What Drives the Difference in Probability of Default from Reduced Form- and Structural Approaches

Abstract: This Master Thesis successfully explains the difference in probability of default implied by Credit Default Swaps, traded by the market, and the benchmark Moody’s EDFTM. The difference is explained by the market price of risk, related to the Girsanov kernel, allowing us to transform the risk neutral measure Q to the physical measure P. This market price of risk is modeled with a log-linear multivariate regression model combined with elastic net, using market data. The predictability of the model is examined. The market price of risk is seen to be mostly dependent on market sentiment, in front of firm specific factors and liquidity. The analysis is made for AB Volvo, Stora Enso Oyj and TeliaSonera AB on data from 2006 - 2014. The work was carried out at Swedbank.