VoxEU Column International trade

US trade margins during the 2008 crisis

If the current “shock” to US trade is similar to previous ones, most of the decline in exports and imports stems from a decline in sales of previously exported goods rather than a decline in the number of products exported. To the extent that is true, trade will bounce back relatively quickly once conditions improve. The alternative view is that the severe credit crunch produced a higher-than-usual share of harder-to-reverse firm exits – potentially dampening the speed of recovery. Even if this did occur, history suggests that it will be concentrated amongst small firms which account for only a small fraction of US exports; US multinationals dominate US trade and these firms have the wherewithal to weather the credit crunch. Should the dollar continue to decline, US firms will broaden the range of products exported and the range of markets reached, putting further downward pressure on the trade deficit.

Since the onset of the current recession in December 2007, seasonally adjusted monthly US merchandise exports and imports have fallen sharply, by 15% and 26% respectively, as of August 2009. As a consequence, the monthly US trade deficit has fallen dramatically during this period, by $29 billion, or 41% of its December 2007 level.1

Figure 1 provides context for these declines by comparing the trends during the current recession with those the US experienced during the last two recessions (which begin July 1990 and March 2001).

Figure 1 US trade during the last three recessions.

Notes: Data are from the US Census Bureau’s website (see text) and are seasonally adjusted. Data are available through August 2009.

While the magnitude of the decline in exports and imports during the current crisis is without precedent, similarities with the previous recessions do exist. For example, a similar sharp proportionate reversal of the US trade deficit also occurred, though more quickly, during the 1990 recession. Then, however, the deficit fell due to a combination of rising exports and falling imports. In this respect, the current crisis looks more like the 2001 recession, where the improvement in the trade deficit was more muted, but where exports and imports declined in tandem.

Table 1 highlights another similarity between the 2001 and 2007 recessions, namely the relatively large contribution of falling commodity imports to the improvement in the trade deficit. In the table, each row summarises the contribution of a particular industry to the overall growth in net exports noted in the final row. Sharp reductions in metals imports – primarily steel (HS 72) – contributed 61% and 73% of the improvements in overall deficit during the 2001 and 2007 recessions, respectively.2

Table 1 Industry contribution towards reducing the US trade deficit, by recession.

Notes: Table records industries' contributions in percentage terms to the change in US net exports across noted in final row across recessions. Positive shares indicate contributions toward increasing net exports. The intervals associated with each recession are 1990.7 to 1991.1, 2001.3 to 2001.12 and 2007.12 to 2009.8). Data are seasonally adjusted by author using the US Census Bureau's X12 method in Eviews and are available on request. See Appendix for analogous breakdown of exports and imports.

Comparison with the Asian crisis

An alternative context for the current decline in US trade – one that focuses on currency devaluation and a severe drop in foreign demand – is the 1997 Asian financial crisis.
Bernard et al (2009) examine the evolution of US trade with five crisis-struck Asian countries (Indonesia, Korea, Malaysia, the Philippines and Thailand) in the months around the 1997 Asian crisis. Seasonally adjusted US net exports fell 50% after the crisis, due to a sharp decline in exports and a moderate increase in imports. The largest export declines occurred among wood, textile and transportation products. Most of the adjustment occurred on the intensive margin.

Another perspective: Extensive versus intensive margins of trade

Another way of decomposing changes in US trade focuses on the degree to which they occur via what has become known as the “extensive” versus “intensive” margins, namely changes in the value of goods that are already imported and exported, and changes in the number of goods exported and imported.

More specifically, the “intensive” margin refers to changes in trade that take place within surviving trade relationships, e.g., the same firm exporting more or less of the same product to the same country. The “extensive” margin, by contrast, tracks changes in trade due to entry and exit, such as a new firm entering the export market, or an existing firm narrowing the range of its export products or destination countries.

Analysis of trade’s extensive and intensive margins increases our understanding of trade patterns and the relative efficiency with which economies allocate resources. To the extent that engaging in international trade requires firms to incur non-recoverable expenses, or develop relationship-specific human capital, high levels of entry and exit may imply greater diversion of resources away from production. On the other hand, relatively fluid entry and exit may reflect re-allocations of resources towards their most efficient use as business conditions change.

A large and growing body of theoretical and empirical work in international trade suggests that trade liberalisation raises aggregate productivity via the extensive margin; as trade costs fall, the least productive firms exit, while the most productive firms expand, and, within surviving firms, the least productive products are dropped. (See, for example, Tybout and Westbrook (1995), Pavcnik (2002), Melitz (2003) and Bernard, Redding and Schott (2009)).

