French exporters and the global crisis

Lionel Fontagné, Guillaume Gaulier, 27 November 2009

a

A

The last quarter of 2008 and the first quarter of 2009 witnessed a sudden, severe, and synchronised drop in world trade (Baldwin and Evenett 2009). Both industrial production and trade fell faster in 2008-2009 than during the Great Depression (Eichengreen and O’Rourke 2009). The annual volume of world trade in 2009 will drop 9%.

In this chapter, we explore the mechanisms that led to such a dramatic collapse in exports using firm-level data on French exporters. The key questions are:

  • Have different firms been differently affected by the crisis, based on their size, their degree of globalisation, or their access to external financing?
  • Has the sectoral and geographic composition of firms’ exports played a role in a trade collapse?

Channels of transmissions of the financial crisis to the real economy

The transmission channels in the real economy are key to understanding this episode. In addition to the synchronised nature of the reduced activity in the OECD, the national economic policies adopted to cope with this depression may have played a role.

  • First, fiscal stimulus packages have been oriented towards non-tradables such as construction and infrastructure, with the exception of fiscal incentives for domestic purchase of new cars.
  • Second, while protectionist tensions have been controlled by the big players in world trade, there is evidence of some murky protectionism.
  • Third, the observed increase in the long-run income elasticity of trade may also be playing a part.

This change can be explained by globalisation in general and more specifically by the fragmentation of supply chains whereby the same component is traded several times before being included in the final product. But if global value chains amplify value added, they do not explain the change in the value of trade. Fragmented supply chains may be consistent with world trade reacting proportionally to a fall in world GDP (Bénassy-Quéré et al 2009). Only a composition effect, where changes in trade marginally fall on the more fragmented sectors can generate a more than proportional reaction of trade to a drop in GDP.

On this point, the sectoral evidence is contradictory. Equipment and capital goods, the car industry, and intermediate goods have been particularly affected by a combination of the inventory cycle, a credit crisis detrimental to durable goods, and a confidence crisis leading to postponement of purchases. Some of the trade crisis ultimately is attributable to a crisis in credit markets that caused investments to fall more than in previous economic cycles.

Fewer exporters or reduced export sales?

The precise significance of such a decline in the value of exports in terms of the extensive and intensive margins of trade is unknown. Do we have fewer firms, exporting to fewer markets (extensive margin) or do we have the same number of firms exporting lower values in all their markets (intensive margin)?

The intuition might suggest that the smallest and most fragile exporters have been pushed out of the market, while the larger and more diversified firms have taken advantage of their size and market power to adjust and eventually pass part of the burden onto suppliers and wholesalers. Also, we would expect that the restriction on trade credit would affect small firms’ exports first. Other dimensions of the dynamics of trade need to be examined. Churning (i.e. rapid change in firm’s export status) is a common feature of the firm-level data. Hence a reduction in the number of exporters could be the result of an increased number of exits, fewer entries, or a combination of the two.

To address these issues, we exploit a data set of individual exporters located in France. Such data enable two types of investigation: a precise description of the characteristics of individual exporters (turnover, employment, productivity, profitability, etc.) relying on information from the best documented firms, which are also the largest ones, and an examination of the dynamics of the distribution of exporters, based on the whole universe of exporters, observing their individual contributions to the value or diversity of exports in the sector to which they belong. Since we are interested in a precise description of the extensive and intensive margin, we adopt the latter approach.

The distribution of French exporters

We start by characterising the distribution of French exporters. France is broadly similar to other countries in that exporting is limited to a very select club of “champions”, flanked by a large number of marginal competitors exporting on an irregular basis (Mayer and Ottaviano 2007).

To illustrate this in the French case, consider the largest exporters (1%) in each sector, using the HS 2 digit system for classification of exported products into 98 sectors. Accordingly, we use the criterion of total value of a firm’s exports relative to the exports of all other firms exporting in the same sector.

We find that the top 1% of firms represents 63% of total French exports. The next four percentiles represent 24% of total exports. The smallest exporters (80%) represent only 3% of the total value. In a nutshell, some 1,000 individual exporters represent two-thirds of the total exports of the 5th largest world exporter.

