Procyclical Capital Regulation and Lending

• Author: RAINER HASELMANN,MARKUS BEHN,PAUL WACHTEL
• Date: 01 April 2016
• Published date: 01 April 2016
• DOI: http://doi.org/10.1111/jofi.12368

THE JOURNAL OF FINANCE •VOL. LXXI, NO. 2 •APRIL 2016

Procyclical Capital Regulation and Lending

MARKUS BEHN, RAINER HASELMANN, and PAUL WACHTEL∗

ABSTRACT

We use a quasi-experimental research design to examine the effect of model-based

capital regulation on the procyclicality of bank lending and firms’ access to funds. In

response to an exogenous shock to credit risk in the German economy,capital charges

for loans under model-based regulation increased by 0.5 percentage points. As a

consequence, banks reduced the amount of these loans by 2.1 to 3.9 percentage points

more than for loans under the traditional approach with fixed capital charges. We

find an even stronger effect when we examine aggregate firm borrowing, suggesting

that microprudential capital regulation can have sizeable real effects.

CAPITAL ADEQUACY REQUIREMENTS are widely considered to be the most effective

tool to ensure the safety and soundness of financial institutions. The determi-

nation of appropriate capital charges, however, is subject to much debate (see,

e.g., Peltzman (1970), Diamond and Rajan (2000,2001), Morrison and White

(2005), Acharya (2009), Admati and Hellwig (2014)). The main rationale for

capital regulation is the existence of guarantees, both implicit and explicit, that

provide banks with an incentive to hold less capital than is socially optimal.

However, regulating capital via a simple capital to asset ratio might incen-

tivize banks to hold portfolios with more risky assets (Koehn and Santomero

∗Markus Behn is at European Central Bank. Rainer Haselmann is at Goethe University and

SAFE. Paul Wachtel is at Stern School of Business, New York University. We appreciate the in-

sightful suggestions made by the Editor,Michael Roberts, two referees, and an Associate Editor. We

further thank Puriya Abbassi, Linda Allen, Tobias Berg, Claudia Buch, Valeriya Dinger, Iftekhar

Hasan, Robert Hauswald, Martin Hellwig, Anil Kashyap, Michael Koetter, Steven Ongena, An-

dreas Pfingsten, Peter Raupach, Jean-Charles Rochet, Amit Seru, Javier Suarez, David Thesmar,

Vikrant Vig, and Dawid Zochowski for helpful comments as well as seminar participants at Bonn

University, Bank of Finland, Bank of Israel, Deutsche Bundesbank, University of Mainz, Uni-

versity of M¨

unster, University of Zurich, the Bank for International Settlements RTF Workshop,

the Bundesbank/DGF/IMF Conference, the Central Banking Workshop, the Dubrovnik Economic

Conference, the German Finance Association Meeting, and the German Research Foundation

SPP-1578 Workshop. Weare grateful to Deutsche Bundesbank, especially to Klaus D ¨

ullmann and

Thomas Kick, for generous support with the construction of the data set and deep insights regard-

ing the German banking sector. Peik Achtert, Birgit Maletzke, Christoph Memmel, and Wolfram

Ohletz provided valuable insights regarding institutional details surrounding the introduction of

asset-specific risk weights by German banks. The views expressed in this paper are those of the

authors and do not necessarily reflect the views of the European Central Bank or its staff. Hasel-

mann thanks the German Research Foundation (Priority Program 1578) and the Research Center

SAFE, funded by the State of Hessen initiative for research LOEWE for financial support. The

authors have no conflicts of interest with respect to The Journal of Finance disclosure policy. The

usual disclaimer on errors also applies here.

DOI: 10.1111/jofi.12368

919

920 The Journal of Finance R

(1980), Kim and Santomero (1988)). Consequently, regulators have been in-

creasing efforts to link capital charges to asset risk. The Basel II framework,

the most important such effort, introduced capital charges for individual loans

that depend on model-based risk estimates from banks’ internal risk models.1

Linking capital charges to asset risk, however, may exacerbate the pro-

cyclicality of lending (see Dan´

ıelsson et al. (2001), Kashyap and Stein (2004),

Repullo and Suarez (2012)). If measures of asset risk are responsive to eco-

nomic conditions, then capital requirements will increase during a downturn.

At the same time, bank capital is likely to be eroded by loan write-offs. Capital-

constrained banks that are unable or unwilling to raise new equity in the

downturn will be forced to deleverage by reducing lending, which exacerbates

the downturn. Thus, well-intentioned regulatory policy can amplify business

cycle fluctuations. While the procyclical features of Basel II have been widely

discussed, we still lack a precise understanding of the link between counter-

cyclical capital charges and the corresponding adjustment of loans.2In this

paper, we exploit the institutional setup around the introduction of model-

based regulation in Germany in combination with an exogenous real sector

shock to identify the procyclical effects of risk-sensitive capital regulation on

banks’ lending behavior and firms’ overall access to funds. The effects on firms’

access to credit are particularly important for assessing the macroeconomic

implications of procyclical capital regulation.