Three lessons

Three lessons emerge from the broader research on the margins of US trade.3

  • First, variation in trade across countries is due mostly to the extensive margin.

The well-known negative “gravity” relationship between trade and distance, for example, is driven almost exclusively by the extensive margin. Both the number of trading firms and the number of traded products decline significantly with distance. If anything, the intensive margin appears to increase with distance.

  • Second, variation in trade across time is dominated by the intensive margin.

This dominance is illustrated in the Appendix Tables, which break down year-to-year changes in total US exports or imports in billions of US dollars (row 10) into the contributions of:

  • Net firm entry and exit (row 3);
  • Net adding and dropping of products and countries (row 6); and
  • The intensive margin (row 9).

As indicated in the final three rows of each panel – which express these contributions in percentage terms – the intensive margin accounts for the largest share of annual trade growth in every year. Across 1993 to 2003, it averages 101% for exports and 114% for imports. Among extensive margins, firm entry and exit is less influential than product-country adding and dropping.

While the intensive margin dominates year-to-year trade growth, its influence declines over longer time periods. Indeed, as indicated in the last column of Table 2, the intensive margin accounts for just 35% and 23% of ten-year export and import growth, respectively.

This waning influence of the intensive margin is intuitive. Because entering and exiting firms (and added and dropped products) tend to be small relative to established incumbents, their entry and exit does not contribute much to changes in exports or imports in the year they enter or exit. On the other hand, because new firms tend to grow relatively fast if they do manage to survive, they get large, thereby increasing the contribution of the extensive margin in the long run.4

  • Third, the intensive margin is more influential in explaining variation in trade between “related” parties (i.e. when the buyer and seller have some sort of corporate relationship) than variation in trade between arm’s-length firms.

This result is also intuitive, as multinationals tend to be large and relatively stable participants in international trade compared with non-multinationals. Thus, while their intensive margin may respond to shocks, they are less likely to enter and exit than non-multinationals.5

US margins during the current crisis

Unfortunately, the detailed, firm-level trade data required to analyse US firms’ responses to the current crisis are not yet available. However, analysis of firms’ reactions to two previous crises encompassed by the data that are available – the 2001 recession and the 1997 Asian financial crisis – suggests that most of the response will be on the intensive margin.

Lessons from the 2001 recession

As illustrated in Figure 2 and Figure 3, the $60 and $72 billion drops in exports and imports over the 2000-2001 episode are mostly due to the intensive margin, i.e., smaller increases and larger decreases among incumbent trading relationships than in non-recession years.6

For both exports and imports, the contribution of firm exits rises, both in the year of the recession and for one or two years afterwards. The product-nation margin – namely the introduction or withdrawal of products to new destinations – is more important for exports than for imports. For exports, the contribution of product-country additions falls and product-country drops rises.

Figure 2 Export intensive and extensive margins, 2001 recession.

Notes: Tables are from Bernard, Jensen, Redding and Schott (2009). See appendix tables for details.

The relative importance of the intensive margin is also evident in the 2001 recession. Appendix Table A2, for example, provides a similar decomposition but is restricted solely to US exports and imports between related parties.7 A comparison of the final three rows of Tables 2 and 3 reveals relatively modest contributions of the extensive margin in the years around the recession.

Figure 3 Import intensive and extensive margins, 2001 recession.

Notes: Tables are from Bernard, Jensen, Redding and Schott (2009). See appendix tables for details.

Lessons from the 1997 Asian Financial Crisis

While there are substantial changes in extensive margins around the Asian crisis, here, too, the intensive margin is most influential, and related-party trade is more resilient to the crisis.

These trends are readily apparent in Figure 4, which compares US trade with the crisis countries identified above, to trade with Rest-of-World (RoW) in the years around July, 1997.

Plots in the first row of the figures display the evolution of three types of trade around the crisis year, 1997. The first is total trade, the second is related-party, or RP, trade (i.e. trade that takes place between a buyer and seller who have a corporate relationship), and “arm's-length” (AL) trade, which is trade among parties that do not share corporate ties. The charts on the left show the facts for US exports to the crisis-stricken Asian nations (Indonesia, Korea, Malaysia, the Philippines and Thailand). For comparison, the charts on the right show the same for other nations. For comparability, each series is normalised to 100 in 1996. The salient points are:

  • Overall US exports to Asia declined by 21% between 1996 and 1998, while exports to RoW increased by 3%.
  • Within Asia, the decline in AL exports was substantially greater than the drop in RP exports, i.e,, 26% versus 4% by 1998.
  • For exports to RoW, the experience of RP and AL trade is similar. The increase in US imports from 1996 to 1998, reported in the last two columns of the figure, roughly mirrors the decline in exports in the first two columns.

