Trade collapse or trade crisis?
Kristian Behrens Gregory Corcos Giordano Mion
21 March 2010
World trade fell dramatically during 2009, as widely documented on this site and elsewhere. But there has been little econometric analysis of the different explanations put forward. This column uses data from Belgium to argue that a fall in demand was the main culprit. It is not a trade crisis – it is a trade collapse.
World trade in manufactures fell by about 30% between the first half of 2008 and the first half of 2009 (WTO 2009), with remarkable synchronisation within the OECD (Araújo and Martins 2009). This trade collapse exceeded the fall in GDP as well as the trade fall predicted by computable general equilibrium models (e.g. Benassy-Quéré et al., 2009) or international real business cycle models (as in Levchenko et al. 2009).
Explanations abound, as illustrated by the comprehensive discussion in Baldwin (2009). Some emphasise a crisis of the supply-side of trade citing such causes as a shortage of trade finance, disruption of global value chains, and increased trade barriers. For others, this fall is simply the flip-side of a fall in the demand for manufactures, postponed purchases of intermediates, and the drawing down of inventories.
Plugging the gap with dataWho is right? Aggregate and broad product categories level analyses are not well suited to provide a clear answer because the different effects and margins of adjustment cannot be separately identified. Econometric analysis of firm-level data is therefore critical to discriminate between these explanations. Yet despite a wealth of statistical analysis, econometric work on firm-level data is scarce.1 Our recent research (Behrens et al., 2010) tries to fill this gap, using data on firm-product-country level Belgian exports and imports as well as a wealth of balance sheet information. We compare the first half of 2008 and the first half of 2009 as the Belgian trade collapse started in November 2008.
Three key findings stand out.
We conclude that the trade collapse did not result from a crisis of trade itself. We can presume that, although exports and imports in Belgium are still suffering to recover, as Schott (2009) puts it: “trade will bounce back relatively quickly once conditions improve”.
- First, the fall in trade overwhelmingly occurred at the intensive margin, i.e., prices charged and quantities sold.
- Second, the absence of composition effects suggests that the drying-up of trade finance, the breaking-up of global value chains, or the drawing-down of inventories were not the main drivers of the fall in trade.
- Third, changes in exports and imports did not systematically deviate from changes in turnover and intermediate consumption respectively.
A fall at the intensive margin of tradeWhich margins mattered most for the trade collapse, firm entry or exit? Adding or dropping products and markets? Price adjustments? Or output scaling? The answer to this question is important. Changes at the intensive margin (price and quantity adjustments) are likely to be less durable and less costly than changes at the extensive margin (i.e., changes in the number of trading firms, products sold, and markets served). This is because the latter entail high irreversible costs, such as those of creating a trading network across different countries.
Table 1 shows that virtually all of the trade collapse during the 2008-2009 crisis was driven by changes in the prices quoted and the quantities shipped. Though both exports and imports fell by about 24%, almost all trading firms remained active, with hardly any change in the average number of countries they traded with, and in the average number of products shipped to or sourced from each country. On the one hand, this result echoes findings on the 1997 Asian crisis and confirms evidence on comparable French data on the current crisis (see Bernard et al. 2009 and Bricongne et al. 2009). On the other hand it highlights the extreme flexibility of business relationships across firms, their input suppliers, and their clients. This is reassuring. A massive reduction in the number of trading firms, countries or products would likely make recovery more costly and sluggish.
Table 1. Changes in the margins of Belgian trade, first semesters of 2008 and 2009
Hardly any systematic differences within or between industriesWe decompose the fall in trade into firm, country and product components, and relate each component to observable firm, industry, country and product characteristics. If trade falls more for certain firms, products or countries, the “composition effects” inform us on the likely determinants.
Firm characteristics include size and productivity, export and import shares, a vertical specialisation index, variables linked to financial structure, ownership indicators, as well as a proxy for inventory capacity. Country characteristics include exchange rate movements, the trading partner's growth rate, as well as EU and OECD dummies.
