Vox columns by Peter Schott (2007) and Fontagné et al. (2008) have argued that developed countries specialise in the production of high-end varieties – expensive varieties of a product which have specific attributes such as reputation, branding, or quality that make them appealing to consumers in spite of their higher price. A few papers have examined the implications of such a specialisation for the labour market (e.g. Verhoogen 2008). However, while policymakers and academics encourage specialisation in high-end varieties, the macroeconomic consequences have not yet been examined. In a recent paper (Martin and Mayneris 2013) we try to fill this gap, and ask how countries’ specialisation in the production of high-end varieties affects the growth and volatility of their exports.
A luxurious procedure to identify high-end variety exporters
Before discussing the results, we wish to emphasise that identifying high-end varieties in the data is a key challenge. How to identify firms selling expensive varieties due to specific quality, reputation, or image, without quantifiable proxies for these characteristics? We propose a two-step procedure based on French firm-level customs data. First, we use the list of members of the Comité Colbert, an organisation composed of the main brands of the French luxury industry. There are 76 members, including brands as famous and expensive as Baccarat, Cartier, Champagne Bollinger, Chanel, Christian Dior, Hermes, Louis Vuitton, and Yves Saint-Laurent. We thus (reasonably) assume that these firms are exporting high-end varieties. Then, we consider the universe of French exporting firms, and we apply the following rule: firms that sell the same products at least at the same price (unit value) as Colbert firms are tagged as high-end variety exporters. Different internal and external checks allow us to validate this procedure. We end up with a sample of 8,379 high-end variety exporters – out of more than 60,000 exporters in total – active in 198 products over the period 2000–2011. Total sales of high-end variety exporters increased from €9 billion to €15 billion over the period, i.e. from 32% of the total French exports of the products we consider to 37% at the end of the period.1
High-end producers do not export to more markets, but do export to more distant ones
We use our classification of high- and low-end variety exporters in France to compare the characteristics of these firms. By definition, high-end variety exporters charge higher prices than low-end ones. In spite of their high price, they export the same quantities on average as low-end variety exporters – in line with the idea that their high price reflects specific attributes valuable to consumers. Furthermore, the distribution of firm-level exports displays huge heterogeneity for both high- and low-end exporters. Among high-end variety exporters, extremely large players co-exist with very tiny niche producers. We also find that high-end exporters do not export to many more markets than low-end exporters. Actually, as shown in Figure 1, while Colbert firms are giant firms, the other high-end exporters tend to export to the same number of destinations as low-end firms, and their product scope is slightly smaller. In fact, the key difference between high- and low-end variety exporters is not the number of destinations served but the average distance of their shipment.
Figure 2 plots the cumulative export shares of low-end firms, Colbert firms, and other high-end firms over different distances. As expected, Colbert firms export to more distant markets on average. This is due to the fact that they are very large firms exporting to many countries. However, while we find that 75% of low-end firm’s exports are sold in markets as close as 2,000 km to France, other high-end exporters sell less than 60% of their exports to such markets.
This micro-fact shows up in the aggregate – as highlighted in Figure 3, aggregate exports of high-end variety exporters are more geographically diversified than exports of low-end variety exporters. The difference holds true if we exclude Colbert firms from our computation.
A tiny elasticity of exports to distance explains the peculiar geography of high-end exports
The difference in geography may have two explanations:
- First, high-end variety exports might be more sensitive to income, and their sales might thus better reflect the distribution of wealth over the world.
- Second, high-end variety exporters may be less sensitive to distance.
To quantify the relative importance of these potential explanations, we compare the sensitivity of high-end and low-end variety exports to different gravity variables. We find that high-end variety trade flows are one-third more sensitive to the per capita income of the destination country. They are also ten times less sensitive to distance. Actually, the effect of distance on high-end exports is almost negligible. We show that this finding cannot be accounted for by alternative explanations such as productivity differences between high- and low-end variety exporters. Based on these estimates, we can then compute the hypothetical geography of low-end variety exports had they the same sensitivity to distance and/or income as high-end variety exports. We show that the lower sensitivity to distance is entirely responsible for the greater geographic diversification of high-end variety exports.2
Macro implication 1: Volatility
Our findings suggest that the impact of specialising in high-end varieties on the aggregate volatility of exports involves two opposing effects. On the one hand, high-end variety exports are more sensitive to income changes. This tends to increase their volatility (consistent with what Berthou and Emlinger 2009 find during the 2008 crisis, for example). On the other hand, high-end exports are more geographically diversified. As long as countries are not hit by a single common shock, this is likely to reduce volatility through a portfolio effect.
One simple way to illustrate these two forces is to compare the country-specific volatilities of high- and low-end variety exports with their aggregate (cross-country) volatility. Over the period 2000–06, we find that high-end variety exports to a country are on average 40% more volatile than low-end variety exports. This is consistent with a higher sensitivity to country-specific income shocks. However, if we look at total exports, both high- and low-end variety exporters face the same level of volatility. Decomposing volatility into country-specific volatility and cross-country diversification, we show that the volatility of high-end exports is significantly dampened by its geographic diversification. More precisely, five-sixths of the effect comes from the fact that high-end exporters export to countries with relatively less positively correlated shocks.
Macro implication 2: Growth
In a final exercise, we show that high-end exporters are more likely to redirect their sales toward fast-growing economies. The correlation we observe between the income growth in a given country and the share of this country in overall exports is positive for both high- and low-end variety exporters, but stronger for the former. Importantly, we find that this stronger positive correlation is entirely driven by their lower sensitivity to distance. Put differently, high- and low-end variety exporters have the same ability to reap the gains from growth in countries neighbouring France. However, if the sources of growth are in distant markets – as has been the case over the last decade with the growing importance of East Asia in world demand – high-end variety firms are better equipped to meet demand there.
Berthou, A, and C Emlinger (2010), “Crises and the Collapse of World Trade: the Shift to Lower Quality”, CEPII Research Center Working Paper 2010-07.
Fontagné L, Gaulier G, Zignago S (2008), “Quality matters: Everything is (not) made in China”, VoxEU.org, 28 March.
Martin, J and F Mayneris (2013), “High-End Variety Exporters Defying Distance: Micro Facts and Macroeconomic Implications”, Paris School of Economics G-MonD Working Paper 35, October.
Schott P, (2007), “How does China compete with developed countries?”, VoxEU.org, 10 October.
Verhoogen, E A (2008), “Trade, Quality Upgrading, and Wage Inequality in the Mexican Manufacturing Sector”, Quarterly Journal of Economics, 123(2): 489–530.
1 Note that we underestimate the weight of high-end varieties since we cannot identify, due to data issues, high-end exporters appearing after 2006 in the data. The 198 products we consider (in the champagne, cosmetics, jewelry, and apparel industries, among others) represent around 10% of total French exports.
2 Including income inequality in the analysis does not change the results.