Taken together, the two previous points suggest that trade within supply chains is more robust to shocks.

  • Import growth is slightly stronger for Asia than RoW (19% versus 17%), and, within Asia, is stronger for RP than AL trade (28% versus 11%).
  • AL and RP trade differ most in terms of the reaction of their intensive margins (+26% versus -1%).

When interpreting these facts, it is important to bear in mind that massive real devaluations by the crisis nations accompanied the Asian crisis.

Figure 4 US exports and imports during the 1997 Asian financial crisis.

Notes: Figures display margins for Asian crisis countries (Indonesia, Korea, Malaysia, the Philippines and Thailand) versus RoW from 1993 to 2000.. All series normalised to 100 in 1996.

Evolution of the margins

To investigate the behaviour of trade in this period more closely, Figure 5 and Figure 6 decompose the total impact on exports and imports into the intensive and extensive margins. Thus, each row of Figure 5 separates the aggregate response of exports into three mutually exclusive and exhaustive components – firms, products per firm, and the intensive margin. That is, changes due to the number of firms engaged in trade, changes due to variation in the number of products exported per firm, and changes due to the average amount exported per product per firm. Figure 6 shows the same for US imports in this period. As a control for general trends – i.e. to help identify the impact of the crisis itself – each figure shows the corresponding figure for US trade with the RoW.

As the first and final rows of the Figure 5 show, the number of firms exporting, as well as their intensive margin, declines substantially more for exports to crisis-hit Asia. The number of firms exporting dropped by -16% (Asia) versus -8% (RoW). The corresponding number for the intensive margins are -2% and +9%. These movements are consistent with the aggregate response associated with the 2001 recession referred to above.

The number of US arm’s-length firms exporting to the crisis countries declines more sharply than the number of related-parties exporters: by -16% versus -7% from 1996 to 1998. Again this suggests that within-firm trade relations are more robust than outside-the-firm relationships, in times of crisis.

A comparison of the intensive margins is even starker: -8% versus +9% for arm’s-length and related-parties, respectively.

The shallower decline in the number of firms exporting to related parties, as well as this increase in the intensive margin, explains the less severe impact of the Asian crisis on overall RP exports. By comparison, the average number of products exported per firm, changes relatively little between 1996 and 1998, for either Asia or RoW.

Contribution of each margin

While the figures are useful for summarising the behaviour of the margins of trade relative to their own past, they do not describe the relative contribution of each margin to overall changes in export or import value.

A decomposition of the margins of US trade with the Asian crisis countries (see the bottom panels of the appendix tables for specifics) reveals that:

  • Exports to the crisis countries declined by $5.6 and $2.7 billion in 1997 and 1998, respectively, before recovering in 1999.
  • Imports from the crisis countries increased by $5.8, $5.1 and $12.6 billion in the three years following the crisis.

In both cases, we find the intensive margin to be most influential in these changes, though the contribution of the extensive margin to 1998 export declines was substantially higher than in other years.

More broadly, the pattern of relatively large percentage changes on the extensive margin, accounting for relatively small shares of the changes in the value of overall trade, is consistent with the idea that exiting firms are small relative to those that survive.

Figure 5 Decomposition of US exports during the 1997 Asian financial crisis.

Notes: Figures display margins for Asian crisis countries (Indonesia, Korea, Malaysia, the Philippines and Thailand) versus RoW from 1993 to 2000. All series normalised to 100 in 1996.

Figure 6 Decomposition of US imports during the 1997 Asian financial crisis.

Notes: Figures display margins for Asian crisis countries (Indonesia, Korea, Malaysia, the Philippines and Thailand) versus RoW from 1993 to 2000. All series normalised to 100 in 1996.

Concluding Remarks

If the current “shock” to US trade is similar to those that have occurred before, most of the decline in exports and imports we’ve seen is due to less intense trade rather than firm and firm-product exit.

To the extent that this is true, trade will bounce back relatively quickly once conditions improve. It is of course possible that the severe shortage of credit available to firms over the past year has led to a higher-than-usual share of harder-to-reverse firm exit, potentially dampening the speed of recovery (see Dougherty 2009, and Chor and Manova 2009). On the other hand, if history is a guide, such exits are most likely concentrated among relatively small firms, compared with the multinationals that dominate US trade and have the wherewithal to weather the credit crunch.

Another factor which may speed recovery is the continued depreciation of the dollar. The decline and increase in the US dollar in the current crisis is similar to, but greater in magnitude than, the trend followed by the dollar in the 1990 recession. In both cases, the dollar was in long-run decline before the recession and increased in value during the recession. Should the dollar resume its decline as the recovery proceeds, the US might experience relatively high firm entry into export markets, putting further downward pressure on the trade deficit.


