VoxEU Column Global crisis International Finance

Running for the exit: International banks and crisis transmission

Cross-border bank lending fell dramatically during the global crisis, but lending to some countries declined far more severely than to others. Recreating the monthly lending flows of the 118 largest international banks, this column finds that banks with head offices farther away from their customers are less reliable funding sources during a crisis, suggesting that the nationality of foreign banks matters.

In the wake of the 2007-2009 economic crisis, the virtues and vices of financial globalisation are being re-evaluated. Financial linkages between countries, in particular in the form of bank lending, have been singled out as a key channel of international crisis transmission. The IMF and the G20 have identified the volatility of cross-border capital flows as a priority related to the reform of the global financial system. Indeed, cross-border capital flows declined substantially in the wake of the crisis, with the collapse of Lehman Brothers being followed by a fall in syndicated cross-border lending by 53% on average.

Figure 1 illustrates, however, that the magnitude of this reduction differed substantially across countries. An important question therefore is why cross-border bank lending to some countries is much more stable than to other countries. In a recent working paper (De Haas and Van Horen 2010), we show that access to borrower information is a key determinant of lending stability in times of crisis.

Figure 1. Distribution of the change in cross-border lending after the Lehman Brothers collapse

Note: This figure shows the distribution across destination countries of the change in the average monthly cross-border syndicated lending inflows after the collapse of Lehman Brothers compared to the pre-crisis period. The pre-crisis period is defined as January 2005 to July 2007 and the post-Lehman period as October 2008 to October 2009. Each bar indicates the number of destination countries that experienced a post-Lehman change in bank lending that falls within the percentage bracket on the horizontal axis. For instance, there were 11 countries to which cross-border syndicated bank lending declined by between 25 and 50 per cent while there were only 2 countries that experienced an increase in cross-border syndicated lending of between 25-50 per cent. In 16 countries (4+12) lending declined by more than 75 per cent.

Distance, borrower information, and lending stability

Banks’ ability to screen and monitor varies across borrowers. Agency problems are more pronounced for opaque companies. When screening and monitoring is difficult, the scope for adverse selection and moral hazard remains high and banks resort to credit rationing (Stiglitz and Weiss 1981). Screening and monitoring intensity also varies over time. An adverse economic shock increases the marginal benefits of screening and monitoring as the proportion of firms with a high default probability increases (Ruckes 2004). Adverse selection and moral hazard will also increase in reaction to firms’ lower net worth. However, banks face difficulties in offsetting increased agency problems if borrowers are opaque. In response to an adverse shock they therefore resort to credit rationing of “intransparent” borrowers in particular (Bernanke et al. 1996).

In a similar vein, we expect that during the recent crisis banks reduced cross-border lending more to countries where they were unable to generate additional borrower information and had to resort to credit rationing instead. We test this hypothesis by using a number of variables that capture the ease with which banks’ can screen and monitor foreign borrowers.

  • A first factor that determines banks’ ability to screen and monitor effectively is distance to the borrower. It is particularly difficult to collect “soft” information on remote borrowers. As a result, banks tend to lend less to far-away clients (Jaffee and Modigliani 1971). In line with such geographical credit rationing, we expect that distant firms saw a sharper decline in the supply of cross-border bank lending during the crisis than less remote companies.
  • Second, the ability to screen and monitor may be positively affected by the establishment of a local subsidiary as local loan officers tend to be better placed to extract soft information from firms. However, while a local subsidiary reduces the physical distance between firm and loan officer, it also creates “functional distance” within the bank. Banks may experience difficulties in efficiently passing along (soft) information from the subsidiary to headquarters (Aghion and Tirole 1997). Whether the presence of a subsidiary makes cross-border lending more stable or not therefore depends on whether the positive effect of the shorter physical distance is offset by the negative effect of a longer functional distance.
  • Third, cooperation with domestic banks can positively affect banks’ ability to generate (additional) information about their clients, as domestic banks tend to have better access to information about the creditworthiness of local firms. We therefore expect that international banks find it easier to lend in times of crisis to countries where they are well-integrated in a network of domestic banks.
  • Finally, we expect that during the financial crisis banks reduced their lending to a lesser extent to countries where they had built up substantial pre-crisis lending experience. A bank that is more familiar with the culture, institutions, and customs in a country may have less problems in generating useful information about borrowers.

