VoxEU Column International trade

High quality imports suffer more during recessions

The volume of world trade has plummeted with the global crisis. This column says that high-quality imports are more responsive to income changes than low-quality imports. This explains why world trade value fell faster during the crisis than world trade volume, which fell faster than GDP.

The volume of world trade plummeted between the last quarter of 2008 and the first quarter of 2009. Recent forecasts by the IMF (2009) predict a reduction of world trade volume by 11% (and value by 23%) for the year 2009. This is by far larger than the 1.3% contraction of the world GDP. The CEPR book edited by Simon Evenett and Richard Baldwin suggests that the availability of trade finance, the increase of protection, and the contraction of inventories contribute to explain the disconnection world trade volumes and the world GDP during the crisis this year. Caroline Freund (2009) also argues that world trade volume is more responsive to GDP variations during global downturns, suggesting that all the above-mentioned hurdles may be specific to bad times.

Another surprising pattern is the decrease of import price indices, associated with the disconnection between import volumes and import values. Joseph Francois and Julia Woerz (2009) argue that part of the decrease in import prices can be explained by the reduction of world commodity prices. But the contraction of markups and individual prices or the selection of varieties according to their price may have also contributed to the decrease of import price indices. Indeed, tougher competition following the contraction of income and demand may lead firms to reduce their profit margins. Also, consumers may shift their demand to low-price varieties in times of recession. Hence, the price evolution within each individual variety and the selection between varieties is responsible for aggregate price variations.

New evidence: A bigger drop for high-quality imports

In an upcoming paper, we test whether variations of income can affect differently the imports of varieties differentiated by their price, therefore contributing to the disconnection between volumes and values when income falls. Theoretical models by Bils and Klenow (2001) and Fajgelbaum et al. (2009) indeed suggest that variations of income are not only associated with variations in the quantity of goods consumed but also variations in the quality of those goods. Households prefer more expensive varieties as they become richer. This implies that a global recession is expected to decrease the demand for high-quality varieties.

Empirically, price differentiation within detailed product categories is usually considered a proxy for quality differentiation (Baldwin and Harrigan 2009, Fontagné et al. 2008, Schott 2008, amongst others). Accordingly, we use international trade data with the greatest level of product disaggregation, provided by the UN, to classify varieties as low or high quality.2 More precisely, on each individual market, varieties with an import price above the average import price observed on the market are classified as high-quality varieties, and those varieties with below-average import prices are classified as low quality.

We use this classification to compute the recent evolution of low and high quality import quantities for the EU (15). The monthly data are by Eurostat Comext are highly disaggregated, yielding some 8,000 product categories.3 Figure 1 reports the evolution of the import quantities for high- and low-quality varieties. The curves show that the decrease of the quantity imported has been larger for high-quality varieties (-23%) than low-quality varieties (-17%), between March 2008 and March 2009. The disconnection between high- and low-quality varieties was more pronounced during the fourth quarter of 2008. This helps explain why import values have decreased more than import volumes or quantity – demand shifted away from more expensive varieties more.

Figure 1. Evolution of import quantity by quality in the EU15

High-quality imports are more responsive to GDP variations

We complete the analysis by estimating the elasticity of import quantity with respect to variations of the GDP for low- and high-quality ladders. We use import quantity and unit values from the BACI database (CEPII) that covers the period 1995-2007. GDP and nominal exchange rates are provided by the World Development Indicators database. Import demand equations are estimated separately for each HS4 level of product disaggregation for 184 exporting and importing countries. We report the median of income elasticities in Figure 2, considering all coefficients.

The results show that a decrease in aggregate income (measured by GDP) reduces the quantity of goods imported (positive elasticity). However, the income elasticity differs significantly across low- and high-quality varieties. The income elasticity is almost 60% higher for high-quality varieties. These results are obtained by considering all countries, but they remain qualitatively similar when the elasticity is estimated for only OECD destinations or emerging economies destinations.

Figure 2. Median estimated income elasticity

The median income elasticity is always above unity for both high-quality (expensive) varieties and low-quality (cheap) varieties. This implies that variations in import quantity are always larger than variations in GDP. This matches recent experience, as trade volume fell much more (-11%) than world production (-1.3%) during the crisis. Applying our estimated elasticity to this decrease in world production, we find that the predicted contraction of world GDP in 2009 should lead to a 1.6% decrease of the import quantity for low-quality varieties, and a 2.6% decrease of import quantity for high-quality varieties.

Figure 2 reports that the income elasticity is larger for OECD destinations compared to emerging markets. The recession has also been worse for advanced economies. Taken together, these facts explain the recent collapse of world trade – countries with more elastic import demand have had the largest recessions. The example of Japan is suggestive. The IMF predicts a -6.2% recession for Japan in 2009, one of the most severe reductions of GDP among advanced economies (IMF 2009). Using the quality distribution of exports for each exporter on the Japanese market and the income elasticity for OECD destinations, we can predict the variation of imports from each country. Results are reported in Figure 3 for a selection of exporters. These predictions show that, due to quality differentiation, the difference between the reaction of Chinese export quantities and export quantities from the most advanced economies may reach more than 2%. The difference in the reaction of export quantity between countries specialised in low and high quality products can explain the decrease of import prices during the crisis.

Figure 3. Predicted change in exports following a 6.2% GDP decrease in Japan, by destination


We show that high-quality imports are more responsive to GDP variations than low-quality imports. This empirical pattern can explain the behaviour of world trade volume, world trade value, and import prices during the last crisis. Our findings also imply that high-quality imports should benefit more from the recovery, due to their larger income elasticity. In other words, countries specialised in high quality exports are expected to suffer more in times of crises but should also recover faster. This has important implications regarding policies in rich countries that promote specialisation in higher-quality product ladders. While this form of specialisation can isolate part of domestic production from competition from the South, it also implies greater volatility, as such exports are more responsive to the world business cycle.


1 This is what we find using the CPB Netherlands Bureau for Economic Policy Analysis

2 We used tariff-line data provided by the UN for the year 2005. In these data, product disaggregation varies according to the market that is considered (between HS6 and HS10 categories).

3 Eurostat uses a Combined Nomenclature at 8 digits.


Baldwin, Richard and James Harrigan (2009), “Zeros, Quality and Space: Trade theory and trade evidence”, mimeo.

Berthou, Antoine and Charlotte Emlinger (2009), “Crisis, trade collapse, and the decrease of import prices”, mimeo.

Bils, Mark and Peter Klenow (2001), “Quantifying quality growth”, The American Economic Review 91 (4): 1006-1030.

Evenett, Simon and Richard Baldwin, “The collapse of global trade, murky protectionism, and the crisis: Recommendations for the G20”, CEPR, 2009.

Fajgelbaum, Pablo, Gene M. Grossman and Elhanan Helpman, “Income distribution, product quality and international trade”, mimeo.

Fontagné, Lionel, Guillaume Gaulier and Soledad Zignago (2008), “Specialization Across Varieties and North–South Competition”, Economic Policy, 23 (53): 51-91.

Francois, Joseph and Julia Woerz (2009), “The big drop: Trade and the Great Recession”, VoxEU.org, 2 May.

Freund, Caroline (2008), “Demystifying the collapse in trade”, VoxEU.org, 3 July.

IMF (2009), World Economic Outlook, April.

Schott, Peter (2008), “The Relative Sophistication of Chinese Exports Economic Policy 53:5-49.

315 Reads