare via increased imports and an extended choice set of available varieties for consumers.
From 1999 to 2008 the total value of imports for all countries has more than doubled. At the same time,
the mean number of supplying countries within an average product category has increased by about
15%, while the number of imported product categories has stayed roughly constant. In sum, about 60%
in the increase of total imports can be attributed to the establishment of new trade linkages with new
goods and/or new trading partners, indicating high gains for consumers as a result of newly available
Since the pioneering work of Krugman (1979, 1980, 1981) the "love for variety" motive and its
implications for consumer welfare have been a key element of the "New Trade Theory". Based on a
monopolistic competition model, first outlined by Spence (1976) and Dixit and Stiglitz (1977), where
a single good is available in different varieties, these models predict that trade leads to an increased
number of varieties available for consumers. In combination with a constant elasticity of substitution
(CES) utility function, trade generates positive effects for consumer welfare via the availability of more
varieties of one good. Since then, economists have tried to quantify empirically the gains for consumers
from newly imported varieties. In his seminal contribution, Feenstra (1994) was the first to assess
empirically the impact of new and disappearing varieties for a single imported good. Based on the
theoretical framework of the New Trade Theory, Feenstra (1994) developed an artificial price index
to measure the impact of traded varieties on consumer welfare. The idea is that new varieties lower
the price index, while disappearing varieties increase the index, where the magnitude depends on the
substitutability between the varieties and their expenditure share. Broda and Weinstein (2006) extend
the approach of Feenstra (1994) and construct an aggregate price index using the full set of imported
products to compute the import price index bias resulting from the omission of new and disappearing
varieties. This approach allows them to quantify the overall impact of imported varieties on consumer
welfare for the United States for the period from 1972 to 2001. Using highly disaggregated trade data,
and the assumption that goods are differentiated across countries, they show that the unmeasured
growth in product variety has been an important source of welfare gains.1Their results indicate an
upward bias of the conventional price index of the magnitude of 1.2% per year which translates into
an overall effect of 2.6% of GDP for the overall period or put differently, consumers are willing to pay
roughly 0.1% of their annual income to have access to a larger set of goods and varieties.
In this contribution we adopt the methodology of Broda and Weinstein (2006) to estimate struc-
turally the gains from imported variety for all 27 European Union member states for the period from
1999 to 2008 by exploring a rich dataset of highly disaggregated trade data at the HS-8 level which allows
us to identify over 10,000 different product categories: to our best knowledge this has not been used
before in another study. The effects on consumer welfare of newly available products are in particular
interesting for European economies, since the European Union consists of many small and medium-sized
economies with high import shares and a high degree of political and economic integration within the
European Union as well as in the world economy. In addition, our study of a set of countries allows us
1The definition that goods are differentiated across countries is based on the theoretical framework of Armington
(1969). Here, a variety is simply a particular good produced by a particular country, e.g., French wine.
to analyse and interpret results across countries, adding another dimension to this approach. We follow
Broda and Weinstein (2006), and construct an artificial import price index for each country. In a first
step we estimate a total of approximately 170,000 elasticities of substitution, one for each imported good
of each country. In a second step we use these elasticities to compute a correction term for each product.
This term captures the effect on the price index that is due to the change in the variety set. Based
on the structural assumption of a Krugman (1980) style economy, we finally calculate the gains from
variety for each single member state. Furthermore, we extend the approach of Broda and Weinstein
(2006) which allows us to calculate the import price index for European Union internal and external
imports. This enables us to identify which trade flows contribute most to the gains from variety for
each member state. Using our highly disaggregated trade data, we also provide extensive descriptive
statistics about the number of trading partners and traded products for the European member states,
which allows us to gain new insights into the trade integration process of all the member states. Finally,
we provide some new robustness measures to build further confidence in our results. Our results can
be summarized as follows: for most countries the biases and hence the gains from variety are positive.
However, the results differ across countries and three different groups can be identified. First, for the
largest four economies in the European Union in terms of GDP, the impact of traded variety is small for
the considered period. This can be explained by smaller import shares and the fact that these economies
have already been strongly integrated into the European Union and the world market in 1999, which is
our base year. Secondly, for all the smaller "old" member states we find modestly positive gains from
imported variety. Finally, for the "newer" member states of the European Union, with the exception of
Malta, the gains are strongly positive, mostly larger than 1% of GDP. This result reflects the effects of
the ongoing integration of these countries into the European single market and into the world trading
system as well as their higher growth rates and higher import shares. For example, variety gains for
Estonia sum up to 2.8% of GDP, which is of the same magnitude as Broda and Weinstein (2006) find
for the United States for their much longer period from 1972 to 2001. Our results show that especially
for fast-growing, less-developed and smaller countries, the establishment of new trade linkages and the
importing of new varieties are an important source of welfare gains via trade. We also find that for
most countries about 70% of the gains stem from intra-European trade, emphasizing the positive effects
and importance of European integration. Third, descriptive statistics of the estimated elasticities of
substitution indicate that they do not differ systematically across countries, which is interesting, given
the different sizes of the economies.
Our paper contributes to two strands of the empirical trade literature. First, beside the approaches
of Feenstra (1994) and Broda and Weinstein (2006) several other studies have tried to evaluate the
effects of new varieties on consumer welfare and the role of trade.2
Feenstra (1992). He shows in a numerical example how trade barriers can affect the number of available
products and reduce consumer welfare. Following the idea of Feenstra (1992), Romer (1994) calibrates
a model with fixed export costs and finds that a substantial reduction in trade barriers will lead to more
exported varieties, resulting in an increase of GDP of up to 20%. Using a similar approach Klenow
and Rodríguez-Clare (1997) construct and calibrate a general equilibrium model with detailed Costa
Rican trade data to quantify the impact of trade restrictions on welfare. Their results suggest that
the gains from trade liberalization can be higher compared with traditional models if the effects of
traded variety are taken into account. In an extension, Arkolakis et al. (2008) provide a more detailed
analysis of the Costa Rican trade liberalization and find that the effect of trade liberalization on product
variety is relatively small, since new products are imported in small quantities. Furthermore, Broda
and Weinstein (2004) document the rapid growth in product variety over the last decade in world trade
and point to the important effects on consumer welfare.3 4Although these studies made important first
steps in analysing and understanding the impact of new traded varieties, their methodologies and data
rest on strict assumptions and provide an unprecise measure compared with the more sophisticated
approach of Broda and Weinstein (2006). Therefore, more recent papers rely on the methodology first
proposed by Feenstra (1994) and extended by Broda and Weinstein (2006). Using detailed market data
A first attempt was made by
2For a more microeconomic perspective on the effects of new varieties on consumer gains see Hausman (1981), Hausman
(1994), and Trajtenberg (1989).
3For a theoretical explanation of the increase in traded varieties also see Yi (2003), Melitz (2003), and Bernard et al.
