The Sensitivity of the Loss Given Default Rate to Systematic Risk: New Empirical Evidence on Bank Loans

We verify the existence of a relation between loss given default rate (LGDR) and macroeconomic conditions by examining 11,649 bank loans concerning the Italian market. Using both the univariate and multivariate analyses, we pinpoint diverse macroeconomic explanatory variables for LGDR on loans to households and SMEs. For households, LGDR is more sensitive to the default-to-loan ratio, the unemployment rate, and household consumption. For SMEs, LGDR is influenced by the total number of employed people and the GDP growth rate. These findings corroborate the Basel Committee’s provision that LGDR quantification process must identify distinct downturn conditions for each supervisory asset class.

Loss given default of residential mortgages in a low LTV regime: Role of foreclosure auction process and housing market cycles

Using loan-level foreclosure auction data we study the loss given default (LGD) of defaulted residential mortgages originated in Korea, a low LTV regime. We find that senior mortgages generate very low loss rates (5–10%) while losses of subordinated claims are in 30–50% range. We document the effects of housing market cycles on loss severity by showing that collateral characteristics that are overvalued during the boom increase loss severity during the market downturn. We also investigate how a broad set of time-of-origination and post-origination information on loan, collateral and borrower characteristics and foreclosure auction process influence the LGD of residential mortgages.     

Mortgage Default: Classification Trees Analysis

We apply the powerful, flexible, and computationally efficient nonparametric Classification and Regression Trees (CART) algorithm to analyze real estate mortgage data. CART is particularly appropriate for our data set because of its strengths in dealing with large data sets, high dimensionality, mixed data types, missing data, different relationships between variables in different parts of the measurement space, and outliers. Moreover, CART is intuitive and easy to interpret and implement. We discuss the pros and cons of CART in relation to traditional methods such as linear logistic regression, nonparametric additive logistic regression, discriminant analysis, partial least squares classification, and neural networks, with particular emphasis on real estate. We use CART to produce the first academic study of Israeli mortgage default data. We find that borrowers features, rather than mortgage contract features, are the strongest predictors of default if accepting icbadli borrowers is more costly than rejecting good ones. If the costs are equal, mortgage features are used as well. The higher (lower) the ratio of misclassification costs of bad risks versus good ones, the lower (higher) are the resulting misclassification rates of bad risks and the higher (lower) are the misclassification rates of good ones. This is consistent with real-world rejection of good risks in an attempt to avoid bad ones.     

Pricing Sovereign Credit Risk of Poland: Evidence from the CDS Market

We analyze the market assessment of sovereign credit risk using a reduced-form model to price the credit default swap (CDS) spreads, thus enabling us to derive values for the probability of default (PD) and loss given default (LGD) from the quotes of sovereign CDS contracts. We compare different specifications of the models allowing for both fixed and time-varying LGD, and we use these values to analyze the sovereign credit risk of Polish debt throughout the period of a global financial crisis. Our results suggest the presence of a low LGD and a relatively high PD during a recent financial crisis.     

Boom–Bust Cycles and the Forecasting Performance of Linear and Non-Linear Models of House Prices

The tremendous rise in house prices over the last decade has been both a national and a global phenomenon. The growth of secondary mortgage holdings and the increased impact of house prices on consumption and other components of economic activity imply ever-greater importance for accurate forecasts of home price changes. Given the boom–bust nature of housing markets, nonlinear techniques seem intuitively very well suited to forecasting prices, and better, for volatile markets, than linear models which impose symmetry of adjustment in both rising and falling price periods. Accordingly, Crawford and Fratantoni (Real Estate Economics 31:223–243, 2003) apply a Markov-switching model to U.S. home prices, and compare the performance with autoregressive-moving average (ARMA) and generalized autoregressive conditional heteroscedastic (GARCH) models. While the switching model shows great promise with excellent in-sample fit, its out-of-sample forecasts are generally inferior to more standard forecasting techniques. Since these results were published, some researchers have discovered that the Markov-switching model is particularly ill-suited for forecasting. We thus consider other non-linear models besides the Markov switching, and after evaluating alternatives, employ the generalized autoregressive (GAR) model. We find the GAR does a better job at out-of-sample forecasting than ARMA and GARCH models in many cases, especially in those markets traditionally associated with high home-price volatility.     

