Empirical Credit Spread Components of Banks and their Coverage in Structural Models – An Integrated Analysis

Abstract

During the global financial crisis and euro crisis, credit spreads of European banks have increased sharply, which has not only been attributed to default risk. Credit spreads can be estimated empirically or using structural models. In the KMV model, a firm defaults if the market value of assets falls below the Default Point (DP) based on liabilities. The Distance-to-Default (DD) measures how many standard deviations the asset value is away from DP. Based on DD, Moody's Analytics calculates a physical probability of default, the Expected Default Frequency (EDF). The paper analyzes the relation between the KMV structural model and empirical credit spread components. In addition to default risk, empirical credit default swap (CDS) spreads compensate risk-averse investors for liquidity risk and common spread risk. Default risk can generally be divided into a diffusion and jump component. Fundamental structural models cover default by a continuous decrease in the asset value below the default barrier, while unexpected jumps or (liquidity) shocks – such as bank runs – are not captured. Comparing model-based bank EDFs and empirical CDS-implied default probabilities, with a special focus on financial market and liquidity crises, the residual risks are derived. In contrast, previous studies primarily consider non-financial firms and phases of low financial market volatility. The present empirical study is the first integrated analysis of the determinants of one-year CDS spreads and the EDF of European banks from 2004 to mid-2015. Using a panel regression with time fixed effects, the residual impact on the logarithmized CDS-implied default probability, that cannot be explained by the logarithmized EDF, is decomposed into a systematic and idiosyncratic component. Since the global financial crisis and especially during the euro crisis, the EDF structural model assesses the default probabilities of European banks to be lower than the CDS market does – due to market-wide risks. The systematic residual impact increases substantially at the outbreak of both financial market and liquidity crises. At the onset of the global financial crisis and euro crisis, in particular market-wide liquidity risks imply a stronger increase in CDS spreads compared to EDFs. Combining market-based signals from European bank CDS spreads with model-based signals in an integrated analysis, the risk of a financial market crisis is derived. Market-wide indicators of funding and liquidity risk provide a statistically significant and economically important explanatory power for the systematic residual impact. Moreover, the risk premium is an essential determinant. Compared to alternative crisis indicators, the systematic residual impact achieves an equally good and during the euro crisis even better performance.