4The increase of product variety over the last decades can also be observed at the national level. Bils and Klenow
(2001) find a strong increase in 106 product categories for the U.S especially over the last 20 years.
about the U.S. automobile market Blonigen and Soderbery (2009) are able to estimate a more precise
measure of welfare gains and show that the Armington assumption, that goods are only differentiated
across countries, hides significant welfare gains. They find that the estimated impact of new net varieties
on consumer welfare is doubled in magnitude when compared with conventional import data. Using
similar methodologies and data another strand of literature emphasizes the positive effect of increased
import variety on productivity, including Feenstra and Markusen (1994), Broda et al. (2006), and
Feenstra and Kee (2008). Considering the European case, Funke and Ruhwedel (2005) provide an
empirical analysis of disaggregated trade data on export variety and economic growth in the Eastern
European countries. Their analysis shows a high correlation between increased imported variety and
economic growth. Second, while the European integration process has received substantial interest in
the literature, the analysis of European Union trade flows and their positive effects on consumer has
been scarce. Smith and Venables (1988) made an important first step in evaluating the positive effects
of a single market in the European Union. Based on a Krugman (1979) model, they show in a numerical
experiment that the creation of a single market has a strong pro-competitive effect at the industry level,
generating large welfare gains for its member states of up to 4% of GDP. In the European trade literature
three prominent lines can be identified. First, several studies have tried to quantify the positive effect
of the introduction of the euro on trade: see Baldwin (2006) for a survey. Second, researchers have
studied the effect of European integration and the role of national borders in intra-European trade
flows, including Nitsch (2000) and Chen (2004). Finally, Buch and Piazolo (2001) and Manchin and
Pinna (2009) study the implications of the Eastern European enlargement in 2004 on the growth and
redirection of trade flows towards the European Union. All these studies rather build on aggregated
trade data and evaluate the extent to which the composition and volume of trade flows are changing,
but do not analyse highly disaggregated trade data which is also a novelty of our approach. Finally,
all these papers neglect the potential positive effect on consumer welfare, which is at the heart of our
The rest of the paper is organized in the following sections. In section 2 we describe the dataset and
provide detailed descriptive statistics of disaggregated trade data and the number of imported varieties
for all the member states. Section 3 briefly reviews the methodology developed by Feenstra (1994) and
Broda and Weinstein (2006) to account for variety changes in price indices before turning to section 4
which presents the results and interpretation for each member state. Section 5 concludes.
2 Descriptive Statistics
To quantify empirically the positive effects on consumer welfare from the availability of new imported
products, our analysis requires highly detailed product information on the quantity and prices of im-
ported goods. Therefore our empirical analysis uses the database from Eurostat which consists of highly
disaggregated trade data defined at the HS-8 level for all EU-27 member states for the period from 1999
to 2008. In this data set about 10,000 products are classified, for which data on the value and tons
of imports are available, which allows us to calculate unit prices for each product. From this database
we collect information on the imports for each single member state from all the trading partners in the
world. We use quarterly data from the first quarter of 1999 to the first quarter of 2008 to rule out
potential seasonality effects. The richness of this data set allows us to provide a precise estimate of the
gains from variety, which is robust to several extensions as discussed in the following sections. Table
1 provides some detailed information on the development of the product categories available from this
data set over time. For the period from 1999 to 2008 the total number of classified products amount
to 13,234. At the same time the number of classified products decreased from 10,428 to 9,699. In
total, 2,950 new products were introduced and 3,838 were deleted, resulting in a total net change of
6,788 classified, products for the overall period. It is also obvious that for each year the numbers new
and excluded products are roughly equal, while the absolute value of reclassification varies over time.
Especially for the years 2002, 2006, and 2007 large product category changes can be observed.5
5This is due to reclassifications that appear regularly. The empirical approach presented below is robust towards such
reclassifications of products, as explained by Feenstra (1991).
Table 1: Summary statistics on product codes at the HS-8 level for the period from 1999 to 2008
Year Total number
of product codes
Source: Eurostat International trade statistics - methodology and classifications 2009.
2.1 Analysis of Aggregated Trade Flows
For our descriptive analysis we start with aggregated trade data at the European level before turning to
the more disaggregated country level. The dynamic economic integration process over the last decade,
within the European Union as well as at the global level, has led to a dramatic increase in imports
for all the member states. From row one in column 1 and 2 in Table 2 we can see that aggregated
imports of the European Union member states from all the trading partners have more than doubled
from 487 billion euros in the first quarter of 1999 to 979 billion euros in the first quarter of 2008. Since
one focus of our study is trade within the European Union we split up imports by European Union,
internal imports and imports from the rest of the world (external imports) in row two and three. We
can see that the European Union internal trade flows account for approximately 60% of all the imports,
underlining the importance of internal European Union trade. Both internal and external import flows
have grown at rapid rates and roughly doubled in this period, while external trade flows have grown
at a slightly higher rate, resulting in a higher import share of non-European Union imports in 2008,
accentuating growing trade with emerging markets like China at the forefront. This strong increase
in imports was accompanied by another phenomenon, a strong increase in the number of imported
varieties. In our analysis a good is defined as an HS-8 product category. Following Armington (1969),
a variety is then assumed to be a particular good from a particular country. Based on this definition
we find a strong increase from 1.67 million to 1.97 million imported varieties during the same period.
Similarly to the value of imports, about two-thirds of new varieties stem from internal imports. This is
interesting, given the relatively small number of potential trading partners for a single product within
the European Union, and highlights the importance of intra-European Union trade as a source of new
product varieties for consumers.
Given the diverse structure of the European economies and their differences in terms of size (GDP),
growth of GDP, absolute imports, and import shares, countries have been affected differently by the
integration process. Since in New Trade Theory the absolute size and growth rate of a country are
an important determinant of the structure and development of import flows, three different types of
countries in the European Union can be identified:6first, the "big four" economies including Germany,
Italy, France, and the United Kingdom with the largest GDP in the European Union; second, the eleven
"small old" high income member states; third the twelve fast-growing less-developed "new" member
states mostly, from Eastern Europe, which joined the European Union in 2004 and 2007, respectively.
In columns three to six in table 2 we show descriptive statistics of aggregated imports for each of these
three blocks. For all the blocks we can see a strong growth in import volume from both European
Union and non-European Union countries, which is highest for the "new member" states. The "big
four" account for roughly half of all the imports but only for a quarter of the imported varieties.
Despite the strong growth in imports, the number of imported varieties only slightly increased from
6In the New Trade Theory, the absolute value of trade between two economies depends on the size of both economies.
456,744 to 496,597, indicating that imports have grown at the intensive margin.7While 65% of the
total imports stem from the EU-27 member states, they only account for 58% of the imported varieties
in 1999, but both shares slightly decreased over time, emphasizing the growing importance of trade
with non-European Union member states over the last decade. We obtain a similar picture for the
"small old" member states, although internal EU-27 imports on average are even more important for
these economies. In contrast to the "big four", imports have also grown along the extensive margin.
This increase in traded variety is mainly due to trade with non-European Union members, since in this
category the number of imported varieties has grown substantially from 217,793 to 292,638. Finally,
for the "new" member states we gain a somewhat different picture. First, trade with other European
member states is very important for this group and amounts to approximately 70% of the total imports.
Second, although the trade value from both European Union and non-European Union members in 2008
was roughly four times larger than in 1999, the fact that the number of varieties imported from other
European Union countries has grown by 50% while the number of varieties from the rest of the world
has been slightly decreasing is striking.
Table 2: Aggregated imports for each subgroup
Total Imports (bn. euros)
Total Imports EU-27 (bn. euros)
Total Imports ROW (bn. euros)
Imported varieties (th.)
Imported varieties EU-27 (th.)
Imported varieties ROW (th.)
Note: All variables are calculated by aggregating each individual variable for each of our defined subgroups.
To obtain a better sense of the forces that have been driving the increase in variety over the last
decade Table 3 shows the data by exporting trading partners. Columns one and two of Table 3 rank the
top 30 out of 189 trading partners by the total number of exported goods categories to all the European
member states. The first column presents the ranking for the first quarter of the year 1999 and the
second column the one for the first quarter of the year 2008. In column 3 the rank of the absolute
increase in the number of exported product categories over the last decade is displayed. In columns
4 to 6 we conduct the same experiment but now for the total value of exports to all the member
states. Interestingly, 18 out of the top 30 supplying countries of exported products and total exports
are European member states, despite the European Union’s many small and medium-sized economies.
