Trade collapse, quality and food exports

Abstract

This article revisits the so-called ‘Collapse in Quality’ hypothesis, according to which, during the 2008–2009 crisis, higher quality goods experienced a stronger export reduction compared to low-quality ones. Using disaggregated trade data from three European countries that traditionally export high-quality food products – France, Italy and Spain – we do not find any econometric evidence supporting this hypothesis. In contrast, we provide preliminary evidence for the concurrent hypothesis, namely that firms reduced their markup to preserve market share.

Keywords:

• 2008–2009 crisis
• trade collapse
• food quality
• markup

JEL Classification:

• F41
• F42
• Q17

Notes

¹ Several possible explanations are put forward in the literature to interpret the trade collapse during the global crisis. Bems et al. (2010) point to the role played by vertical trade, and the subsequent strong reduction of trade in intermediates. Ahn et al. (2011) find evidence that bank credits play a key role in the international transactions. By contrast, Levchenko et al. (2010) proved the existence of a ‘compositional effect’, namely sectors that experienced a larger reduction in domestic output lowered more their trade. Interestingly, although trade protection increased during the recent crisis, it is by far not considered among the main reasons of the trade collapse (Baldwin and Evenett, 2009). See Levchenko et al. (2010) for a review on the main empirical evidence.

² Following Colantone and Crinò (2011), we aggregate these elasticities at the NACE 4-digit level, by taking the median value across all corresponding HS 3-digit products. Before estimating Equation 3, as usual, we drop varieties with unit value that fall below the 5th and above the 95th percentile.

³ Note that, by construction, prices are the sum of the quality and price-adjusted quality components.

⁴ These additional results are available from the authors upon request.

⁵ Markup is computed from Amadeus data. Ideally, we would need data on prices and marginal costs to measure markup. However, marginal costs are difficult to observe. Thus, we estimate markup for firm f in year t, following Chen et al. (2009) and many others, using the information on turnover and variable costs:

⁶ We select France among the three exporting countries, as it is the only one with information on international trade activities in the Amadeus database, allowing us to compare the markup of domestic and exporting firms.