Resilient to the crisis? Global supply chains and trade flows

Carlo Altomonte, Gianmarco Ottaviano

27 November 2009



According to the most recent IMF estimates (IMF 2009), the ongoing recovery will drive a wedge between output and trade. Output is supposed to shrink by ‘only’ 1.1% at the end of 2009 (-3.4% in advanced economies), but world trade is forecast to still experience a drop of -11.9%. While other estimates put the latter figure at –9% (WTO, World Bank), it is indisputable that during 2009 official figures recording trade flows will fall much more than GDP.

Apart from its magnitude, the fall in trade in 2009 has also been quite homogeneous across all countries (more than 90% of OECD countries have exhibited simultaneously a decline in exports and imports exceeding 10%, as noted by Araujo and Olivera Martins 2009). This fall has also been very fast, with trade virtually grinding to a halt in the last month of 2008.1 These facts led Baldwin and Evenett (2009) to qualify the drop in trade during the crisis as “severe, sudden and synchronised”.

It’s the global supply chain, stupid! Or is it?

A number of transmission mechanisms have recently been proposed to account for these three attributes of the contraction of trade flows, many of which impinge upon the role that global supply chains might have played in exacerbating the drop in global demand.

The basic argument is that in a world characterised increasingly by vertical specialisation, goods are produced sequentially in stages across different countries – so-called international supply chains. The constituent parts and components of a final good crosses borders several times before the final product reaches the consumer; at each border crossing, the full value of the partially assembled good is recorded as trade. As a result, for a given reduction in world income, trade should decline “not only by the value of the finished product, but also by the value of all the intermediate trade flows that went into creating it”.

O’Rourke (2009), with his Barbie-doll example, has been the first to doubt whether, as a result of fragmentation, changes in world trade should necessarily outweigh changes in world GDP.2 Even if the Barbie parts cross the border twice in the production of a final doll that sells for $20 in the US, the final sales and total trade should contract by the same percentage; a 50% drop in US Barbie sales reduces world Barbie trade by 50%.

More recently, Fontagnè et al. (2009a) have provided a more structured analysis confirming the insight of the Barbie-doll example. First of all, they give a very simple accounting example showing that, if relative prices are held constant, fragmented trade flows within global supply chains should react proportionally to a fall in world GDP. Then they validate and generalise this finding via a simulation based on a multi-country, multi-sector CGE model. Their simulation shows that, if all trade flows are deflated by their specific prices (rather than the world GDP deflater) and GDP flows are aggregated at the world level using current exchange rates (as done for trade flows) rather than PPIs, the measured drops in trade volumes and GDPs converge to roughly comparable values, -2.4% and -2.6% respectively.

This implies that the extensive presence of supply chains does not automatically explain why world trade overshot the world GDP drop; other explanatory factors are needed. These may include:

  • The collapse in internal demand and production, affecting current and future level of (tradable) inventories worldwide;
  • Fiscal stimulus plans with a relatively stronger support of non-tradable sectors, like construction and infrastructures (Bénassy-Quéré et al. 2009);
  • The rise of ‘murky’ protectionism; and
  • The problems of trade finance with financial spreads still well-above ‘normal’ (i.e. pre-crisis) market rates (Auboin, 2009).

Trade finance and liquidity constraints

Do the above arguments mean that global supply chains are totally neutral as a transmission mechanism of the crisis from GDP to trade? Of course not. In all likelihood, however, the channels are much more complex than originally thought, and entail important compositional effects.

For the sake of argument, let us take the following story based on the idea that a relatively large part of the overreaction of trade has been caused by the sudden drying up of liquidity in trade finance. Auboin (2009) notes that, in the second part of 2008, spreads on short-term trade credit facilities suddenly soared to between 300 to 600 basis points above LIBOR, compared to 10 to 20 basis points in normal times, leading to a virtual freeze of important trade deals throughout the globe, with supply chain operations being disrupted by lack of financing, especially for developing country suppliers.