This reflects the high degree of economic integration and proximity within Europe. Not surprisingly,
the largest four economies of the European Union are also the top suppliers in terms of the number of
products and the total exports to the EU-27. However, over time the rank increase in exported product
categories has been relatively modest compared with the absolute export values for these countries. In
particular, the United Kingdom seems to have profited less from the increased export possibilities over
the last decade. Besides, fast-growing emerging markets like China, India and Turkey, and in particular
small and medium- sized European countries have been the driving force of new product suppliers for
the EU-27, as can be seen from column 3. Given the small size of many Eastern European economies,
only four out of twelve make it into the list of the top 30. Nevertheless, all these countries have improved
their relative position over time and some countries like Poland and the Czech Republic have experienced
one of the largest increases in terms of exported products among all the other countries, ranked 2 and 3
in the absolute increase in the number of varieties. This underlines the growing importance of Eastern
European economies as a source of new products and varieties. On the other hand high-income non-
European Union members like the United States, Japan, Norway and Switzerland have become less
important as a suppliers of new products.
7Intensive margins of trade mean that more products of the same category have been traded. This is in line with the
findings of Besedes and Prusa (2007) for high-income countries.
Table 3: Ranking in terms of the number of goods imported by the EU-27 member states
Notes: The ranking is based on the aggregated number of products exported by a single
country to all European Union member states at the HS-8 level. The total number of
trading partners is 189.
2.2Analysis of Disaggregated Trade Flows
So far our analysis has relied on aggregated data on the European level. We now focus on disaggregated
country data to provide a more detailed picture of the development of imported varieties for each single
member state. Given our assumption that products are differentiated across countries, there are two
potential sources for new varieties. First, an entirely new goods category is imported. Second, the
number of supplying countries of an individual good is increasing. Therefore, Table 4 includes some
descriptive statistics regarding the number of imported product categories and the average trading
partners for a single product. Larger and high-income countries import a larger set of goods from a
more diverse set of countries. This is in line with the New Trade Theory first outlined by Krugman
(1979,1980) and what other empirical studies by Hummels and Klenow (2002) and Broda and Weinstein
(2004) have found.
From columns 1 and 2 we can see that the total number of imported product categories is relatively
constant over time for all the countries: while for most of the “old” member states the number of imported
goods decreases slightly, modest increases can be observed for some of the “new” member states. Hence,
the importing of new products has played a role in extending the set of available products for consumers
in some of these countries. Also note that many of the old member states already imported in nearly all
the product categories at the HS-8 level in 1999.8At the same time the average number of supplying
countries of each individual good has increased for all the countries, except for Hungary and Malta
(columns 3 and 4). The increase has been largest for the “small old” and the “new” member states.
Consequently, the overall variety growth can be mainly attributed to the effect of an increased number
of trading partners for a specific product category. Taken together, this translates into an overall
8This limits the possibility of an increase in the number of varieties via the new goods dimension. “New” goods may
essentially be classified in already existing goods categories, resulting in an potential underestimation of variety growth.
increase of imported varieties for all the countries except Cyprus, Hungary, and Malta, as can be seen
from columns 5 and 6. This is in line with what our previous analysis has already shown. The increase
in imported varieties is largest for the “new”, modest for the “small old”, and relatively small for the
“big four” member states. Although, there are similar numbers of new and disappearing varieties for
most countries (columns 7 and 8), from column 9 and 10 we can see that the value of new varieties is
much higher than the value of disappearing ones, especially for the “new” member states. This indicates
that new varieties have played a more important role than disappearing ones in the consumer budget
Table 4: Variety of EU-27 imports from worldwide trading partners (1999-2008)
Value (mil. euro)
Notes: A good is defined after the HS-8 classification and a variety is defined as a good from a particular country.
For convenience we ordered the table according to our definition of our three subgroups of European Union member
In Tables 5 and 6 we again split up the data by European Union internal and external imports to
evaluate further the different sources of traded varieties for each country. In Table 5, specifics about the
imports stemming from other European trading partners are depicted, whereas in Table 6 the imports
from non-European countries are displayed. Comparing the first two columns of Table 5 with those of
Table 6 we can see that the number of imported product categories from the rest of the world compared
with internal imported varieties is lower for all the countries. Especially, large and “old” member states
already import nearly all the classified product categories from the EU-27 at the beginning of our
period. For example, France imported 9,860 out of 10,428 product categories in 1999. On the other
hand some “new” member states have experienced a strong increase in imported product categories
stemming from other European member states. A case in point is Latvia, which only imported in 5,892
product categories in 1999 but in 7,007 categories in 2008. Similar to the number of imported goods, the
average number of supplying countries as displayed in columns 3 and 4 per good is higher for internally
traded products for all the countries, except for the Netherlands. This is interesting, given the small
number of potential trading partners within the European Union. While for the “big four” the difference
between internal and external average trading partners is relatively small, it is high for all the other
European Union members. In 2008, Germany imported each good on average from 8.79 European Union
members and from 8.55 non-European members, while for the Czech Republic the average was 7.04 for
internally but only 4.10 for externally imported goods.
Table 5: Variety of EU-27 imports from European Union internal trading partners (1999-2008)
Value (mil. euro)
Note: A good is defined after the HS-8 classification and a variety is defined as a good from a particular country.
For convenience we ordered the table according to our definition of our three subgroups of European Union member
This fact stresses the importance of internal European Union trade partners as a source of varieties
for smaller members of the European Union. Over time, the mean number of supplying countries for each
individual good has increased for all the countries, as well from internal and external trading partners.
Here, two points are striking: for the “small old” member states the average supplying country has
grown relatively strongly from non-European Union members, while for the “new” members the average
trading partners increased with other European Union members. Taking the effect of imported product
categories and new trading partners together we obtain the following results for imported varieties
(columns 5 and 6). First, for the “big four” economies neither the EU-27 nor the rest of the world has
been an important source of newly traded products or varieties over the last decade. Second, the small
member states have gained access to new varieties via the establishment of new trade linkages with new
trading partners at the global level, while the European Union and imports of entirely new product
categories only had a minor effect. Finally, “new” member states have benefited from both the import
of new product categories and new trading partners within the European Union, but less from trade
with countries outside the European Union. Regarding the number of new and disappearing varieties,
they are roughly equal in most cases for internal as well as for external imports. Again, the absolute
value of newly imported varieties is strictly higher compared with disappearing varieties.
Table 6: Variety of EU-27 imports from non-European Union trading partners (1999-2008)
Value (mil. euro)
Note: A good is defined after the HS-8 classification and a variety is defined as a good from a particular country.
For convenience we ordered the table according to our definition of our three subgroups of European Union member
In section 4 we provide a more detailed discussion and interpretation of the possible economic forces
that are driving these results.
3 Empirical Strategy
In this section we briefly review the methodology used to determine the gains from variety for the
consumers. It is mainly developed by Feenstra (1994) and by Broda and Weinstein (2006).
3.1 The Variety Gains from Trade
We follow Feenstra (1994) to derive an exact price index for a CES utility function of a single good
with a constant number of varieties. This index is then extended by allowing for new and disappearing
varieties. Finally, we show how to construct a aggregate import price index based on the contribution of
Broda and Weinstein (2006). We start with a simple CES utility function with the following functional
form for a single imported good. To define a variety of a good we assume that imports of one good g
are treated as differentiated across countries of supply, c:
where C denotes the set of available countries and hence of all potentially available varieties.