Expected Default Probabilities in Structural Models: Empirical Evidence

We apply a set of structural models (Black and Cox 1976; Collin-Dufresne and Goldstein 2001; Ericsson and Reneby 1998; Leland and Toft 1996; Longstaff and Schwartz 1995; Merton 1974) to estimate expected default probabilities (EDPs) for a sample of failed and non-failed UK real estate companies. Results are generally consistent with models’ predictions and estimates of EDPs for different models are closely clustered. The results of z-scores and synthetic ratings misclassify 33% of the total sample in contrast to 8% misclassification by structural models. Further analysis of EDPs based on logistic regressions suggests the observed misclassification of the companies by structural models is due to special company management and/or regulatory circumstances rather than limitations of these models.      

Modeling severity risk under PD–LGD correlation

In this article, a generic severity risk framework in which loss given default (LGD) is dependent upon probability of default (PD) in an intuitive manner is developed. By modeling the conditional mean of LGD as a function of PD, which also varies with systemic risk factors, this model allows an arbitrary functional relationship between PD and LGD. Based on this framework, several specifications of stochastic LGD are proposed with detailed calibration methods. By combining these models with an extension of CreditRisk+, a versatile mixed Poisson credit risk model that is capable of handling both risk factor correlation and PD–LGD dependency is developed. An efficient simulation algorithm based on importance sampling is also introduced for risk calculation. Empirical studies suggest that ignoring or incorrectly specifying severity risk can significantly underestimate credit risk and a properly defined severity risk model is critical for credit risk measurement as well as downturn LGD estimation.     

Industry and time specific deviations from fundamental values in a random coefficient model

The paper analyzes the relationship between stock prices and fundamentals for a large sample of US stocks in the last 10 years using a random coefficient model. Heterogeneity and omitted variable bias are properly taken into account with model coefficients being allowed to vary across time and industries. The random coefficient model allows to track waves of reliance on analysts’ forecasts and nonfundamental stock price components across time and clearly identifies the growth of the nonfundamental component in the long 1991–2000 swing.      

Credit Risk and the Term Structure of Lease Rates: A Reduced Form Approach

Previous research either assumes default free leases or leases subject to default risk using a structural approach. However, structural credit risk models suffer from a common criticism that the firm’s asset value process is unobservable. We develop a reduced form credit risk model for leases that avoids making assumptions regarding unobservable asset valuation processes. Furthermore, we assume a correlated market and credit risk that provides us with a simple analytic formula for valuing defaultable lease contracts. Numerical analysis reveals that tenant credit risk can have a substantial impact on the term structure of leases. Finally, we use the model to demonstrate the implied lease term structure for a set of retail and financial firms in the Fall of 2000.

Explicit tests of contingent claims models of mortgage default

This paper provides explicit and powerful tests of contingent claims approaches to modeling mortgage default. We investigate a model of frictionless default (i.e., one in which transactions costs, reputation costs, and moving costs play no role) and analyze its implications-the relationship between equity and default, the timing of default, its dependence upon initial conditions, and the severity of losses. Absent transactions costs and other market imperfections, economic theory makes well-defined predictions about these various outcomes.The empirical analysis is based upon two particularly rich bodies of micro data: one indicating the default and loss experience of all mortgages purchased by the Federal Home Mortgage Corporation (Freddie Mac), and a large sample of all repeat sales of single family houses whose mortgages were purchased by Freddie Mac since 1976.     

Bank Liability Structure,FDIC Loss,and Time to Failure: A Quantile Regression Approach

Deposit insurers are particularly concerned about high-cost failures. When the factors driving such failures differ systematically from the determinants of low- and moderate-cost failures, a new estimation technique is required. Using a sample of more than 1,000 bank failures in the U.S. between 1984 and 2003, I present a quantile regression approach that illustrates the sensitivity of the dollar value of losses in different quantiles to my explanatory variables. These findings suggest that reliance on standard econometric techniques results in misleading inferences, and that losses are not homogeneously driven by the same factors across the quantiles. I also find that liability composition affects time to failure.

Housing-Price Cycles and Prepayment Rates of U.S. Mortgage Pools

Empirical mortgage prepayment models generally have trouble explaining differences in mortgage-prepayment speeds among pools with similar interest rates on the underlying mortgages. In this article, we model some of the sources of termination heterogeneity across mortgage pools, particularly the role of regional variations in housing prices in generating atypical prepayment speeds. Using a sample of Freddie Mac mortgage pools from 1991 to 1998, we compare two classes of empirical models: a rational option-pricing model using a backward-solving pricing algorithm and an empirical hazard model. In both empirical estimation strategies, we find evidence that differences in house-price dynamics across regions are an important source of between-pool heterogeneity. This finding is then shown to be robust to alternative ways of parameterizing pool heterogeneity in mortgage termination models.