The second main characteristic, directly linked to the very large presence of small exporters in the distribution, is churning (numerous entries and exits). Since not every firm exports every month, we look at annual export activity. Some 20,000 exporters enter each year, and as many exit. Thus, over ten years, it is possible to observe 300,000 different exporters, with a maximum of 100,000 each year, and a maximum of 50,000 each month. What is the dynamics related to these switchers? Only 35% of entries correspond to firms that are observable the following year’s statistics. But three years after entry 20% of entrants have survived. Hence, the survival rate increases very fast after a very low start in the first year.

Crisis impact on individual firms – the theory

Against this empirical evidence, the expected impact of the crisis on individual exporters is uncertain.

The very large presence (in terms of numbers) of small exporters may well lead to the decimation of this group. However, the large concentration of exports (in terms of value) should lead the largest firms to contribute to the overall decline more or less in proportion of their overwhelming presence in total exports. All in all, a story of numerous small exporters exiting the market and the large ones contributing to a large share of the observed drop is what we would expect.

The policy implication is clear. Such an outcome would have long-lasting, detrimental effects on exports. Small firms, often entrants, could be expected to grow rapidly in the future as their probability of death would fall greatly as time passes. From today’s small players the market will sort out the champions of tomorrow. If these small players are massively hurt by the crisis, for instance due to finance drying up, then the crisis will leave its footprint on export performance for many years to come.

Crisis impact on individual firms – the facts

However, if we look at the data, we see something quite different. Detailed data on French exporters during the turmoil (Bricongne et al. 2009) show that the number of exporters has been reduced only slightly by the crisis, while the value of total exports has sunk greatly. The extensive margin of trade has only slightly contributed to the drop in French exports, the bulk of the observed decline affecting the intensive margin and, more precisely, the drop in the value exported by the top 1% of exporters.

This is illustrated in Figure 1. From 2004 to the end of 2007, the value of French exports has increased regularly notwithstanding the downward trend in the number of exporters. During the full semester of turmoil, i.e. from the last quarter 2007 to the first quarter of 2009, the value of French exports reduced dramatically, while the reduction in the number of exporters was more or less in line with the pace observed since 2000. Our first conclusion then is that most of the adjustment has taken place at the intensive margin, via a reduction in the value exported.

Figure 1. Total value of French exports and total number of French exporters, 2000-M1 to 2009-M4

Note: Chapters 98 and 99 of the HS2 are dropped. 3-months moving averages. Left scale: million euros. Right scale, number of exporters in monthly data.

Source: French customs data, Bricongne et al (2009).

However, we cannot simply count the number of exporters. In a multi-market settings – where exporters start by serving the most profitable markets and export up to the marginal market where entry costs are only just covered by local sales – a violent shock in world demand should lead exporters to refocus on their most profitable markets. The pertinent extensive margin of trade in this case would be the elementary market (a firm exporting to a destination market within the same sector).

There are about 95,000 individual French firms exporting at least once a year, but only 50,000 exporting firms in the monthly data (not all exporters export each month as already stressed). Monthly exports by destination and product category are observed for the period January 2000 to April 2009. There is considerable seasonality in these data and the number of working days is also an important determinant of monthly exports. Accordingly, we apply the coefficient of adjustment used by the French customs to broad categories of products, and focus on year-on-year variations (month m of year t is compared to the same month of year t-1).

With monthly data, churning is amplified and consequently the calculation of growth rates is more difficult. Indeed, at the detailed level of a firm exporting to a market in the same sector, in a given month, calculating growth rates by considering only flows observed in two subsequent months would lead to a large and undesirable selection of flows.

To get around such problems, we rely on the so-called mid-point growth rates. With this method, elementary trade flows in a sector each month can be classified into four types:

  • Created (positive extensive margin),
  • Destroyed (negative extensive margin),
  • Increased (positive intensive margin), and
  • Decreased (negative intensive margin).

Ultimately, we need to match two dimensions of analysis – distribution of exporters by size, and trade margins at the elementary flow level.

With these tools in hand, we investigate the causes of French export failure.

The microeconomic causes of export engine failure

The worse month in our sample was February 2009, with a recorded 27.5% year-on-year drop in exports. At most, 20% of this drop is due to missing flows, that is, to exporters having stopped serving at least one destination market in a given sector, or due to lack of entry of new exporters.

Decomposing these latter effects we observe that entries have not so much been reduced, but that the key explanation is lack of additional exits, compared to the “good times”. The remaining and main part of the drop is accordingly due to the intensive margin, with a dominant contribution of the top 1% of exporters. More than a third of this deterioration is attributable to trade in intermediate goods. Equipment goods (excluding aircraft) and the automobile industry are next two broad contributors. These three sectors overall account for 82% of the champion’s losses, as illustrated in Figure 2.

