“Credit Risk of Financial Institutions has Increased through Basel II”

Reported probability of default and actual default rate under the standard and the model-based approach

(The interview appeared in SAFE Newsletter Q4 2014)
 

Rainer Haselmann joined Goethe University and the Research Center SAFE as a Professor of Finance, Accounting and Taxation in October 2014. Previously, he was a Professor of Finance at the University of Bonn. His research focuses on banking regulation, lending and portfolio allocation. After having earned a Ph.D. in 2006 from the Leipzig Graduate School of Management, Rainer Haselmann worked as a postdoctoral researcher at the Columbia Business School and University of Mainz. From 2009 to 2011, he was an Assistant Professor at the Bonn Graduate School of Economics.

Which research questions are you currently focusing on?

My research is about financial intermediaries with a strong focus on the questions of how financial institutions react to regulation and how this reaction affects the real economy, mainly via bank lending. A related field is the examination of how certain regulations come into place, taking a political economy perspective on regulation in the financial sector.


A recent paper of yours focuses on model-based capital regulation that was introduced by the Basel Committee (Basel II). What did this regulation aim for?

Banks have an interest to reduce their equity holdings to a minimum. This is why regulators impose capital charges that reflect the riskiness of loans. The main idea of Basel II was to steer banks’ behavior. When a bank invests in a risky asset or provides loans, it should have an adequate equity buffer to bear possible losses. Under Basel I, all bank assets were classified into five broad risk buckets with each of them having a fixed risk weight assigned. The most important novelty of the Basel II regulation was to allow banks to develop their own internal rating for each client which would then be used to determine how high the capital buffer for this exposure should be. One reason behind allowing banks to assess borrower risk is that they have inside knowledge about their clients and, therefore, should be able to evaluate their risks better than outsiders such as external rating agencies.


However, they have different incentives…

Exactly. Banks have a strong incentive to under-evaluate risks in order to economize on capital charges. For this reason, the banks’ risk models have to be certified by the supervisor on a portfolio basis. This way, so the idea behind this exercise, strategic understatement of risk would be impossible.


What are your findings?

In our paper, we look at the introduction of model-based capital regulation in Germany, which was introduced step-wise because of the time-consuming certification process. The long introduction phase allows us to observe, simultaneously, both regulatory designs within the same bank. We were able to compare portfolios that have been shifted to the new regulation with portfolios that were still under the standard approach but would eventually be shifted. What we find is that the estimated ratings within the standard approach predict the probability of defaults quite well, while the risk models under the new model-based regulation tend to considerably under-predict actual defaults (see Figure 1). This is, of course, quite surprising. The main idea of the new regulation had been to come to a more sophisticated risk assessment. But the outcome was directly opposed. 


How do you explain these results? Did the banks manage to manipulate the models despite the certifying process?

This is difficult to say. What can give us an indication are the interest rates that the banks charged their clients on these loans. We find that the interest rates under the model-based approach are higher than under the standard approach and tend to reflect the actual default patterns and, thus, the real riskiness of the loans (see Figure 2). So, interest rates and reported ratings show opposite patterns. This, in fact, suggests that the banks were quite aware of the riskiness of their loan portfolios.

So, why is the regulated equity buffer not linked to the interest rate charged?

Banks could circumvent such a regulation by charging fees or changing other contractual features instead of adjusting the interest rate. It is nearly impossible to design a regulation which is geared towards getting a regulated entity to act against its own interest.

What lessons need to be drawn from your findings? Should regulation be less complex in order to allow for better supervision?

In theory, a highly complex regime like model-based regulation is clearly better than simple categories for risk assignment. But in practice, especially when enforcement costs are high, this might not be the case. However, I cannot prove that simplicity is always better than complexity. As a reaction to our results, you can also think of even more complex regulation that is supervised by many more regulators. But I strongly believe that we would benefit from simplicity in the regulation of banks.

What are the implications of your results for financial stability?

When Lehman collapsed in September 2008, a considerable number of the large German banks had already introduced model-based risk regulation while most cooperative and savings banks remained in a standard approach similar to Basel I. The large banks benefitted from the regulation because they were able to reduce capital charges and, thus, to expand their lending – potentially at the expense of smaller banks. We find that banks that opted for the new approach increased their lending by about 9 percent relative to banks that remained under the traditional regime. In other words, this regulation subsidized larger banks, which seems rather paradoxical given the systemic risk associated with larger banks. All in all, our results suggest that the aggregated credit risk of financial institutions has increased through the regulatory framework of Basel II. So, the regulator has failed to meet the objective of better detecting default risks and increasing financial stability.

Behn, M., Haselmann, R., Vig, V. (2014)
“The Limits of Model-Based Regulation”,
SAFE Working Paper No. 75


Selected Publications by Rainer Haselmann

Hackbarth, D., Haselmann, R., Schoenherr, D. (2014)
"Financial Distress, Stock Returns, and the 1978 Bankruptcy Reform Act",
forthcoming in Review of Financial Studies.

Althammer, W., Haselmann, R. (2011)
“Explaining Foreign Bank Entrance in Emerging Markets”,
Journal of Comparative Economics, Vol. 39, Issue 4, pp. 486-498.

Haselmann, R., Wachtel, P. (2011)
“Foreign Banks in Syndicated Loan Markets”,
Journal of Banking and Finance, Vol.35, Issue 10, pp. 2679-2689.

Haselmann, R., Pistor, K., Vig, V. (2010)
“How Law Affects Lending”,
Review of Financial Studies, Vol. 23, Issue 2, pp. 549-580.

Haselmann, R., Wachtel, P. (2010)
“Institutions and Bank Behavior”,
Journal of Money, Credit and Banking, Vol. 42, Issue. 5, pp. 965-984.

Haselmann, R., Herwartz, H. (2009)
“The Introduction of the Euro and its Effects on Investment Decisions”,
Journal of International Money and Finance, Vol. 29, Issue 1, pp. 94-110.


SAFE profile page of Rainer Haselmann