1 Seasonally adjusted monthly aggregate trade figures are available starting in 1989 from the U.S. Census Bureau’s website. As illustrated in Figure 1 below, the declines are even larger if one starts from the post-recession-onset peak exports and imports.

2 In Table 1, petroleum is included in mineral products (HS 27). For a similar breakdown with respect to exports and imports, see Appendix Table 1.

3 The results described in this section and the next are from Bernard, Jensen, Redding and Schott (2007, 2009). The detailed, firm-level data for such analyses are from the U.S. Census Bureau’s Longitudinal Firm Trade Transactions Database (LFTTD); see Bernard, Jensen and Schott (2009) for further details. These data are currently available for 1992 through 2005.

4 See, for example, Eaton et al. (2008).

5 See also Obashi (2009).

6 Differences between the decline in net exports in Table 2 versus Table 1 are due to their coverage of different periods as well as the exclusion of trade data in the former table for transactions that cannot be matched to a firm. For more information, see Bernard, Jensen and Schott (2009).

7 In U.S. trade data, exporting firms are considered “related” to their foreign counterparty if either owns at least 10% of the other. For imports, the threshold is 6%.


Bernard, Andrew B., J. Bradford Jensen, Stephen J. Redding and Peter K. Schott (2009) “The Margins of US Trade,” American Economic Review Papers and Proceedings 99(2):487-493.

Bernard, Andrew B., J. Bradford Jensen, Stephen J. Redding and Peter K. Schott (2007) “Firms in International Trade,” Journal of Economic Perspectives 21(3):105-130 (Summer).

Bernard, Andrew B., J. Bradford Jensen and Peter K. Schott (2009) "Importers, Exporters and Multinationals: A Portrait of Firms in the US that Trade Goods," in T. Dunne, J.B. Jensen and M.J. Roberts (eds.), Producer Dynamics: New Evidence from Micro Data (University of Chicago Press).

Chor, Davin and Kalina Manova (2009) “Off the Cliff and Back? Credit Conditions and International Trade During the Global Financial Crisis”, Mimeo.

Dougherty, Carter (2009), “Trade Finance Shrivels, Pushing Downturn Deeper”, New York Times (Tuesday, March 3).

Bernard, Andrew B., Stephen J. Redding and Peter K. Schott (2009) “Multiple-Product Firms and Trade Liberalization,” NBER Working Paper 12782 (revised).

Eaton, Jonathan, Marcela Eslava, Maurice Kugler and James R. Tybout (2008), “The Margins of Entry into Export Markets: Evidence from Colombia,” in The Organization of Firms in a Global Economy, ed. Elhanan Helpman, Dalia Marin and Thierry Verdier, 231-272. Cambridge: Harvard University Press.

Melitz, Marc J. (2003) "The Impact of Trade on Intra-Industry Reallocations and Aggregate Industry Productivity," Econometrica 71, November 2003, pp. 1695-1725.

Obashi, Ayako (2009), “Resiliency of Production Networks in Asia: Evidence from the Asian Crisis,” ERIA Discussion Paper Series 2009-21.

Pavcnik, Nina (2002) "Trade Liberalization, Exit, and Productivity Improvement: Evidence from Chilean Plants," Review of Economic Studies, 69(1), 245-76.


Appendix Table 1a: Decomposition of US exports, extensive and intensive margins, 1993-2003 (annual, 5-year & 10-year horizons).

Notes: Tables are from Bernard, Jensen, Redding and Schott (2009). Data are from the LFTTD. Panels decompose total change in US exports or imports ($ billion) during the noted periods according to noted firm activities. Rows 1 to 3 summarise the contribution of firm entry into and exit. Rows 4 to 6 summarise changes in firms' product-country combinations. Rows 7 to 9 summarise the growth and decline of continuing product-country exports or imports at continuing importers. Bottom panel reports percentage contribution of each net margin in terms of the total change in imports. Each column summarises growth over the noted time interval. Years are July to June, e.g., 2001-2 is July 2001 to June 2002.

Appendix Table 1b: Decomposition of US imports along extensive and intensive margins, 1993-2003.

Note: See previous tables.

Appendix Table 2a: Decomposition of US ‘related parties’ exports along extensive and intensive margins, 1993-2003.

Note: See previous tables.

Appendix Table 2b: Decomposition of US ‘related parties’ imports along extensive and intensive margins, 1993-2003.

Notes: See previous tables.

6,299 Reads