Mg,c,tis the subutility derived from the imported variety c of good g in period t and dg,c,t> 0 is the
corresponding taste or quality parameter. The elasticity of substitution among varieties is given by σg
and is assumed to be larger than one. Using standard cost minimization gives us the minimum unit-cost
where pg,c,tis the price of variety c of good g in period t and?dc,tis the vector of taste or quality
parameters. Ig,t⊂ C is the subset of varieties of good g imported at time t. Suppose the set of available
product varieties Ig,tin period t and t − 1 is identical, the taste parameters?dc,tare also constant over
time and ? xtand ? xt−1are the cost-minimizing consumption bundle vectors for the varieties of one good
for given the price vectors. In this case Diewert (1976) defines an exact price index as a ratio of the
minimum cost functions
Pg(? pg,t,? pg,t−1,? xg,t,? xg,t−1,Ig) =
where the price index does not depend on the unknown taste parameters?dc,t. Sato (1976) and
Vartia (1976) have derived the exact price index for our CES unit-cost function. It can be written as
the geometric mean of the individual price changes
Pg(? pg,t,? pg,t−1,? xg,t,? xg,t−1,Ig) =
where the weights are calculated using the expenditure shares in the two periods:
So far we have assumed that all varieties of one good are available in both periods to calculate
the exact price index. As our data also include new and disappearing varieties we use the price index
developed by Feenstra (1994) which allows to incorporate new and disappearing product varieties given
by the following proposition.
Proposition: For every good g, if dg,c,t= dg,c,t−1for c ∈ Ig= ((Ig,t∩Cg,t−1); IG?= ∅, then the exact
price index for good g with change in varieties is given by
Πg(? pg,t,? pg,t−1,? xg,t,? xg,t−1,Ig)=
Pg(? pg,t,? pg,t−1,? xg,t,? xg,t−1,Ig)
; r = t,t − 1.
The idea of the Feenstra (1994) index is to correct the conventional price index Pgby multiplying it
with an additional term which measures the influence of new and disappearing varieties and is called the
lambda or Feenstra ratio. If r = t, the numerator of this term quantifies the impact of newly available
varieties as λg,tis the ratio of expenditures on varieties available in both periods relative to the entire set
of varieties available in period t. Hence, λg,tdecreases when new varieties appear and so does the price
index. On the other hand the denominator of the lambda ratio captures the impact of disappearing
varieties. They lower λg,t−1and the index is increased. Secondly, the exact price index depends on the
elasticity of substitution between varieties. If we observe a high elasticity of substitution, the additional
point of view this is intuitive, since new and disappearing products will only have a minor influence on
the welfare of consumers if there exist close substitutes, i.e. if the varieties are homogeneous. Having
derived the exact price index for one good, we can now aggregate the imported goods to an aggregate
import price index as in Broda and Weinstein (2006). This is done by building a geometric mean of the
price indices. The aggregate import price index is then given by
σ−1will approach unity and the influence on the price index is small. From an economic
Π(? pt,? pt−1,? xt,? xt−1,I)=
where the weights wg,tare defined as above. Equation (9) shows that the aggregate exact import
price index is the product of a conventional import price index, CIPI(I), and the aggregated lambda
ratios. Consequently, the following measure, called endpoint ratio (EPR) can be used as an indicator of
the upward bias of a conventional price index compared to the corrected price index. It is the ratio of
the corrected import price index and the conventional import price index:
Using a simple Krugman (1980) structure of the economy, the inverse of the endpoint ratio can be
weighted by the share of imports on the GDP to get the gains from variety:
− 1 =
is the import share.
In order to compute the exact import price index we have to estimate the elasticity of substitution
between varieties of each good. Therefore we briefly review the estimator developed by Feenstra (1994).
Given our utility function (1), we can derive the import demand equation for a single variety using
expenditure shares s as defined above.9Taking logs and first differences results in:
∆lnsg,c,t= ϕg,t− (σg− 1)∆lnpg,c,t+ εg,c,t,
where σg is equal across countries, ϕg,t= (σg− 1)ln[φg,tM(dt)/φM
since dtis unobserved and εg,c,t= ∆ln dg,c,t. The export supply equation in logs and first differences
is specified by
g,t−1(dt−1)] is a random effect
1 + ωg∆lnsg,c,t+ δg,c,t.
where ωg≥ 0 is the good specific inverse supply elasticity10(assumed to be constant across coun-
tries) and δg,c,tis an error term. To identify the elasticity of substitution we have to assume that the
the error terms between the demand and supply curve (εg,c,t,δg,c,t) are uncorrelated after controlling
9Using shares helps to avoid the problems of measurement error of unit-value indices as pointed out by Kemp (1962)
10If ωg = 0 we get the special case of a horizontal supply curve
for good and time specific effects. To take advantage of this assumption we first eliminate the random
terms ϕg,tand ψg,tfrom equations (12) and (13) by taking differences relative to a reference country k:
∆keg,c,t= −(σg− 1)∆kln pg,c,t+ εk
1 + ωg∆klnsg,c,t+ δk
where ∆kKg,c,t= ∆Kg,c,t− ∆Kg,k,tfor K = (lnp,lns), εr
δg,r,t. We can now use the assumption of the independent error terms to multiply (14) and (15) and
dividing by (1 − ρg)(σg− 1) to obtain
Yg,c,t= θ1,gX1,g,c,t+ θ2,gX2,g,c,t+ ug,c,t,
g,c,t= εg,c,t− εg,r,tand δr
?+ ug,c,t or(16)
with obvious definitions of θ1,g and θ2,g. Since the error term ug,c,tis correlated with the prices
and expenditure shares in X1,g,c,tand X2,g,c,t, we do not get a consistent estimator for θ1,g and θ2,g.
However, Feenstra (1994) shows how to exploit the panel structure of the data to get a consistent
estimator by averaging (17) over all t. Hence, we can use the GMM estimator developed by Hanson
(1998) to run a regression on the transformed equation of (17) to estimate θ1,gand θ2,gconsistently.
Yg,c,t= θ1,gX1,g,c,t+ θ2,gX2,g,c,t+ ug,c,t
where upper bars on variables denote sample means over t.11Once, we have consistent estimators
of θ1,gand θ2,gwe can calculate the elasticity of substitution σg:
As long as θ1,g> 0, σgcan be estimated as
and in either case,
? σg= 1 +
?2? ρg− 1
1 − ? pg
In this section we discuss the results of our estimation of the consumer gains from an increased imported
product variety. Furthermore, we show where these gains come from geographically and provide some
robustness measures for our results.
4.1The Gains from Variety in the Countries of the European Union
The final aim of our empirical analysis is to evaluate equation (11), which quantifies the variety gains
from trade with respect to GDP. In a first step we use equation (7) to calculate the lambda ratios for each
11Feenstra (1994) points out that θ1,g and θ2,g can not be estimated separately if the two vectors X1 and X2 are
proportional. Hence, the following identification condition must hold
where c,r and j denote different countries. In words, there must be some differences in the relative variances of the
demand and supply curves across countries.
imported product category of each country. Summary statistics of these ratios are presented in Table
7: for example, the median lambda ratio12for Ireland is 0.96 < 1, expressing that the typical imported
product category in Ireland experienced a positive variety growth of about 4%.13Using the lambda
ratios as a measure of variety growth is more sophisticated than just counting new and disappearing
varieties, as shown previously in Table 4. This measure also accounts for the importance of different
varieties to the consumer budget decision by using expenditure shares as weights.14
Table 7: Lambda ratios
Note: Goods are defined at the HS-8 level.
