Analysis of Concentration Risks in Credit Portfolios

In this empirical study, I explore two hypotheses about concentration risk in credit portfolios. The first hypothesis examines the inequality of ratios of granularity adjustments (GAs) and mean exposures using data from the company called “Lending Club”. I created four sample portfolios with increasing name concentration risk, calculated the optimal probabilities of default, and used the Gordy & Lütkebohmert method to compute GAs. After bootstrapping 400 portfolios from the initial four and applying the Kolmogorov-Smirnov test, I supported hypothesis #1. For the second hypothesis, I used data from the Central Bank of the Russian Federation to estimate economic capital add-ons for corporate loan portfolios, applying Pykhtin's multifactor model. By comparing the slopes of the regressions of economic capital add-ons on Herfindahl-Hirschman Index values given the Russian and British data, I found a statistically significant difference as indicated by the Chow test. In the conclusion section, I discuss the results, limitations, and potential practical implementations.