Taste differences, home markets, and the gains from trade

Raphael Auer

19 September 2010



In his 2008 Nobel Prize Lecture, Paul Krugman announced that the initial motivation behind his analysis of the home-market effect was in order “to bury Burenstam Linder, not to praise him” (Krugman 2008). Linder's famous conjecture in 1961 was that "[t]he more similar is the demand structure of two countries, the more intensive, potentially, is the trade between these two countries" (Linder 1961). I argue here that Krugman succeeded in part his original intention because Linder's conjecture is actually two fold – by confirming one part, Krugman disproved the other. In doing so I provide support for the recent quantifications of the gains from trade and find that, contrary to Linder, the volume of trade is entirely unaffected by the distribution of foreign tastes.

What is the home-market effect?

The home-market effect, as outlined by Paul Krugman (1980), is the property that distinguishes the "new" trade theory most clearly from supply-sided motives for trade. This effect relates bilateral net trade flows to the across-country differences in the distribution of consumer tastes.

How do such taste differences affect the volume of trade and the welfare gains from liberalisation?

There is now ample empirical evidence for the home-market effect and thus also for the existence of cross-country taste heterogeneity. Yet while there is ample discussion and informal notion in the literature suggesting the differences between theoretical predictions and empirical estimates of trade volume can be explained by unmeasured cross-country differences in consumer tastes, there is very limited formal analysis of how such preference heterogeneity affects the aggregate volume of trade and the welfare gains from liberalisation.

In a recent study (Auer 2010) I revisit the Linder hypothesis and document that, owing to the way in which the domestic industrial composition adjusts to trade, the effects of trade liberalisation in a structural model of across-country demand heterogeneity may be entirely unaffected by the distribution of foreign tastes and may also coincide with those in a representative agent framework.

A structural model of demand by heterogeneous consumers

To demonstrate this result, I first develop a preference framework in which consumers are heterogeneous in their valuation of attributes, goods are heterogeneous in the level of the attribute they deliver, and in equilibrium good attributes and consumer valuations tend to be matched assortively.

I nest these preferences in a model of the international economy featuring iceberg transportation costs and two countries that differ in the distribution of consumer valuations. The model comprises the standard representative agent framework of Dixit and Stiglitz (1977) as a special case, thus allowing us to directly evaluating the impact of taste heterogeneity on trade flows, industry composition, and the welfare gains from trade.1

The latter impact is shown to depend crucially on the degree to which the domestic industry composition can adjust to counteract the mismatch between the attribute composition of imports and the domestic distribution of tastes, i.e. on whether countries are completely specialised or not.

The invariance of relative competition to trade

For the case of incomplete specialisation, I find that the class of models of international trade deriving from Dixit and Stiglitz (1977) and Krugman (1980) gives rise to a result on the invariance of relative ideal price indices to trade that is akin to the factor price equalisation theorem in the classical theory of trade.

In the latter constant returns economy, costless trade equates good prices and any equilibrium featuring incomplete specialisation requires that factors of production receive exactly the same reward across countries.

In the increasing-returns economy of the “new” trade theory, the adjustment of firm profits is attained via relative firm scale. Irrespective of whether markets are open or not, the free firm "attribute-entry" condition requires that the set of competitors (adjusted for trade costs) producing a certain type of good is proportional to the number of consumers with a preference for the type of good in question.

It is then shown that the interplay of the free attribute-entry conditions at Home and in Foreign requires that the domestic industry structure must adjust such as to exactly counteract the mismatch between the attribute composition of imports and the domestic distribution of tastes, in turn leaving the relative "toughness" of competition unaffected by trade.

Implications for the gains from trade and Linder’s hypothesis

A first main implication of this finding is that the recent quantifications of the gains from trade based on the representative agent framework (see Broda and Weinstein 2004, Broda et al. 2006, and Akorlakis et al. 2008), hold exactly even if the representative agent is actually not the correct description of the underlying consumption decisions.