The summary statistics of the lambda ratios show that, for the largest four EU members, the
growth in imported variety has been moderate, with median lambda ratios of 0.98 and 0.99, indicating
a weighted variety growth of 1% or 2%. For the “small old” member states: the median lambda ratios
are on average slightly lower, ranging from 0.95 in the case of Greece to 1.00 for Luxemburg, indicating
that the variety growth ranges from 0% to 5% in this country group, while for most of these countries,
the observed variety growth lies between 3% and 4%. The “new” member states have experienced an
even larger increase in imported varieties. The descriptive statistics of Section 2 hinted at this result
already, but it is now confirmed by the lambda ratios: the median lambda ratio can be as low as 0.79,
as in the case of Latvia or 0.83 in Lithuania, indicating a variety growth of about 25%. Exceptions
are Malta, Hungary, and Poland with median lambda ratios of 1.00 or 0.99 and therefore with only
small or zero growth in imported variety. From the quantiles it is also obvious that there is substantial
variation across products and countries, emphasizing the importance of using product-specific measures
to calculate accurate gains from imported varieties.
It is important to understand that this growth in variety does not directly lead to consumer gains.
Most importantly, the degree of differentiation within the different product categories is crucial to the
consumers: to make a simple example, it is not important to consumers how many different varieties of
car fuel are available. Fuel is a very homogeneous good, thus it does probably not matter to the consumer
whether it is imported from Norway, Nigeria, Saudi Arabia, or from all three countries. Within a CES
12Here the mean can be somewhat misleading due to the existence of outliers reaching high absolute values.
13Calculated as 1/0.96=4.2%.
14There are fewer lambda ratios than product groups: Some lambda ratios cannot be defined at the HS-8 level since
there is no common variety at the beginning and the end of the chosen time period. We then follow Broda and Weinstein
(2006) and define the lambda ratio at the SITC-5 level.
framework, this homogeneity is expressed by a high value of the elasticity of substitution. Considering
equation (6), it is obvious that for a high value of the elasticity of substitution for a specific product group
the second term on the right hand side will converge to one. In this case, the price index is not being
corrected and the consumers do not gain from the additional varieties (or lose from the disappearing
ones). On the other hand, consumers do care about different varieties within a very differentiated
product group, say for example trainers, furniture, or cars. Expectedly, these product categories exhibit
low elasticities of substitution and therefore new varieties lower the price index substantially. As a
next step, we estimate the elasticities of substitution for every imported product category of each
country following our system of equations (18). Table 8 reports descriptive statistics of the estimated
Table 8: Estimated elasticities of substitution
7,631 7.94 0.55
8,698 16.90 3.81
Czech Republic7,526 13.54
Note: Elasticities are estimated at the goods level, which is defined at the HS-8 level.
Our estimation of the elasticities of substitutions reveals median elasticities between 3.4 and 4.9 in
the different countries. They are of a similar magnitude to those in other contributions, for example in
Broda and Weinstein (2006), Broda et al. (2006) or, Berry et al. (1995). Based on the assumption of a
Krugman type economy, this translates into median mark-ups of between 42% and 25%. Note that the
means are heavily influenced by some outlier elasticities since the elasticities are bounded from below by
1 but are not bounded from above. However, these outliers do not affect our overall results of the gains
from variety: high elasticities just indicate very homogeneous goods that have no impact on the price
index in equation (6). Our results also show that there are no apparent systematic differences between
median elasticities across different countries: for example between small and large countries or between
“old” and “new” member states. This is interesting given the different structures of European Union
economies in terms of size, growth rate and development and does not support theoretical predictions,
as for example in Melitz and Ottaviano (2008). Furthermore, we test whether our estimated elasticities
make sense from a practical view and therefore categorize the elasticities after the classification of Rauch
15A total of 2,093 estimated elasticities in Malta may seem like too few considering that this country imported 5,517
goods in 1999. However, some product categories in very small countries are imported from very few trading partners and
for only a very short time span. For these goods it is not possible to estimate the elasticity of substitution. We drop these
product categories. See Feenstra (1994) for more information about this estimation technique. Note that this is only a
serious issue in the case of Malta and Cyprus, two small and somehow untypical “new” member states. We decided to
report the results all the same.
(1999).16We find that our estimates fit the expectations well: homogeneous product categories exhibit
a median elasticity of 4.8, reference priced products of 4.3, and homogenous products of 4.0.17Using
our estimated elasticities of substitution and the lambda ratios we can calculate corrected price indices
following equation (6) for each of the product categories. Following equation (9), these indices can be
aggregated into a corrected import price index. The ratio of the conventional import price index and
the corrected import price index then results in the EPR as displayed by equation (10). It is worth
explaining the intuition behind this EPR: if this fraction is lower than 1, it means that the changing set
of imported varieties has lowered the import price index. In that case, the consumers profit from lower
unit costs of imports. These lower costs are the source of the welfare gains. On the other hand, if the
EPR is above 1, the import price index is increased by the changing variety set. Thus, the disappearing
varieties are more important to the consumers than the new varieties and the result will be a welfare
loss. Said differently, if we calculate (1/EPR) − 1, we obtain the bias of the conventional import price
index; if the bias is positive there is an upward bias and if it is negative there is a downward bias in
the conventional price index. The EPR and the upward bias of the conventional import price index are
displayed in columns 1 and 2 of Table 9.
Table 9: Import Price Index Bias and the Gains from Variety
Note: Estimates are based on the definition of a good at the HS-
8 level. A variety is defined as a particular good from a particular
The biases in the “big four” countries are relatively small in magnitude. In France, the change in the
imported variety increases the import price index by 1.67% over the whole period, that is, consumers
actually suffer from a slight decrease in the choice of different varieties. In Germany, this decrease is
also present but very close to zero, thus we do not observe a relevant change in the set of varieties
from the perspective of consumers in this country. In Great Britain and Italy, on the other hand, the
newly imported varieties lead to a slight decrease in the import price index. Considering the “small
old” economies of the European Union, we observe that the import price indices decrease due to the
16Rauch (1999) classifies goods as homogeneous if they are traded on organized exchanges, as reference priced if the
goods can be identified by referring to list prices meaning that prices can be quoted without mentioning the name of the
manufacturer, and as differentiated if products differ over a multitude of dimensions including for example a brand name
or the place of selling.
17This also holds for the individual countries in our data set.
new varieties in all the countries except in the cases of Finland and Luxembourg. The magnitude of
this decrease is a bit larger on average than in the larger countries, with Denmark, Greece, and Spain
experiencing a decrease of more than 1.5% in the import price index over the considered time span. The
more accented differences can be observed if we consider the “new” members: in all these countries, with
the exception of Malta, the change in the variety set translates into lower import prices. Furthermore,
the magnitude of the correction in the price index is much larger, with Estonia and Latvia experiencing
lower import prices of over 3%, while in Bulgaria, Lithuania, Romania, and Slovakia, the bias is larger
than 2%. For Poland we observe the lowest positive bias with 0.97%, while in all the other countries it
is larger than 1%.
The bias in the import price index quantifies the drop in import prices due to newly imported
varieties. To obtain the total gains for consumers in relation to the total economic activity, we have to
weight these price decreases (or increases) by the share of imports on this total activity. Based on the
assumption of a Krugman style economy, we compute the gains from variety as a fraction of GDP as
in equation (11). In column 3 of Table 9 we report the import share of each country and in column 4
we calculate the gains from variety (GFV). The GFV can be interpreted in the following way: for an
example, the EPR in the Netherlands is 0.992, indicating an upward bias in the conventional price index
of 0.79% over the whole period. Weighting this bias by the import share of 51% results in an overall
gain from variety through imports of 0.40% of GDP. Put differently, consumers in the Netherlands are
willing to spend 0.40% of their GDP in the year 2008 to have access to the larger set of imported
varieties of 2008 instead of the set of 1999. Not surprisingly, the smaller countries in the sample exhibit
higher import shares, a fact well documented in the literature.18Also note that the “new” member
states exhibit a large dependence on imports, mostly between 50% and 80% of GDP.
