Working paper series // Frankfurt School of Finance & Management 136
Credit risk measurement and management become more important in all financial institutions in the light of the current financial crisis and the global recession. This particularly applies to most of the complex structured financing forms whose risk cannot be quantified with com-mon rating methods. This paper explains the risk associated with leveraged buyout (LBO) transactions and demon-strates the implementation of a new rating method based on a logistic regression (logit func-tion), a rating system commonly used by banks. The system estimates probabilities of default for various time horizons between three months and two years. Input variables contain information about the transaction (based on financial covenants) as well as macroeconomic parameters. The most important factor is a firm’s cyclicality. Leve-rage and capital structure are statistically significant and are also utilized in this ratings sys-tem, however they are far less important compared to cyclicality when this method is em-ployed. The validation results demonstrate a very good calibration and discriminatory power between defaulting and non-defaulting LBO transactions.
Logistic Regression Logit Credit Risk Credit Risk Modeling Rating Probabili-ty of Default PD Basel II Rating Validation Rseudo-R-Square Alpha Error Beta Error Minimum Classification Error Cumulative Accuracy Profile Curve CAP Receiver Operating Characteristic ROC Area Under the Curve AUC Brier Score Bootstrapping Leveraged Buyout LBO Buyout Leveraged Finance Private Equity