Export prices are frequently interpreted as reflecting the quality of a nation’s exports. Hummels and Klenow (2005), for example, showed that richer countries export the same good at somewhat higher prices than less rich ones. More recently, we have seen that the same country charges different prices for the same product in different destination markets. This variation is not random, Baldwin and Harrigan (2007) show that prices charged by a country for a particular good are positively related to distance between the exporting and importing country. Naturally, this result would not be surprising at all for export prices including transportation cost – but these studies always build on “free-on-board” prices that exclude transportation, insurance costs and customs duties. At first glance, it is not clear why these prices should be in any relationship with distance, let alone a strong positive correlation.

Why do export prices differ?

There are two different theoretical explanations for this phenomenon. The simpler one is that the same firm charges different prices in different markets because of, for example, the varying intensiveness of competition across markets (for example Melitz and Ottaviano 2008). In particular, any exporting firm is in a disadvantaged position compared to local firms in a far away country; the farther away the country, the larger this competitive disadvantage. As a result, firms should absorb part of the transportation cost in their free-on-board prices to remain competitive in far away export markets. This argument predicts a negative relationship between distance and within-firm export prices.

The second explanation starts from the idea of self-selection, in the fashion of the heterogeneous firm literature. Firms producing higher quality (and more expensive) goods are more likely to export to distant markets. This leads to a composition effect. As on average higher quality goods are exported to more distant markets, the observed average price will also be increasing in the distance between the two trading countries even if all individual firms charge the same price across export markets, as Baldwin and Harrigan (2007) argue.

Within-firm price differences

Bilateral and country level trade data are not suitable to distinguish between these two explanations. However, with the help of recently available firm-product-destination level trade datasets, it is possible to follow individual firms’ export prices, and compare their characteristics to average export prices. In recent research (Görg et al. 2010) we exploit Hungarian customs data to shed light on this question. This dataset includes both quantities and export receipts, so we can calculate export unit values which are interpreted as export prices. We normalised unit values by using their mean for every product to make them comparable across products. These normalised unit values show the price charged by a firm in a destination market relative to the average price charged for that product by all Hungarian firms in all export markets.1

Before going down to the firm-level, we checked the aggregated pattern to relate our results to earlier works. In Figure 1 we show the aggregate relationship between distance and these normalised unit values, averaged for each of Hungary’s top 100 destination markets over all firms and products. The positive relationship between distance and average export prices, documented for a large sample of countries by Baldwin and Harrigan (2007) is replicated. A simple regression shows that export prices are 12% higher when distance from Hungary is doubled.

Figure 1. The relationship between average Hungarian export prices and distance.

Note: The size of the circles is proportional to total export volume.

This aggregated result does not allow to infer, whether it is an outcome of selection or within-firm price differences. This requires a different normalisation, using the average price at the firm-product level. Figure 2 shows how the price of the firm in one market differs from its own average price in all markets. The pattern of these prices reveals whether firms charge different prices for the same product in different markets.2 The relationship is clearly (and significantly) positive. The slope of the fitted line, however, is only 5.5% compared to 12% for the aggregated data. Given that firms charge higher prices for the same product in more distant markets, the difference between the slopes should be the consequence of selection.

Figure 2. The relationship between within-firm price differences and distance.

Note: The size of the circles is proportional to total export volume.

A high quality beer might solve the problem

The positive relationship between within-firm export prices and distance calls for further investigation. The Melitz-Ottaviano (2008) model mentioned above, where firms absorb transportation costs partly, predicts a negative relationship. So how can our finding explained? One possibility is that firms ship the good apple out, as Alchian and Allen (1964) argued in a slightly different context. If transportation costs are a function of the physical quantity of products rather than their value, then firms will export higher quality products than sold in the domestic market. This requires that firms are able to differentiate their own goods even within quite narrow product categories.

Imagine a brewery. It sells its low quality products to domestic consumers and exports them to nearby markets, and ships the premium brand to both nearby and far away markets. This leads to a within-firm composition effect. On average higher quality beer arrives to more distant markets. The same brewery may try to build a different image for the same beer in different countries. Maybe it is optimal to position it to a mass market category with its low marginal cost at home; but the firm may try to present it as an exotic, high-value variety in far-away countries, where marginal cost includes high transportation costs. This example may also fit with other research because we cannot exclude that some firms’ free-on-board prices include a part of the mark-up on the transportation cost. Moreover, within-firm price differences are not significant for homogeneous goods, what gives some credibility to the quality driven pricing.


Firms charge different prices for the same product in different international markets. But while trade theorists are developing models that can reflect this diversity, current models at hand do not square with our empirical results either qualitatively or quantitatively. The most intricate issue is how the geographical distance between countries – a proxy for transportation costs – affects export prices. We have found that properly defined and measured within-firm export prices are in positive relation with distance. This calls for new approaches and models to separate the effects caused by quality pricing and price discrimination.


1 Though the dataset consists of information for several years, we have chosen 2003 for this analysis. The results are similar for other years. Also, we have included only direct exports by manufacturing firms, as trade theories are most relevant for this sector. One product in these data is a 6-digit Harmonised System category.

2 This figure only shows firm-product combinations with at least 6 observations.


Alchian, Armen A and William R. Allen (1964), University Economics, Belmont CA: Wadsworth Publishing Company.

Baldwin, Richard E and James Harrigan (2007), “Zeros, Quality and Space: Trade Theory and Trade Evidence”, CEPR Discussion Paper 6368.

Görg, Holger, László Halpern, and Balázs Muraközy (2010), “Why do within firm-product export prices differ across markets?”, CEPR Discussion Paper 7708.

Hummels, David and Peter J Klenow (2005): “The Variety and Quality of a Nation’s Exports”, American Economic Review, 95(3):704-723.

Melitz, Marc J and Gianmarco I P Ottaviano (2008), “Market size, Trade and Productivity”, Review of Economic Studies, 75(1): 295-316.

2,520 Reads