For example, if there are two goods wine and beer, assume that the French population consists mostly of wine lovers, while the opposite is true for Germany. German exports are then too "beer-intensive" for the typical French consumer, but this is offset by the French industry specialising into the wine segment. On the consumer side, this implies that while the group of French beer lovers gains relatively more from the imported German varieties, the domestic industry responses favour French wine lovers. In equilibrium, these two effects exactly offset each other and all consumers benefit from trade in the same proportion irrespective of the distribution of tastes abroad.

The second main implication of this finding is that the volume of trade is entirely unaffected by the distribution of foreign tastes. Notions along the lines of Linder or Armington (1969) disregard the fact that demand for imports is not determined by preferences alone, but also by how well these preferences are served by the domestic industry. With trade, the home-market effect implies that a lower domestic valuation for an attribute is associated with an over-proportional reduction in domestic production of goods embodying the attribute and consequently, with high import demand for precisely those goods that are in low overall demand.2

An aside on the history of economic thought in international trade

It is interesting to view the above mentioned results in relation to the evolution of economic thought in international trade. Linder (1961) actually made two conjectures. The first is that domestic demand is a prerequisite to export (coined the home-market effect by Krugman 1980) and the second one (successively known as the Linder hypothesis) is that across-country taste differences impede trade.

Somewhat paradoxically, I document that Linder's second conjecture does not hold true precisely because the first one does, i.e., that the home-market effect is always such that it exactly offsets the mismatch between the composition of imports and the distribution of domestic demand.

This sheds a new light on Paul Krugman's original analysis of the home-market effect. He initiated his study in order "to bury Burenstam Linder, not to praise him" (Krugman 2008). Contrary to what Krugman argues later in his Nobel Prize speech, it turns out that he actually partly succeeded in his original intention as a by-product of confirming the other, probably more consequential, part of Linder's conjecture.


Armington, Paul S (1969), "A theory of demand for products distinguished by place of production", IMF Staff Papers 16:159-176.

Arkolakis, Costas, Pete Klenow, Svetlana Demidova, and Andres Rodriguez-Clare (2008), "The Gains from Trade with Endogenous Variety", The American Economic Review, 98(4):444-450.

Auer, Raphael (2010), "Consumer Heterogeneity and the Impact of Trade Liberalization: How Representative is the Representative Agent Framework?," Working Papers, 2010-13, Swiss National Bank.

Broda, Christian, and David Weinstein (2006), "Globalisation and the Gains from Variety", The Quarterly Journal of Economics, 121(2): 541--585.

Broda, Christian, Josh Greenfield and David Weinstein (2006), "From Groundnuts to Globalisation: A Structural Estimate of Trade and Growth", NBER Working Paper 12512, September.

Dixit, Avinash V, and Joseph E Stiglitz (1977), "Monopolistic Competition and Optimum Product Diversity", The American Economic Review, 67(3):297-308.

Krugman, Paul R (1980), "Scale Economies, Product Differentiation, and the Pattern of Trade", The American Economic Review, 70(5):950-959.

Krugman, Paul R (2008), "The Increasing Returns Revolution in trade and Geography", Nobel Prize Lecture, December 2008.

1 A key assumption of the model is that firms can decide with what kind of good to enter the market and that therefore, attribute-entry is directed towards the distribution of consumer tastes.
2 Linder (1961) famously argued that across-country taste differences impede the volume of trade and the gains from liberalisation. It is also noteworthy that the analysis also implies that taste heterogeneity may only matter little for the effect of liberalisation even if specialisation is complete. For example, if parameters are such that specialisation is complete but only marginally so, the volume of trade and the welfare gains from trade are still equal to the one prevailing in a representative agent framework.



Topics:  International trade

Tags:  international trade, home-market effect, consumer tastes

Deputy Head, International Trade and Capital Flows, Swiss National Bank and Research Affiliate, CEPR