Product variables follow a product classification as well as Rauch's (1999) measure of product differentiation (see Behrens et al. 2010 for a discussion of the methodology and data).
Our econometric analysis provides a number of striking results. The intensive margin fall has been extremely evenly spread across products, including consumer durables and capital goods. The same applies to destination and origin markets, though the fall has been slightly less pronounced for EU partners and/or higher-growth countries.
Finally, no firm composition effects appear on the export side. On the import side, larger and/or more indebted firms (especially with trade credit as opposed to financial credit), and firms relying on a larger share of imported intermediate inputs, reduced imports a bit more. Yet even then firm characteristics explain less than 25% of the change in imports. Furthermore, cross-industry heterogeneity is hardly relevant for exports and completely irrelevant for imports.
A comparable fall in domestic operationsFinally, we consider changes in exports-to-turnover and imports-to-intermediates ratios at the firm level. This allows us to compare the magnitude of the fall in domestic activity with that of the fall in international trade. On aggregate, the ratios of exports and importer over production did not fall during the period we consider (as can be seen from Figures 1 and 2) so confirming the evidence provided in Eaton et al. (2009) for most OECD countries. Crucially, our micro analysis reveals no cross-industry pattern, while firm characteristics have almost no explanatory power. The exports-to-turnover ratio falls more among firms with higher vertical specialisation indices, but by less than 2 percentage points.
Figure1. Monthly export over production value ratio from January 2005 to June 2009
Figure 2. Monthly import over production value ratio from January 2005 to June 2009
ConclusionDetailed micro trade data from Belgium allow for a sharp analysis of the trade collapse. Our econometric study shows that the 25% trade fall was very evenly spread across industries, across firms within industries and across products and countries. Finally, domestic sales and purchases fell equally fast with no systematic variation across firms. We conclude that trade is not in crisis per se and so it would be better to talk about a trade collapse rather than a trade crisis. Further investigation of the fall in demand is certainly needed. Sector biases in fiscal stimuli, a fall in commodity prices, or substitution patterns among consumers may be among the suspects.
ReferencesAraújo, Sónia and Joaquim O Martins (2009), “The Great Synchronisation: tracking the trade collapse with high-frequency data”, in Richard Baldwin (ed.), “The Great Trade Collapse: Causes, Consequences and Prospects”, VoxEU.org.
Baldwin, Richard (ed.) (2009), “The Great Trade Collapse: Causes, Consequences and Prospects”, CEPR.
Behrens, Kristian, Gregory Corcos and Giodarno Mion (2010), “Trade Crisis? What Trade Crisis?”, mimeo LSE.
Bénassy-Quéré Agnès, Yvan Decreux, Lionel Fontagné, and David Koudour-Casteras (2009), “Economic Crisis and Global Supply Chains”, Document de travail CEPII, 2009-15.
Bernard, Andrew, Jensen J Bradford, Stephen Redding and Peter K Schott (2009), “The Margins of US Trade”, American Economic Review Papers and Proceedings
Bricongne, Jean-Charles, Lionel Fontagné, Guillaume Gaulier, Daria Taglioni, and Vincent Vicard (2009), “Firms and the global crisis: French exports in the turmoil”, mimeo CEPII.
Eaton, Jonathan, Sam Kortum, Brent Neiman, and John Romalis (2009), “Trade and the Global Recession”, mimeo.
Levchenko, Andrei, Logan Logan and Linda Tesar (2009), “The collapse of US trade: In search of the smoking gun”, VoxEU.org, 27 November.
Schott, Peter K (2009), “US trade margins during the 2008 crisis”, in Richard Baldwin (ed.), “The Great Trade Collapse: Causes, Consequences and Prospects”, CEPR.
1 To the best of our knowledge, only Bricongne et al. (2009) make use of some firm-level data. They examine the fall in French trade among various classes of exporter size, and various sectors which depend to different degrees on external finance. They do not, however, exploit individual firm characteristics to discriminate between the above-mentioned explanations.