Under this assumption we would have a scenario in which the liquidity channel has led trade to overshoot the fall in demand, with the effect being larger within supply chains, as the trade financing of these operations is typically managed by large international financial institutions, particularly hit by the crisis.3

In this scenario, we would still obtain a severe, sudden and synchronised drop in trade flows, with the effects correlated with (but not caused by) the behaviour of global supply chains.

Moreover, under the same scenario, we would also observe that, during the crisis,trade falls more along the intensive margin (i.e. value per trade) than the extensive margins (i.e. number of traders). The reason being that, if the overreaction of trade was caused relatively more by liquidity constraints than by a disruption of supply chains, the above effects would lead to a reduction in the volume of trade, but not necessarily to a similar reduction in the number of traders worldwide.

This is exactly what Bricongne et al. (2009) find in a paper analysing the behaviour of French exporters during the crisis. Relying on monthly data for individual French exporters observed until April 2009, the authors find that the drop in French exports is mainly due to the intensive margin of large exporters, with small and large firms evenly affected once sectoral and geographical specialisation are controlled for. Interestingly, they also find that firms (small and large) in sectors more dependent on external finance are the most affected by the crisis.

Long-lasting relations

Equally plausible stories suggest that trade flows within supply chains are more, rather than less, resilient to large adverse shocks like the current crisis. Such resilience would derive from the fact that setting up organised supply chains entails some sunk costs, so firms would prefer to adjust the entire chain along the intensive margin (i.e. reducing volumes), rather than the extensive margin (i.e. disrupting part of the supply chain).

Moreover, even if some adjustment along the extensive margin has to be made (e.g. by dropping some suppliers), it could well be that some long-term contractual relationships within supply chains are more difficult to sever in the short run. Finally, it is also possible that large multinational corporations at the centre of several supply chains could alleviate the liquidity constraints of suppliers, thus protecting the entire supply chain from external finance shortages.

Although a precise distinction between general trade flows and those flows happening within supply chains is difficult to make, on the basis of the available macro data, the foregoing considerations are consistent with two pieces of evidence observed in US and European data.

New evidence

In Europe, the process of east-west integration has triggered the emergence of international networks of production involving, in particular, German and Italian companies investing in the new member states of Central and Eastern Europe (CEECs).4

Figure 1 looks at the most recent trade data on average year-on-year monthly growth rates for the four biggest European economies (France, Germany, Italy and the UK), which are also the biggest exporters to the CEECs. At the world level, all these countries have experienced negative growth rates in their total exports, with little differences among them (average monthly rates ranged from -12% and -15% from July 2008 to July 2009.).

However, when looking at the trade flows with the CEECs, the figure shows that, until March 2009 (the worst moment of the crisis), Italian and German exports had fallen much less than those of France and the UK. Since trade between Germany, Italy and the CEECs takes place within supply chains to a larger extent than that of France and the UK, one may find here an indirect confirmation of the resilience of supply-chain-related trade flows during the crisis.

Clearly, resilience does not necessarily mean that trade within these international value-added chains is insulated from the crisis (differences between German and Italian trade flows on one hand, and French and UK ones on the other, have recently disappeared). The supply chain trade, however, might have reacted later to the shock.

Figure 1 Growth rates in exports to Central and Eastern Europe.

Source: authors’ calculations on Eurostat data.

Along the same lines, Bernard et al. (2009) analyse the behaviour of US exports at the time of the Asian crisis, when trade slumped rapidly and trade finance was also severely hit. They show that, overall, US exports to Asia declined by 21% between 1996 and 1998, while exports to the rest of the world increased by 3%. Within Asia, however, the decline in arm’s length exports was substantially greater than the drop of trade undertaken within supply chains (-26% versus -4% by 1998), while two figures evolved in a similar way in the case of exports to the rest of the world. This is again evidence consistent with the idea that, in a crisis context, trade undertaken within supply chains does not necessarily overreact to a drop in demand, but rather exhibits some degree of resilience.