The differences in import shares in combination with the correct import price indices result in a
particular pattern of the GFV. The gains from variety in the “big four” of the European Union are very
small. In France, a loss in variety of 0.40% of GDP occurs, while in the other three countries the gains
or losses are closely around zero. Regarding the “small old” members, we observe that the gains from
variety are strongly positive in most cases and around 0.5% of GDP. In this subgroup consumers of
Denmark and Spain enjoyed the highest gains relative to GDP with 0.74% and 0.59%. The exceptions
are Finland and Luxemburg which basically neither gain nor lose from the changed variety set. Again,
the results for the “new” members of the European Union are the most striking ones. In nearly all
the countries, the GFV surpass 1% of GDP. The consumers in Estonia gained 2.80% of GDP in the
last 10 years by the increased imported variety. High GFV above 1.5% of GDP are also found in
Bulgaria, Latvia and Slovakia. Hungary, Lithuania, Romania, and Slovenia exhibit modest gains that
still lie above 1% of GDP. In Cyprus and Malta, the gains are small or even negative, a result of their
dependence on only a few trading partners like Italy or France. In Poland on the other hand, the small
GFV of only 0.34% of GDP are a result of the relatively low import share.
4.2 Geographical Origin of the Variety Gains from Trade
We are now interested in the question of where these GFV come from geographically. In particular,
we would like to know whether the majority of these gains stems from trade with other European
Union economies, or whether other countries, addressed here as “Rest of the World” (ROW) contribute
a substantial share to these gains. The methodology used and presented in section 3 allows us to
compute the EPR for each trading partner, or, more appropriately here, the EPR stemming from trade
with a group of countries. For each country group i, in our case the European Union and ROW; thus
i = EU,ROW, the EPR is computed as follows:
where Wigtis the ideal log-change weight of country group i on good g. Note that, in multiplying all
these EPRi’s, the total EPR as reported in Table 9 results:
EPR = EPREUEPRROW.
18In fact, the variation of import shares is one of the only significant differences between small and large countries that
can be well documented, according to Rose (2006).
The bias in the price index can then be calculated as described above and the results for the 27 EU
members are depicted in Table 10. Column 1 and 2 depict the EPR ratio resulting from the imports
from other EU member states and from the ROW, respectively while columns 3 and 4 display the bias
in the import price index resulting from these imports. For example, German consumers very slightly
gain from the change in the variety set imported from its European trading partners (a decrease in the
price index of 0.02%, as depicted in column 3) but lose from the change in imported variety of its ROW
partners (an increase of 0.15% in the price index as displayed in column 4). In Estonia, the country
with the largest gains from variety, the bias in the price index for imports from other EU members
amounts to 2.89% while the bias stemming from imports of ROW countries accounts for only 0.66%.
This pattern can be observed for many countries, old as well new member states. All these countries
experience lower import price indices through both, imports from other European Union members and
from ROW countries. However, in all these countries, the upward bias is much higher for the European
imports. Some other countries as for example the Czech Republic or Ireland gain from the the higher
variety from intra-European trade but lose part of these gains due to the lower variety imported from
ROW countries. In Finland, this loss even dominates the gains due to imports from the European
Union. The only exception is France, which experiences variety losses in imports from both blocks.
Table 10: Geographical origin of the Gains from Variety
Note: Estimates are based on the definition of a good at the HS-
8 level. A variety is defined as a particular good from a particular
4.3 Interpretation of the Results
Our results can be summarized as follows: the average bias of the import price index for the “big four”
is slightly negative at -0.18%.19Consumers in the four largest economies of the EU consequently lose
slightly from the change in the set of imported varieties and this loss stems from trade with the ROW
countries. For the “smaller old” member states we estimate positive gains with an average of 1.24%.
Although these countries have profited from both internal and external imports our results show that
more than 70% of the gains can be attributed to within European trade. This is interesting, given
19We calculated the weighted average bias of the import price index using the size of each country in term of its GDP.
This is done to obtain a clearer picture of the differences between the three country blocks.
our somewhat different results from section 2, where we find a strong increase in the number of newly
imported varieties from non-European members but at the same time observed only a modest increase
in imported varieties from trade within Europe. This fact again underlines our approach to take into
account the weights of each product in the consumption decision instead of simply counting the number
of new varieties. Finally, for the “new” member states, we find a large average bias in the import price
index of 1.68%. These countries have benefited substantially from internal European trade over the last
decade, which accounts for 90% of the total gains from imported varieties.
One explanation for this pattern of the gains from variety makes use of the ongoing process of
European integration as well as globalization in general: the "big four" countries were already playing
a key role in the global economy at the beginning of our period and had well- established trade links as
well within the European Union as well as within the global trading system. Consequently, access to new
varieties via new trade linkages has been limited, given their already diverse structure of imports in 1999.
Hence, we observe that most trade has been growing at the intensive margin, resulting in relatively low
gains. Besides these reasons, the smaller import shares also play a role in these countries. In addition,
some trade diversion from ROW towards the European Union may have taken place, adding another
possible explanation for our slightly negative results for some of these countries especially concerning
imports from ROW.
For the high income "small old" member states, we observed in section 2 that their import diversity
is somewhat more limited compared with the largest countries. The increase over the last decade
has been more substantial and their large import shares make imports, and the imported variety in
particular, an important source of welfare gains. Nonetheless, the trading of these countries with other
European members was already well diversified by the year 1999 and especially the European Union has
been an important source of imports for decades. Furthermore, most of these countries have been part
of the European Union for a longer period and had already adopted important institutions like the single
market programme before 1999. This in combination with the proximity to the other member states
can explain the slower growth rate of new trade linkages within the European trade network compared
with the "new" members.
These "new" member countries, on the other hand, were less integrated into the world trading
system in 1999 and consequently have taken advantage of the dynamic globalization process over the
last decade to diversify and extend their imported product set at the global level. In contradiction to the
stylized facts,however, our estimated results for the gains show that most of the gains stem from internal
European imports. This can be reasoned by the fact that the new trade linkages from non-European
members only had a minor impact on the gains from variety, due to their relatively small share of
expenditure in the overall consumption bundle. Finally, our results reveal an expected process, namely
that over the last decade, the new member states have caught up with the older ones, regarding the
integration into the European market. Consequently, the European Union has been an important source
of new products and varieties for these countries. With the reorientation of the transition economies
towards "old" Europe in combination with the reduction in trade barriers and the adoption of important
European Union institutions during the accession period, trade linkages of these countries with all the
other EU-27 members have grown at a rapid rate, resulting in substantial consumer welfare gains due
to the existence of a more diverse set of products and varieties.
5 Robustness Checks
Since our empirical approach rests on some strong assumptions we address four important issues to
build further confidence in our results in this section. One issue is the dependence of the results on the
estimated elasticities of substitution. Secondly, we would like to talk about the level of disaggregation
of the trade data ans therefore about how a variety is defined. As a third point, we argue that we obtain
welfare gains of consumers although we use all the imports to estimate the GFV instead of only imports
of consumer goods. Fourth, since the methodology used above only focuses on imported varieties and
neglects changes in the domestic variety, we provide a short discussion on how this effect may change
The Degree of Substitutability
Estimating elasticities of substitution from trade data is not an easy task. Due to the data restrictions,
several strong assumptions have to be made to identify this parameter. It would be beyond the scope
of this paper to discuss this in detail.20However, we would like to assess the impact of potential biases
on the variety gains from trade.
Table 11: Robustness of the results using different values for the elasticity of substitution
ˆ σ +50%
σ = 2
σ = 3
σ = 4
σ = 8
Note: Elasticities are estimated at the goods level, which is defined at the HS-8 level.
