Tobias Berg

Does Contingent Capital Induce Excessive Risk-Taking?

In this paper, we analyze the effect of the conversion price of CoCo bonds on equity holders’ incentives. First, we use an option-pricing context to show that CoCo bonds can magnify equity holders’ incentives to increase the riskiness of assets and decrease incentives to raise new equity in a crisis in cases in which conversion transfers wealth from CoCo bond holders to equity holders. Second, we present a clinical study of the CoCo bonds issued so far. We show that (i) almost all existing CoCo bonds are designed in a way that implies a wealth transfer from CoCo bond holders to equity holders at conversion and (ii) this contractual design is reflected in traded prices of CoCo bonds. In particular, CoCo bonds are short volatility with a magnitude five times greater than that which can be observed for straight bonds. These results are robust and economically significant. We conclude that the CoCo bonds issued so far can create perverse incentives for banks’ equity holders.

A Certification Model for Regulatory Arbitrage:Will Regulatory Arbitrage Persist under Basel III?

Based on anecdotal evidence, regulatory arbitrage was a major catalyst of the recent financial crisis. However, regulatory arbitrage is both theoretically and empirically not yet adequately explored. This article develops a theoretical model of regulatory arbitrage. The model is based on a set up where regulatory risk weights act as a certification of the riskiness of a bank’s opaque balance sheet. When regulatory risk weights are misspecified, regulatory arbitrage for the (informed) banks exists. The authors empirically demonstrate that regulatory arbitrage opportunities between the trading business and the loan business existed under the Basel I and Basel II framework and are likely to subsist under the Basel III regime. Their analysis is based on the so-called asset correlation parameter that plays a crucial role for capital requirements under Basel II and Basel III.

Estimating Equity Premia from CDS Spreads

We propose a new approach to estimate the equity premium using CDS spreads and structural models of default. Our estimates yield equity premia of 6.50% for the U.S., 5.44% for Europe and 6.21% for Asia based on 5-year CDS spreads from 2003-2007. Due to some conservative assumptions these estimates are upper limits for the equity premium. Using 3-, 7- and 10-year CDS maturities yields similar results and offers an opportunity to estimate the term structure of risk premia. Although our estimator is developed in a Merton framework it is robust with respect to model changes. In fact, we obtain similar results when extending the approach to a first-passage-time framework, strategic default frameworks or a framework with unobservable asset values (Duffie/Lando (2001)).

What Explains the Difference in Leverage between Banks and Non-Banks?

Banks have much more leverage than nonbanks. In this article, we use a joint sample of banks and nonbanks between 1965 and 2013 to analyze the determinants of this leverage difference. We find that a single factor, asset risk, is able to explain up to 90% of this difference. Banks’ assets consist of a diversified portfolio of nonbank debt. Therefore, banks have much lower asset risk than do nonbanks. Because asset risk is a major determinant of capital structure choice, this factor is able to explain a large fraction of the difference between bank and nonbank leverage.

Loan Officer Incentives, Internal Rating Models and Default Rates

There is increasing reliance on quantitative complex models, such as internal ratings based (IRB) models for bank regulation, with much resources being spent on model validation exercises. We argue that a significant cost of IRB models that is not well understood or monitored is the change in loan officer incentives down the line. Using proprietary data on almost a quarter million loan applications, we show loan officer incentives significantly skew ratings even if the quantitative model is correct and there is no subjectivity in the system. These incentive effects have a first order effect on bank profitability. Incentives influence the hard information reported by loan officers and thus change the link between hard information and default probabilities in a way not captured by regular model validation exercises. Banks and regulators need to take these effects into account when using internal ratings for risk assessment and regulation.

Extracting the Equity Premium from CDS Spreads

The capital asset pricing model that relates a stock’s expected risk premium to the risk premium on the market portfolio and the stock’s beta coefficient relative to the market is one of the pillars of modern finance. Unfortunately, expected equity risk premia are not observable, and extracting them from realized returns produces very noisy estimates at best. In this article, Berg and Kaserer offer a new approach based on analyzing spreads on credit default swaps (CDS). A “structural” default model computes a firm’s probability of default as a function of the mean and standard deviation of the changes in the firm’s asset value and how far that value can fall before it would precipitate insolvency. This (risk-neutral) probability can be deduced from the market spread on the firm’s CDS. Moody’s KMV’s expected default frequency (EDF) is an estimate of the empirical probability of default. Combining the risk-neutral and the empirical default probabilities allows the expected equity risk premium to be extracted. The article develops the theory and applies it to the individual stocks contained in the CDX.NA.IG CDS index.