Abstract
- K. Huber examines the difficulties faced by the German bank Commerzbank in the wake of the 2008 crisis to measure the effects of a credit crunch on the real economy.
- The author shows that the difficulties of a major bank can have macroeconomic consequences for the regions that depend on it.
What is the impact of a decline in credit supply on the economy when a bank or banking system finds itself in difficulty? This question is crucial for regulators: if credit supply plays an important role, this justifies stronger prudential regulation upstream of crises and more significant easing when crises occur. The question is not trivial: credit supply can only play an important role if it is not perfectly substitutable by other sources of financing; the difficulties of an isolated bank can only have consequences if its customer relationships are not immediately replaced by others with a competing bank.
To answer this question, economists study specific events affecting banks and compare companies or regions that are more or less connected to them. For example, J. Peek and E. Rosengren (2000) examine the effects of the Japanese banking crisis on US commercial real estate companies that were financed to varying degrees by the US branches of these banks. A. Khwaja and A. Mian (2008) study the consequences of the 1998 nuclear tests in Pakistan, following which the authorities decided to restrict dollar withdrawals, thereby placing greater constraints on banks that relied most heavily on the dollar for financing.
This paper by K. Huber (2018), published like the previous two in the American Economic Review, follows in this tradition. Huber focuses on the case of Commerzbank in Germany during the 2008 financial crisis. This major bank, which accounted for 9% of non-financial credit in Germany at the onset of the crisis, was characterized by a very late divestment of its exposures to US mortgage-backed securities and lost 68% of its equity between 2007 and 2009. These losses forced it to restructure its business in Germany more significantly than its competitors, despite a relatively healthy domestic credit portfolio: Commerzbank’s domestic credit outstanding fell by 17% compared to its competitors between 2009 and 2012.
In order to establish a causal link between this « supply shock » and economic activity, empiricists face at least two difficulties.
First, they must measure a causal effect that is a priori bidirectional. The slowdown in the real economy can cause a contraction in credit supply, as can the reverse. To overcome this difficulty, the author uses microeconomic data on bank credit to measure each company’s dependence on Commerzbank. This rate is measured for a company as the number of Commerzbank branches that have granted it loans, divided by the total number of branches that have financed it. The paper thus shows that a company that is completely dependent on Commerzbank has an employment level 5.3% lower than a similar company that is independent of this bank, over the period 2009-2012. The key is the comparability of these two fictitious companies. Indeed, if the companies financed by Commerzbank were structurally weaker or more vulnerable to the crisis, their decline in activity could not be attributed to a decline in credit supply. The study clearly shows that the degree of dependence on Commerzbank is independent of characteristics such as the sector, size, or age of its borrowers. Figure 1 shows the divergence in employment between two companies that differ only in terms of their banking relationship with Commerzbank.
Employment trends for the average company with a banking relationship with Commerzbank compared to the average company without such a relationship, normalized to 1 in 2006
Secondly, credit supply and economic activity may be driven by common underlying factors. This question arises when the author seeks to measure the macroeconomic effects of dependence on Commerzbank by comparing the performance of districts (Landkreiser) with different levels of dependence on Commerzbank. Commerzbank has a very specific geographical footprint. For example, districts where Commerzbank has expanded may be more vulnerable due to less protective local legislation.
To overcome this difficulty, one of the original features of the paper is that it exploits the partial exogeneity of Commerzbank’s network in Germany. At the end of World War II, the Allies, anxious to limit the centralization of the German economy, divided Commerzbank into three entities, each attached to a different « banking zone » and head office. A tripolar bank was created with Düsseldorf, Frankfurt, and Hamburg as its epicenters. Sixty years later, the districts closest to these three cities are still, on average, more dependent on Commerzbank. Huber therefore uses the distance of each district from the nearest of these three cities as an exogenous measure of dependence on Commerzbank. An important detail is that the author controls for the distance to each of these three cities. The aim is to measure the effect of having a banking relationship with Commerzbank, not the effect of being located a certain distance from Frankfurt or Hamburg. This analysis allows Huber to conclude that a one standard deviation increase in a district’s dependence on Commerzbank leads to a 0.8% decline in employment over the period.
Conclusion
The use of microeconomic data and specific banking events allows economists to provide fairly accurate estimates of the economic cost of a banking crisis. In turn, these estimates provide quantitative evidence for the ongoing debate on the appropriate level of prudential regulation.