We use a unique dataset with detailed information on international syndicated loans to test whether the above factors did indeed determine the stability of cross-border bank lending during the crisis. We recreate the monthly lending flows of the 118 largest international banks to all countries of operations. For each bank-destination country pair we calculate the change in average monthly cross-border bank lending in the year after the collapse of Lehman Brothers compared to the pre-crisis period January 2005-July 2007. We then use OLS and logit regressions to analyse to what extent access to borrower information determined the change in bank lending during the crisis.

Empirical findings

We find that during the crisis banks were better able to keep lending to countries that are geographically close, in which they are well integrated into a network of domestic co-lenders, and in which they had gained experience by building relationships with (repeat) borrowers. For emerging markets, where trustworthy “hard” information is less readily available and a local presence may be more important, we also find (weak) evidence that the presence of a local subsidiary stabilises cross-border lending. These findings hold for lending to both first-time and repeat borrowers. Interestingly, the stability of syndicated lending to bank borrowers was not affected by any of these factors. Agency problems and mistrust in the market for long-term inter-bank lending were simply too large for banks to mitigate them in any meaningful way. As a result, the sudden stop in cross-border lending was significantly larger for banks than for non-banks.

Because international banks were more inclined to keep lending to some countries than to others, we document substantial variation in the severity of the sudden stop. Figure 2 illustrates that most countries were unable to offset the decline in cross-border lending through increasing domestic syndicated lending. The left-hand pane shows that there were only a few countries – India, China, Japan – where increased lending by (often state-owned) banks more than compensated for the drop in cross-border inflows. The right-hand pane shows that most countries experienced a decline in total syndicated lending very similar to the decline in cross-border syndicated lending (observations on the 45º line). This imperfect substitutability between cross-border and domestic syndicated loans implies that the results we document in this column are likely to have had severe consequences for the total lending supply in the destination countries.

Figure 2. A comparison of cross-border and total syndicated lending

Note: This figure compares the change in cross-border syndicated lending to a country (horizontal axis) with the change in total syndicated lending (cross-border plus domestic syndicated lending) in that country. Lending change is the percentage change in average monthly lending in the post-Lehman compared to the pre-crisis period. The pre-crisis period is defined as January 2005 to July 2007 and the post-Lehman period as October 2008 to October 2009. The left-hand pane shows all 60 destination countries included in our dataset whereas the right-hand pane zooms in on those countries that experienced a decline in both cross-border and total syndicated lending. Countries that experienced a percentage change in domestic lending that was exactly equal to the percentage change in cross-border lending are on the 45º line. Countries where domestic lending shrank faster (slower) than cross-border lending are to the right (left) of this line.

Policy implications

Our results bear on the policy debate on financial globalisation and in particular on whether and how countries should integrate with global financial markets. A key feature of cross-border lending that has been a focus of debate, and further underlined by the recent crisis, is its unstable character. Our findings provide some first answers to the question of when cross-border lending is particularly volatile and when it is not. Perhaps somewhat controversially, we find that banks that are further away from their customers are less reliable funding sources during a crisis. Clearly, policymakers not only need to make a decision on whether to open up their banking system but also to whom.

A second finding is that international banks with a local presence on the ground may be more stable providers of credit. For emerging markets that are considering to open up their banking system this implies that stimulating banks to “set up shop” may kill two birds with one stone. Not only do foreign bank subsidiaries provide for a relatively stable credit source themselves, but their presence may also stabilise the cross-border component of bank lending. Rather than imposing capital controls to reduce the volatility of cross-border lending, countries may thus contemplate allowing international banks to also set up a local affiliate.

The views expressed in this column are those of the authors only and do not necessarily reflect the views of the EBRD, De Nederlandsche Bank or their respective Boards.


Aghion, P and J Tirole (1997), “Formal and Real Authority in Organisations”, Journal of Political Economy, 105:1-29.

Bernanke, B, M Gertler, and S Gilchrist (1996), “The financial accelerator and the flight to quality”, Review of Economics and Statistics, 78:1-15.

De Haas, R and N Van Horen (2010), “Running for the Exit: International Banks and Crisis Transmission”, EBRD Working Paper 124 and DNB Working Paper 279.

Jaffee, DM and F Modigliani (1971), “A Theory and Test of Credit Rationing”, American Economic Review, 59:850-872.

Ruckes, M (2004), “Bank competition and credit standards”, Review of Financial Studies, 17:1073-1102.

Stiglitz, J and A Weiss (1981), Credit rationing in markets with imperfect information, American Economic Review, 71:393-410. 

210 Reads