Probability of Default (PD) Model to Estimate Ex Ante Credit Risk

A genuine measure of an ex ante credit risk links borrowers’ financial position with the odds of default. Comprehension of borrower’s financial position is proxied by the derivatives of its filled financial statements, i.e. financial ratios. To measure an ex ante credit risk, one needs a forward-looking estimate. We identify statistically significant relationships between the shortlisted financial ratios and the subsequent default events. To estimate the odds of the borrower to default on its obligations, we simulate its probability of default at a horizon of one year. We horse run the constructed PD model against the alternative measures of ex ante credit risk that the related literature on bank risk-taking widely uses: credit quality groups and credit spreads in interest rates. We compare the results obtained with the PD model, and with the alternative approaches. We find that the PD model predicts the default event more accurately at a horizon of one year. We conclude that the developed measure of ex ante credit risk is feasible for estimating the risk-taking behaviour by banks and analysing the shifts in portfolio composition with the sufficient degree of granularity. The model could be used in applied research as the tool for measuring ex ante credit risk based on micro level data (credit registry).