Bancos

J. Finan. Intermediation 17 (2008) 295–314 www.elsevier.com/locate/jfi

Bank capital structure and credit decisions
Roman Inderst a,b,c,d , Holger M. Mueller c,d,e,∗
a University of Frankfurt (IMFS), Germany b London School of Economics, UK c CEPR, UK d ECGI, Belgium e New York University, USA

Received 28 December 2005 Available online 21 March 2008

Abstract This paper argues that banksmust be sufficiently levered to have first-best incentives to make new risky loans. This result, which is at odds with the notion that leverage invariably leads to excessive risk taking, derives from two key premises that focus squarely on the role of banks as informed lenders. First, banks finance projects that they do not own, which implies that they cannot extract all the profits. Second, banksconduct a credit risk analysis before making new loans. Our model may help understand why banks take on additional unsecured debt, such as unsecured deposits and subordinated loans, over and above their existing deposit base. It may also help understand why banks and finance companies have similar leverage ratios, even though the latter are not deposit takers and hence not subject to the sameregulatory capital requirements as banks. © 2008 Elsevier Inc. All rights reserved.
JEL classification: G21; G32

1. Introduction This paper calls into question a common benchmark in the literature, namely, that all equityfinanced banks have first-best incentives to take on risks. We argue instead that all-equity financed banks are inefficiently conservative in their credit decisions. Leverage is thereforebeneficial, at
* Corresponding author at: Department of Finance, Stern School of Business, 44 West Fourth Street, New York, NY 10012, USA. Fax: +1 212 995 4233. E-mail addresses: r.inderst@lse.ac.uk (R. Inderst), hmueller@stern.nyu.edu (H.M. Mueller).

1042-9573/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.jfi.2008.02.006

296

R. Inderst, H.M. Mueller / J.Finan. Intermediation 17 (2008) 295–314

least up to a certain point, as it induces banks to take on more risks, thereby mitigating their excessive conservatism. Our theory of optimal bank leverage may help understand why financial institutions, in contrast to non-financial firms, are so highly levered. The argument we present is based on the role of financial institutions as providers of loans and isnot mechanically linked to their role as deposit takers. This is important, for deposit-taking financial institutions have substantial liabilities over and above their deposit base, e.g., in the form of subordinated debt. For example, in 2000 commercial banks’ nondeposit liabilities (e.g., notes and bonds) accounted for 26.8 percent of their total liabilities (Saunders and Cornett, 2003). Ourmodel builds on the model of informed lending by Inderst and Mueller (2006, 2007). Two assumptions are key in that model. First, as banks do not own the projects in which they invest, competition ensures that they cannot extract all of the proceeds from the project. Second, banks are sophisticated lenders who conduct a credit risk analysis before originating new loans. Similar to Stein’s (2002) notionof “soft information,” the loan officer receives an informative but non-contractible signal about the project to be financed. Loan officers, whose primary task is to make informed credit decisions using their own judgment, can draw on personal experience from making loans to similar firms in the past.1 Unlike this paper, the focus in Inderst and Mueller (2006, 2007) is on the contract design with theborrower. Inderst and Mueller (2006) focus on the optimal security design in a setting with a continuum of cash flows, while Inderst and Mueller (2007) focus on the optimal use of collateral in contracts with the borrower. In both papers, lenders are assumed to be all-equity financed. Hence, the issue of the lender’s own capital structure is ignored. Since all-equity financed banks are too...