The first column shows the results from above using the estimated elasticities ˆ σ while columns 2
and 3 show how the bias changes if we increase or decrease our estimates by 50%.21The idea of this
experiment is to assess whether a possible bias in the estimated elasticities can have a substantial effect
on our results. Even though we assume a large upward and downward bias of 50% in every estimated
elasticity, our results remain relatively constant. For some countries the bias quite expectedly becomes
large if the much lower sigmas are used.22
In columns 4 to 7 of Table 11 we use a fixed value for the elasticities of substitution for all the
product groups. We present these results here to demonstrate the potential bias of using a single value
for all the elasticities of all the product groups. Consider for example Italy. With the estimated values
of the sigmas (column 1), new varieties lower the import price index by 0.76%. As Table 4 shows, the
median sigma in Italy is 4.60. However, using a similar elasticity of, say 4, for each product results in a
bias of 2.11%, almost three times larger. Even if we use a common elasticity as high as 8 the bias is still
higher than the results when using the estimated elasticities that vary for each product group (0.91%
compare with 0.76%). Thus, if we are interested in finding the "true" gains from imported variety, it is
of central importance to estimate the elasticities for each product category. To explain the differences
in the results for the gains from variety between the fixed and the estimated elasticities of substitution
20Interested readers are referred to Feenstra (1991), which is the more detailed working paper version of Feenstra (1994),
as well as to an appendix available on Robert C. Feenstra’s website. Furthermore, Soderbery (2009) discussed some of
the properties of this estimator.
21For example, if ˆ σ = 1.5, we use σu= 1.75 as an upper bound and σl= 1.25 as a lower bound. Thus, we first substract
1 from the estimate, then increase or decrease the remainder by 50% and then add 1.
22Recall that lower values for the elasticity of substitution imply higher gains since consumers value the availability of
new varieties more.
consider the following example: when using the fixed values of sigma, highly homogenous products with
a high import share and a high elasticity of substitution like gasoline will substantially bias the "true"
gains from variety upwards given the misspecification of the elasticity of substitution when fixed values
The Definition of a Variety
Next, we analyse the extent to which our results depend on the definition of a variety. This is of course
a central issue: the more detailed the data, the higher the “turnover” of varieties and thus, potentially,
the higher the gains from variety. Table 12 presents the results for different levels of aggregation of the
data. The HTS-8 results from above are shown in column 1. As mentioned before, about 10.000 product
categories are defined at that level. The EPRs of data at the HTS-6 level are presented in column 2. At
this level of aggregation about 6.000 categories are defined. Column 3 then shows the results of HTS-4,
which defines only slightly over 1.000 products. The bias and GFV for these levels of aggregation are
then shown in columns 4 to 6 and 7 to 9, respectively. Comparing the bias and the GFV of HTS-6
and HTS-8, we conclude that the results are sufficiently robust: for many countries, the bias is slightly
lower using HTS-6 compared with using HTS-8, as expected. For other countries, however, this is just
reversed. This result may be counterintuitive at first sight. However, this is perfectly possible since we
also reestimate the elasticities of substitution at this level of aggregation. The median sigmas at the
HTS-6 level are lower than those at the HTS-8 level (results omitted here). This is an intuitive result
since more broadly defined product categories generally yield lower substitutability of the contained
varieties. Hence, by using less disaggregated data, we may miss some variety growth: the varieties
observed, however, are estimated as being more differentiated and therefore they contribute more to the
gains from variety.
Table 12: Robustness of the results using different aggregation levels of the data
Note: Different levels of product aggregation are based on the Hármonized system nomenclature of the World
Considering the results using HTS-4, we observe that the bias is always much lower in magnitude
compared with HTS-6 or HTS-8. With the restriction to only 1.000 product categories we lose much of
the information of the variety change that actually occurs in the more disaggregated data. This effect
also dominates the opposite effect of the slightly lower elasticities in the HTS-4 case (results omitted).
Most importantly note that our results remain fairly constant in terms of relative size across countries.
Hence our qualitative conclusions remain valid.
Ultimately, this leads to the question of what the optimal definition of a variety should be. Blonigen
and Soderbery (2009) argue that the variety gains from trade as estimated above are underestimated
since trade data hides some variety growth. They show, using very detailed market data of the U.S.
autombile market, that the gains from variety are 50% higher if these more disaggregated data are
used instead of standard trade data. In the same vein is a comment by Bernard et al. (2009), who
argue that even new (and still very scarce) firm-level data would imply higher variety gains from trade,
because since every firm produces different varieties instead of “just” every country. To sum up, our
results presented above may be a lower bound and the actual gains may even be higher due to data
limitations. On the other hand, Table 12 has also shown that an opposing force, namely the higher
estimated elasticities of substitution exists, if more detailed data are used.
The Effects on Domestically Produced Varieties
One central issue remains to be discussed when speaking about the variety gains from trade. Using
the model described in Section 2, we implicitly assume that domestic and foreign goods cannot be
substituted. That is, a change in the variety of imported goods does not affect the domestic economy,
or more specifically the variety of domestically produced goods. This is the same stark assumption as
Broda and Weinstein (2006) use. It is not hard to find a model that addresses this issue theoretically.
For example in Melitz (2003), more productive foreign firms crowd out the less productive domestic
firms, leading to a decrease in domestically produced varieties. As Baldwin (2006) or Arkolakis et al.
(2008) show, the total variety consumed in a country can even decrease after trade liberalization in
such a model.23Empirically, the effect on domestic production is harder to assess due to the lack of
availability of disaggregated domestic production data.
In a very recent contribution however, Ardelean and Lugovskyy (2009) address this issue. They
set up a simple model, where varieties can be substituted on two levels: on the first level, domestic
varieties can be substituted with a constant elasticity. The same is possible within a foreign good. On
the second level, foreign and domestic varieties are substitutable by another, sensibly a lower, elasticity
of substitution. Thus, depending on the magnitude of the elasticities, foreign varieties could replace
domestic varieties upon trade liberalization. The authors then quantify a potential bias that results from
ignoring this possible substitution. Using data on U.S. manufacturing sectors, they find that in some
sectors, such as electronics, the variety change is even underestimated by as much as 90%. Thus, trade
liberalization even led to a larger increase in variety if the domestic sector is taken into account. On
the other hand, for other sectors, like machinery & transportation, the variety change is overestimated
by 40% neglecting domestic variety. On average, the bias in the variety change is small, accounting for
an overestimation of 8%.
For our results presented above this means that the stark separation of domestic and foreign varieties
does not lead to a systematic bias. Specifically, we do not systematically overestimate the gains from
variety due to neglecting the effect on the domestic variety. We may overestimate the gains stemming
from some product categories but underestimate the gains stemming from others. Whether this leads
to a bias in total is difficult to say. Also, it is not easy to answer whether this potential bias is higher
for some countries than for others. To address these important questions, further research, for example
using more detailed data sets that are restricted to some product categories, is necessary but beyond
the scope of this paper.
Over the last decade the member states of the European Union have been part of a dynamic economic
integration process as well within the European Union as well as at the global level, resulting in a
strong increase in imported products and varieties. In this paper we adopt the methodology outlined
by Feenstra (1994) and Broda and Weinstein (2006) analyse and estimate the positive effects of variety
growth on consumer welfare for all the European Union member states for the period from 1999 to 2008.
23Of course, one still had to weight these verieties by the expenditure shares and the degree of substitutability. Thus
these results themselves do not imply that the gains from variety would be negative.