Special interests and sheer luck

While any conclusion must wait for more data to become available, there are good reasons to believe that the rise of global supply chains has not necessarily been the main cause of the recent “severe, sudden and synchronised” fall in global trade flows. Based on the available evidence, one may even be tempted to conclude that, under certain circumstances, international networks of production may also display some degree of ‘resilience’ to adverse shocks like the current crisis: supply-chain-related trade flows may react later (rather than sooner) to an adverse shock. Their fall may be smaller and, eventually, their recovery may happen faster relative to overall trade flows.

The observed resilience of supply chains may arise from some intrinsic attribute of production chains, as argued above. Alternatively, it may be the outcome of the political economy. Fearing that a collapse of supply chains would set off a sudden process of de-globalisation and implosion of international trade, governments may intervene in favour of supply chains. For example, the massive bail-outs of large financial institutions have helped their best customers, among them the big players within supply chains. Finally, of course, this indirect support of supply chains may have also been an unintended consequence of financial bailouts implemented for very different reasons.

De-regulation vs re-globalisation

There are too many blind spots in our current understanding of the nature and operation of international supply chains. Once data become available, the current crisis should give us material to substantially improve our understanding. In the end, it may well be that we discover that without supply chains, things would have been much worse than they actually were; and that this crisis may eventually boost rather than cripple the globalisation process.


1 The annualised rates of growth between October and December 2008 were -43% for the US, -81% for Germany, -38% for China (Yi, 2009), with an OECD average negative growth rate between October 2008 and March 2009 of -21% (Araujo and Oliveira Martins 2009).

2 Kevin O’Rourke makes this point is a famous blog entry that uses the example of a Barbie doll.

3 Moreover in years before the crisis some supply chains had abandoned the traditional instruments of letters of credit, preferring to regulate transactions directly through an open account balance, given the abundance of liquidity. With the financial crisis, and the ensuing drying up of liquidity, those supply chains had to rely again on traditional instruments of trade finance (e.g. letters of credit), but at much higher costs.

4 Germany and Italy are the two largest investors in Central and Eastern Europe, and their trade flows with the area, contrary to other EU countries, are mainly driven by trade in intermediates.


Araujo, Sonia, Joaquim Oliveira Martins. (2009). “The Great Synchronisation: What do high-frequency statistics tell us about the trade collapse?” Vox, 8 July

Auboin, Marc. (2009). “The challenges of trade financing.” Vox, 28 January.

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

Bénassy-Quéré, Agnès, Yvan Decreux, Lionel Fontagné, and David Khoudour-Casteras. (2009). “Economic Crisis and Global Supply Chains”, CEPII Discussion paper No. 2009-15, Paris.

Bernard, Andrew, J. Bradford Jensen, Stephen J. Redding and Peter K. Schott. (2009). “The Margins of US Trade.” American Economic Review Papers and Proceedings 99: 487-93.

Bricongne, Jean-Charles, Lionel Fontagné, Guillaume Gaulier, Daria Taglioni and Vincent Vicard. (2009). “Firms and the global crisis: French exports in the turmoil.”, mimeo.

IMF (2009). World Economic Outlook, October.

O’Rourke, Kevin. (2009). “Collapsing trade in a Barbie world”, blog post, Irish Economy

Yi, Kei-Mu. (2009). “The collapse of global trade: the role of vertical specialization”, in Baldwin and Evenett (eds.), The collapse of global trade, murky protectionism, and the crisis: Recommendations for the G20, Ed. Vox, CEPR, London.



Topics:  International trade

Tags:  great trade collapse, international supply chains

Associate Professor of Economics, Bocconi University and Visiting Fellow, Bruegel

Professor of Economics, LSE; Non-Resident Senior Fellow, Bruegel; and CEPR Research Fellow