VoxEU Column International trade

Dissecting the effect of credit supply on trade

On the back of the global crisis came the global trade collapse, as international trade fell 15% between 2008 and 2009. Was this a result of credit lines being cut or did demand simply disappear? This column uses Peruvian export data to argue that bank credit played a smaller role than suggested by previous estimates.

One of the most striking aspects of the Great Recession was the Great Trade Collapse. According to the IMF Global Data Source, between the first quarter of 2008 and the first quarter of 2009 international trade fell by 15%, while real world GDP fell by 3.7%.

The debate on the causes of this trade collapse is far from settled. The emerging consensus is that the drop in international demand played a central role (see Baldwin 2009 on this site). In particular, the demand for durable consumption and investment goods, highly represented in international trade, was severely affected. Moreover, the concurrent price collapse in commodities, a major source of exports for many emerging markets, further contributed to the drop in world trade.

However, an open question is whether additional factors contributed to the dramatic trade collapse. In particular, additional factors may explain why the drop in trade was so widespread across countries and export products.

Several economists have pointed towards credit shortages, caused by the 2008 financial crisis, as a potential culprit of the trade decline (Amiti and Weinstein 2009a, 2009b; Chor and Manova 2010a, 2010b; Fontagné and Gaulier 2009). The idea is that banks cut credit during economic downturns, and the resulting credit shortages amplify the contraction in economic output and trade. Indeed, this insight has a long history and is believed to explain the sharp and sudden decline in output during the Great Depression (Friedman and Schwartz 1963; Bernanke 1983). However, to date it is unclear how much of the output decline during the financial crisis of 2008 was due to bank credit shortages, or whether the credit shortages had a particularly severe effect on exports relative to domestic sales.

In recent research (Paravisini et al. 2011), we address these questions by analysing the impact of the 2008 financial crisis on Peruvian exports. We find that Peruvian banks played an important role in the international transmission of the crisis, and, through this channel, the international credit crunch negatively affected Peruvian export performance. Although the credit crunch had a first order effect on firms’ exports, it explains less than 15% of the decline in its volume during the 2008 crisis. In addition, we find no evidence to suggest that exports are more sensitive to credit than domestic sales.

The role of banks in the international transmission of the financial crisis

International funding to banks in Peru declined sharply after July 2008 (Figure 1a). The decline was due to the reversal of portfolio capital flows from Peru to developed economies, a phenomenon known as “flight to quality” that typically affects emerging economies following sharp increases in uncertainty in international financial markets. As a consequence of the capital-flow reversals, banks that depended heavily on foreign funding prior to the crisis reduced the amount of lending to Peruvian firms. The loan growth rate by these banks declined during the twelve months following July 2008 relative to the previous year (what we call “Pre” and “Post” periods, respectively). The opposite was true for banks that were less reliant on foreign funding (see Figure 1b.).

Figure 1. Reversal of capital flows and Peruvian banking sector

a. Foreign funding into the banking sector
b. Lending by banks with low and high foreign funding dependence

Source: Superintendencia de Bancos y Seguros, Peru. High and low foreign dependence refers to the bank’s ratio of foreign liabilities to assets relative to the mean across banks in 2007.

Effect of bank credit crunch on the supply of exports

The volume of exports by firms borrowing from banks affected by the credit crunch declined by 25% relative to exports by firms borrowing from unaffected banks during the post period. However, we have to be careful in attributing this decline solely to credit shortages. The reason is that banks develop expertise in lending to exporters concentrated in certain economic sectors or geographical markets. The export performance of firms borrowing from affected banks is driven by both changes in credit supply and demand fluctuations in those markets the bank specialises in.

For example, in 2007 over 31% of exports by Peruvian firms borrowing from HSBC went to the US, as opposed to only 3% of exports by firms borrowing from Banco Santander. HSBC Peru was also highly dependent on foreign funding before the capital-flow reversals while Santander was not (see Table 1). Thus, it is difficult to distinguish whether the poor relative export performance of HSBC borrowers was due to the credit shortage, or due to a disproportionately large decline in export demand from the US.

Table 1. Foreign funding dependence by commercial banks (December 2007)

Bank
 
Foreign liability/assets
HSBC
 
17.7%
Mibanco
 
16.8%
BBVA Continental
 
12.2%
Citibank
 
10.3%
Interamericano
 
7.5%
Financiero
 
7.3%
Credito
 
6.2%
Wiese-Scotia Bank
 
6.0%
Interbank
 
5.5%
Santander
 
2.2%

 

Source: Superintendencia de Bancos y Seguros, Peru

To disentangle the effect of credit shortages on export performance, we compare exports of the same product and to the same destination by firms borrowing from banks with high and low foreign funding dependence. For example, consider two firms that export Men’s Cotton Overcoats to the US. One of the firms obtains all its credit from Bank A, which was negatively affected by the credit crunch, while the other firm obtains its credit from Bank B, which was not. Changes in the demand or price of overcoats in the US should affect exports by both firms in a similar way. Also, any change in the prices of inputs (e.g., cotton, wages in the garment sector) equally affects both firms. In our example, the change in exports of Men’s Cotton Overcoats to the US by a firm that borrows from Bank A relative to a firm that borrows from Bank B isolates the effect of credit on exports.