Identifying the effects of model-based capital charges on lending is a difficult

task for four reasons. First, a bank’s risk assessment and lending decision

for a specific borrower are likely to be jointly determined, which means that

the relationship between changes in risk estimates and lending adjustments

suffers from endogeneity. Second, economic downturns are likely to affect both

a firm’s loan demand and the lender’s evaluation of its credit risk. Therefore, it

is essential to disentangle a shift in a firm’s loan demand from a shift in loan

supply. Third, a downturn might affect larger banks, which are more likely

to use risk models to determine capital charges in contrast to smaller banks,

which are likely to retain the traditional approach to capital regulation. Thus,

it can be difficult to determine whether a change in bank lending is driven by

countercyclical changes in capital charges or by the way in which the bank is

affected by a recession shock. Finally, microlevel data on loan balances, along

with information on the regulatory approach used by the bank to determine

capital charges for the respective loans, are difficult to obtain.

1More than 100 countries have implemented Basel II (see Financial Stability Institute (2010))

and many of the world’s larger banks are now using internal rating models to determine capital

charges for individual credit risks.

2Terminology is as follows: capital charges are countercyclical if they increase during an eco-

nomic downturn. If banks react by reducing loans, then lending and credit availability are pro-

cyclical and amplify the business cycle. Procyclical features of Basel II are discussed by Borio,

Furfine, and Lowe (2001), Goodhart, Hofman, and Segoviano (2004), Gordy and Howells (2006),

Rochet (2008), and Gersbach and Rochet (2013). Brunnermeier (2009) and Admati and Hellwig

(2014) discuss how procyclical features of the regulation contributed to the financial crisis.

Procyclical Capital Regulation and Lending 921

We address these issues by examining changes in bank lending in Germany

surrounding the credit risk shock that followed the failure of Lehman Brothers

in September 2008. Economic conditions changed suddenly and dramatically

following the Lehman event, causing a steep and abrupt decline in real sec-

tor expectations. In response to this exogenous increase in credit risk, banks

had to adjust their internal risk estimates, implying an increase in capital

charges for loans under model-based capital regulation. In contrast, capital

charges for loans under the traditional approach to capital regulation, which

do not depend on economic conditions, were unaffected by the credit risk shock.

The fact that only loans under model-based regulation were affected by the

exogenous increase in capital charges allows us to identify its effects on bank

lending behavior. For this purpose, we use data from the German credit reg-

ister, which includes detailed information on the regulatory approach and the

bank’s internal risk estimate for each loan.

German banks began implementing the Basel II regulations before the credit

risk shock. They were allowed to choose between the model-based (i.e., internal

ratings-based, or IRB) approach and the traditional or standard approach (SA)

with fixed risk weights. Those banks that chose model-based regulation (IRB

banks) phased in the new approach over time. Thus, in September 2008, IRB

banks were using the IRB approach for some of their loan portfolios, while other

portfolios were still subject to SA. Since German firms typically have multiple

banking relationships, many firms found themselves in the IRB pool of one

bank and the SA pool of another IRB bank. This institutional setup allows us

to control for heterogeneous effects of the credit risk shock on individual firms’

loan demand or individual banks’ loan supply. In particular, firm fixed effects

control for any demand shocks where our estimates compare the change in IRB

loans to the change in SA loans to a given borrower (compare with Khwaja and

Mian (2008)), while bank fixed effects control for bank-specific supply shocks

and rely on within-bank variation in the adjustment of IRB and SA loans

(compare with Jim´

enez et al. (2014a)). We investigate whether, within firms

that have multiple lending relationships with IRB banks, loans that use the

IRB approach are adjusted differently from loans that remain under SA.

We find that increases in capital charges caused by procyclical regulation

have a strong and economically meaningful impact on the adjustment of loans

over the credit risk shock. In our sample, capital requirements for loans under

IRB rose by about 0.5 percentage points as banks adjusted their internal risk

estimates (given that the minimum capital requirement is 8% of risk-weighted

assets (RWAs), this is a large change). In contrast, capital charges for loans

subject to SA remained constant. Our estimates indicate that, in response to

the shock, IRB banks reduced loans to the same firm by 2.1 to 3.9 percentage

points more when capital charges for the loan were based on internal ratings

(IRB) than when they were based on fixed risk weights (SA).

We further examine whether the increases in capital requirements associ-

ated with the credit risk shock affect the aggregate availability of bank credit

to firms. Interestingly, we find that procyclical effects are even stronger at the

firm level. A firm with 82% of its loans subject to IRB prior to the shock (the 75th