Our results show that for most countries the import price index is biased upwards due to the
omission of newly imported varieties. This gives rise to positive welfare gains to consumers stemming
from an increased product variety. However, our analysis also reveals substantial differences across
countries. Based on the assumption of a Krugman type economy, we can hardly identify any gains from
newly imported varieties over the last decade for the largest four countries of the European Union. On
the other hand, these gains are more significant for the smaller and especially younger member states
of the European Union. Here, our results suggest positive welfare gains of up to 2.8% of GDP, as in
the case of Estonia. This fact demonstrates that especially for smaller and fast-growing economies the
creation and extension of trade linkages can be an important source of welfare, a fact often neglected in
the discussion about the positive effects of globalization and economic integration.
To shed further light on the source of these gains, we develop an empirical strategy that allows us
to identify the extent to which intra-EU and non-EU imports contribute to the gains from variety. Our
analysis shows that between 70% and 100% can be attributed to increased variety imports from other
European Union members. Imports from other countries did not contribute much to these gains; on
the contrary, according to our results these imports often even contributed negatively, thus mitigating
the positive effects of variety growth in the total imports. Thus, the ongoing integration of countries
within the European Union positively benefits consumers with availability of an increased consumption
set. Specifically, these predominantly stem from European varieties accentuating the economic benefits
of the European integration process. Consequently, our empirical study is also interesting for future
European Union accession candidates.
Finally, we provide a sensitivity analysis and conclude that our results are reasonably robust to
other specifications. We show that the estimation of different elasticities of substitution for different
product categories is a central issue. Additionally, using aggregated data may hide significant growth
along the extensive margin leading to an underestimation of the gains from variety. While our study
solely focuses on the consumption side, the methodology and data can also be easily implemented to
analyse the positive effects of variety growth on production and productivity and may be an interesting
field for future research.
Ardelean, A., Lugovskyy, V., 2009. Domestic productivity and variety gains from trade. Journal of
International Economics 80, 280–291.
Arkolakis, C., Demidova, S., Klenow, P. J., Rodríguez-Clare, A., 2008. Endogenous variety and the
gains from trade. American Economic Review 98 (2), 444–450.
Armington, P. S., 1969. A theory of demand for products distinguished by place of production. Inter-
national Monetary Fund Staff Papers 16, 159–178.
Baldwin, R., 2006. The euro‘s trade effect. ECB Working Paper Series 594.
Bernard, A. B., Eaton, J., Jensen, J. B., Kortum, S., 2003. Plants and productivity in international
trade. American Economic Review 93 (4), 1268–1290.
Bernard, A. B., Jensen, J. B., Redding, S. J., Schott, P. K., 2009. The margins of U.S. trade. NBER
Working Paper 14662.
Berry, S., Levinsohn, P., Pakes, A., 1995. Automobile prices in market equilibrium. Econometrica 63 (4),
Besedes, T., Prusa, T. J., 2007. The role of extensive and intensive margins and export growth. NBER
Working Paper 13628.
Bils, M., Klenow, P. J., 2001. The acceleration in variety growth. American Economic Review 91 (2),
Blonigen, B. A., Soderbery, A., 2009. Measuring the benefits of product variety with an accurate variety
set. NBER Working Paper 14956.
Broda, C., Greenfield, J., Weinstein, D. E., 2006. From groundnuts to globalization: A structural
estimate of trade and growth. NBER Working Paper 12512.
Broda, C., Weinstein, D. E., 2004. Variety growth and world welfare. American Economic Review 94 (2),
Broda, C., Weinstein, D. E., 2006. Globalization and the gains from trade. Quarterly Journal of Eco-
nomics 121 (2), 541–585.
Buch, C., Piazolo, D., 2001. Capital and trade flows in europe and the impact of enlargement. Economic
Systems 253, 183–214.
Chen, N., 2004. Intra-national versus international trade in the european union: Why do national
borders matter? Journal of International Economics 63 (4), 93–118.
Diewert, W. E., 1976. Exact and superlative index numbers. Journal of Econometrics 4 (2), 115–145.
Dixit, A. K., Stiglitz, J. E., 1977. Monopolistic competition and optimum product diversity. American
Economic Review 67 (3), 297–308.
Feenstra, R. C., 1991. New goods and index numbers: U.s. import prices. NBER Working Paper 1902.
Feenstra, R. C., 1992. How costly is protectionism? Journal of Economic Perspectives 6 (3), 159–178.
Feenstra, R. C., 1994. New product varieties and the measurement of international prices. American
Economic Review 84 (1), 157–177.
Feenstra, R. C., Kee, H. L., 2008. Export variety and country productivity: Estimating the monopolistic
competition model with endogenous productivity. Journal of International Economics 74 (2), 500–518.
Feenstra, R. C., Markusen, J. R., 1994. Accounting for growth with new inputs. International Economic
Review 35 (2), 429–447.
Funke, M., Ruhwedel, R., 2005. Export variety and economic growth in eastern european transition
economies. Economics of Transition 13 (1), 25–50.
Hanson, G. H., 1998. Market potential, increasing returns, and geographic concentration, mimeo, Uni-
versity of Michigan.
Hausman, J. A., 1981. Exact consumer surplus and deadweight loss. American Economic Review 71 (4),
Hausman, J. A., 1994. Valuation of new goods under perfect and imperfect competition. NBER Working
Hummels, D., Klenow, P. J., 2002. The variety and quality of a nation’s exports. NBER Working Papers
Kemp, M. C., 1962. Errors of measurement and bias in estimates of import demand parameters. Eco-
nomic Record 38, 369–372.
Klenow, P. J., Rodríguez-Clare, A., 1997. Quantifying variety gains from trade liberalization, working
Paper, University of Chicago.
Krugman, P. R., 1979. Increasing returns, monopolistic competition, and international trade. Journal
of International Economics 9 (4), 469–479.
Krugman, P. R., 1980. Scale economies, product differentiation, and the pattern of trade. American
Economic Review 70 (5), 950–959.
Krugman, P. R., 1981. Intraindustry specialization and the gains from trade. Journal of Political Econ-
omy 89 (5), 959–973.
Manchin, M., Pinna, A. M., 2009. Border effects in the enlarged eu erea: Evidence from imports to
acession countries. Applied Economics 41 (14), 1835–1854.
Melitz, M. J., 2003. The impact of trade on intra-industry reallocations and aggregate industry produc-
tivity. Econometrica 71 (6), 1695–1725.
Melitz, M. J., Ottaviano, G. I. P., 2008. Market size, trade, and productivity. Review of Economic
Studies 75 (1), 295–316.
Nitsch, V., 2000. National borders and international trade: Evidence from the european union. Canadian
Journal of Economics 33 (4), 1091–1105.
Rauch, J. E., 1999. Networks versus markets in international trade. Journal of International Economics
48 (1), 7–35.
Romer, P. M., 1994. New goods, old theory, and the welfare costs of trade restrictions. Journal of
Development Economics 43 (1), 5–38.
Rose, A. K., 2006. Size really doesn’t matter: In search of a national scale effect. NBER Working Paper
Sato, K., 1976. The ideal log-change index number. Review of Economics and Statistics 58 (2), 223–228.
Smith, A., Venables, A., 1988. Completing the internal market in the european community. European
Economic Review 32 (1), 1501–1525.
Soderbery, A., 2009. Investigating the asymptotic properties of import elasticity estimates, working
Paper, University of California, Davis.
Spence, M., 1976. Product selection, fixed costs, and monopolistic competition. Review of Economic
Studies 43 (2), 217–235.
Trajtenberg, M., 1989. Welfare analysis of product innovations, with an application to computed to-
mography scanners. Journal of Political Economy 97 (2), 444–479.
Vartia, Y. O., 1976. Ideal log-change index numbers. Scandinavian Journal of Economics 3 (3), 121–126.
Yi, K.-M., 2003. Can vertical specialization explain the growth of world trade? Journal of Political
Economy 111 (1), 52–102.