Using this approach, we find that credit shortages led to a 13% decline in exports by firms borrowing from affected banks relative to firms borrowing from unaffected banks. This means that credit shortages explain approximately half of the 25% relative export decline by those firms, while the rest results from differences in international demand. This implies that banks highly dependent on foreign funding specialise in lending to firms that suffered a disproportionate larger drop in international demand for their products.

Are exports to distant markets more sensitive to credit shortages?

International trade involves long freight times and, therefore, may require additional credit to finance the time between shipment and payment. For this reason, distance to destination markets is often used to proxy for funding needs that are specific to export activities.

While we find that the decline in exports was larger for more distant export markets, this was not because exports to these markets were more sensitive to credit shortages. Instead, firms borrowing from affected banks suffered a larger decline in international demand for their products, and this led them to scale back their international operations. Since distant markets typically represent a small share of exporters’ international business, these were the first ones to be discontinued.

Again, it is important to account for the heterogeneous decline in international demand across products and destinations. When comparing exports for the same product and destination, we find that the credit shortage had a disproportionately large effect on exports to close destinations. The reason is that banks cut lending disproportionately to small firms, and small firms tend to export to neighbouring countries.

The quantitative importance of the bank-credit channel

We compute the sensitivity of exports to bank credit. A 10% drop in the firm’s stock of bank credit generates a reduction of 2.3% in its intensive margin of exports, that is, the annual volume of exports of a given product-destination flow that a firm continues exporting after the crisis. Credit also affects the extensive margin of trade. A 10% drop in the stock of credit reduces by 3.6% the number of product-destination markets served by a firm during a year.

These sensitivities are useful to gauge the importance of the credit channel on the overall drop in Peruvian export growth during 2008. The value growth of Peruvian exports declined from 11% during the 12 months prior to the capital-flow reversal (July 2008), to -22.4% in the following 12 months. Most of this decline can be attributed to the reduction in the international price of Peruvian products, since during the same period, the volume growth of exports volume declined by 12.8 percentage points (from 3.2% to -9.6%; see Table 2). Our results imply that the bank-credit channel can account for less than one-sixth of the overall decline in the volume of Peruvian exports during the 2008-2009 crisis.

Table 2. Growth of exports before and after the capital flow reversal

 
 
Volume (kg)
 
Missing volume
of trade
% of missing trade explained by
credit crunch
 
 
Pre
Post
 
Total
 
3.2%
-9.6%
 
-12.80%
15.10%
 Intensive
 
2.1%
-2.2%
 
-4.30%
26.92%
 Extensive
 
1.2%
-7.4%
 
-8.60%
9.01%

 

Note: Intensive margin corresponds to volume of export flows active in the Pre and Post periods. Extensive correspond to volume of new and discontinued export flows. Flows are defined as a firm-destination-product (HS 4-digit) export. Pre and Post correspond to the 12 months before and after July 2008.

Conclusion

Our study suggests that the reduction in bank credit had a first order effect on exports, but the bulk of the export decline is attributed to other factors – i.e., the international price decline and lower demand for Peruvian products. The importance of the credit channel is smaller than previously estimated – Amiti and Weinstein (2009a), for example argue that close to one-third of the drop in Japanese exports during the 1990s can be attributed to finance. And since we find no relationship between the sensitivity of exports to finance and factors related to export specific costs (i.e., shipping, export insurance, freight time, etc.), the estimated sensitivities are likely to pertain the overall usage of credit by the firm and not exclusively the funding of export activities.

References

Amiti, Mary, and David Weinstein (2009a), “Exports and Financial Shocks”, NBER Working Paper 15556.

Amiti, Mary, and David Weinstein (2009b), “Exports and Financial Shocks: New Evidence from Japan”, VoxEU.org, 23 December.

Baldwin, Richard (ed.) (2009), The Great Trade Collapse: Causes, Consequences and Prospects, A VoxEU.org Publication, 27 November.

Bernanke, Ben (1983), “Non Monetary Effects of the Financial Crisis in the Propagation of the Great Depression”, The American Economic Review, 82(4):901-921.

Bricongne, Charles, Lionel Fontagne, Guillaume Gaulier, Daria Taglioni, and Vincent Vicard (2009), “Exports and Financial Shocks”, Banque de France Working Paper 265.

Chor, Davin, and Kalina Manova (2010a) ‘Off the Cliff and Back? Credit Conditions add International Trade during the Global Financial Crisis.’ NBER Working Paper 16174.

Chor, Davin, and Kalina Manova (2010b) “Off the Cliff and Back? Credit Conditions

and International Trade during the Global Financial Crisis”, VoxEU.org, 15 Febuary.

Friedman, Milton and Anna Schwarz (1963), A Monetary History of the US, 1867-1960, Princeton University Press.

Fontagné, Lionel and Guillaume Gaulier (2009), “French Exporters and the Global Crisis”, in Richard Baldwin (ed.), The Great Trade Collapse: Causes, Consequences and Prospects, VoxEU.org. 27 November.

Paravisini, Daniel, Veronica Rappoport, Philipp Schnabl, and Daniel Wolfenzon (2011), “Dissecting the Effect of Credit Supply on Trade: Evidence from Matched Credit-Export Data”, NBER Working Paper 16975.

 

1,155 Reads