The application of data envelopment analysis to credit measurement

In competitive markets only the strong survive. For an institution tosurvive it needs to achieve high levels of performance through continuedimprovements and learning. Credit risk measurement has come underintense scrutiny with the recent Basel II Accord, an Accord that obligesbanks to seek more efficient means for the management of their credit risk.In this dissertation I examine Data Envelopment Analysis (DEA), itsextensions and applications to credit risk measurement. DEA is aperformance measurement technique used to evaluate relative efficiency ofa group. It is a quantitative model with a solid mathematical andeconomic underpinning; solving several linear programs simultaneously. Itsgreatest advantage is a feature that allows it to process multiple inputsand multiple outputs, thus uncovering relationships which remain hiddenfrom other methodologies. We consider DEA as a useful tool for improvingcredit risk measurement within peer group analysis, a supplement orcomplement to the credit ratings and a validation tool for credit ratings.We apply DEA models to two credit risk measurement areas: CorporateCredit Risk and Country Risk. In the first application we apply DEA tocorporate entities and compare efficiency to corporate credit ratings.Overall we find efficiency and credit rating have common elements. Thisconfirms the common sense notion that corporates receiving better creditratings are more efficient. In the second application we apply DEA tocountries. We identify those countries that are performing better than therest based on their efficiency. We discover that when a country’s efficiencyis compared to country credit rating, efficiency and credit rating aremeasuring two different but equally important aspects. We find thatefficiency is a good indicator of relative economic acceleration. We observethat efficiency has the potential to identify future improvements in countrycredit quality. Based on the application of DEA to Corporate and CountryCredit Risk, we recommend the use of DEA as a complementary tool tocredit rating models and a means of facilitating better measurement andmanagement of credit risk within banks.