Figure 2. Contribution of negative growth to the top 1% exporters sales’ growth rates January 2008 to April 2009, by broad sector

Note: Chapters 98 and 99 of the HS2 are dropped. Exporters are ranked according to the value of their exports within a sector.

Source: French customs data, Bricongne et al. (2009) calculations.

Overall, the top 1% of exporters contributes to 67.4% of the drop in sales in existing markets, which is not so very different from their overall contributions to exports. This suggests that the drop in the value shipped by exporters may be quite evenly distributed.

No conditional differences between small and big firms

To check the respective contributions of the geographic and sectoral orientation of exports, and the size of exporters, we can decompose the variance of the drop into monthly exports. While the uncorrected growth rates of exports exhibit large differences between small and large exporters (large exporter being more severely hit), after correcting for the orientation of exports these differences vanish. There is ultimately no difference between small and large firms, with the exception of the black February already referred to, where the largest firms were actually the most affected.

As well as this descriptive analysis, it is useful to perform multivariate regressions to identify the respective roles of the various suspects. We control for sectoral and geographical determinants of the drop in individual French firms’ exports (such as demand in foreign markets, by sector) during the crisis and investigate the impact of the sectoral dependence on external finance. We control whether this impact is robust to controlling for the degree of globalisation in the sector.

We can confirm that the differences in year-on-year growth rates for monthly exports of individual firms (to individual destinations) are not significantly different across size groups.

Three facts are worth highlighting:

  • Firms of all sizes have been affected fairly equally by the crisis when destination country and the sector dimensions of the decline are controlled for.
  • The second empirical fact is that the crisis has hit firms in sectors relying on external finance more severely, irrelevant of firm size.
  • The third piece of evidence is that this effect of external finance is robust if we control for the fact that more globalised sectors (where intermediate imports account for a larger share of the value of final production), have also been hit more severely.

Conclusion and implications for the future

Detailed evidence on what has happened to individual exporters during the turmoil suggests that most of the adjustment has taken place at the intensive margin, through a reduction in the value of existing flows. Clearly, the exit of exporters from the market is not the explanation. If we differentiate among exporters of different sizes, it appears that the largest ones have been the most affected. This unexpected outcome is due to the presence of the largest exporters in the most affected sectors and in the most affected destination markets. If we take this into account, then all firms have been evenly hit.

This result has several implications.

  • First, our results cast doubt on the case for the withdrawal of trade credit hurting the smallest players first.
  • Second it suggests that being more globalised may not be an advantage in the context of a global crisis.

The big players are typically more global and this has not helped them. On the contrary, firms belonging to more globalised sectors, other things been equal, have suffered more.

  • Third, the blow to exporters may not have long lasting effects on aggregate export capacity, since the reservoir of small and promising firms has not been decimated by the crisis.

Our micro-analysis produces some less good news. The other side of the coin in terms of the evolution observed, may be that the bulk of exits has been delayed and may appear later, this time concentrated on the smallest exporters. Given the delay between the decision to enter and first export, some entries may materialise. Also, sectors dependent on external finance have been more severely hit.

Were this result to be confirmed by other studies, it would mean that, at some point, the financial sector stopped fulfilling its traditional role. This would produce long lasting effects. Firms, and especially large firms, may be tempted to reorganise their activity to get rid of future problems, and the future growth of international supply chains will be threatened. While the trade crisis is not the result of de-globalisation, the role of external finance in the differentiated impact of the global crisis on individual sectors may lead to industrial strategies departing from the pre-crisis global factory.

References

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

Eichengreen, Barry, O’Rourke Kevin H., 2009 A Tale of Two Depressions. Vox.

Bénassy-Quéré Agnès, Decreux Yvan, Fontagné Lionel, Koudour-Casteras David, 2009, Economic Crisis and Global Supply Chains, Document de travail CEPII, 2009-15.

Mayer, Thierry, Ottaviano, Gianmarco, 2007, The Happy Few: The internationalisation of European firms, Bruegel Blue print series, 3.

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

Topics: International trade
Tags: firm-level trade data, great trade collapse

Professor of economics in the Paris School of Economics, Université Paris I Panthéon Sorbonne
Economist at the Banque de France