Credit analysis in times of change: On the benefits of internal and external ratings from a historical and current perspective

How valid are credit ratings? Although they have proven useful over time, they should not serve as a substitute for independent reasoning and analyses by investors and lenders
In the closing panel discussion of the symposium on Oct. 27, 2021, jointly organized by The Center for Financial Studies, Goethe-University Frankfurt, and the Institute for Banking and Financial History (IBF), Jens Schmidt-B├╝rgel, Managing Director of Moody's Germany, also said, that for rating agencies trust and reputation are essential. However, expectations about what ratings can accomplish should not be overestimated. He stressed that ratings have the status of opinions. Users of external ratings should always read the underlying analyses as well. Matthias Danne, a member of the Executive Board of DekaBank, which uses ratings in various contexts - as an issuer and as an investor, from the treasury to capital market business - explained that ratings provide a "feel-good" factor in investment decisions and that there is great confidence in external ratings. It is not completely clear how rating agencies make their conclusions, he said. Aggregating rating assessments to a letter abbreviation means  sacrificing the nuanced picture but is necessary to shape market liquidity. Senior Expert Hiltrud Thelen-Pischke, Steinbach, recalled that the use of external ratings in Europe has a relatively short history. Furthermore, banks had just started to work with internal ratings in the 1980s, when regulation in the context of Basel II increasingly focused on external ratings. Since the global financial crisis, the European Securities and Markets Authority ESMA tried to reduce reliance on external ratings. Steffen Kern, ESMA's chief economist, postulated that rating agencies should be transparent about the fundamentals of their credit ratings, as he had already pointed out in his previous presentation on rating and regulation in times of financial crisis and pandemic. He once again stressed that a broader range of ratings could help to ensure sound assessments as 90 percent of ratings are currently issued by the top three rating agencies. Kern also pointed out that it was only in the context of the global financial crisis rating agencies were identified as important but at the same time problematic market players. Credit ratings were subsequently subjected to regulation, measuring their quality. As a result, ratings so far have proven reliable in the crisis caused by the Covid19 pandemic. In his commentary, Jan Krahnen, Director of the Leibniz Institute for Financial Market Research SAFE and Professor of Finance at Goethe University Frankfurt, suggested reviewing the assumption of a market structure for ratings that is more or less concentrated on three providers, might be problematic. In general, he recommended that to question regulation from a methodological angle. Krahnen also suggested investigating the accuracy of ratings in long-term empirical studies. Katja Langenbucher, Chair of Civil Law and Banking Law at the Goethe University Frankfurt and Associate Professor at SciencesPo Paris, argued from a different perspective for a closer look at credit scoring and its regulation. She stated that with the increasing use of AI or Big Data in the assessment of credit applicants, either by growing in-house data pools or the purchase of external data and thus via algorithms, a variety of potentially discriminating information comes into play. However, current regulatory initiatives classifying credit scoring with AI as "high-risk AI systems" are inadequate or inefficient, as normative fundamentals are not properly defined yet. Furthermore, there is a risk that by focusing on banking supervision potentially discriminatory actors such as credit reporting agencies would be overlooked, she said.
In his commentary, Mark Wahrenburg, Chair of Banking at Goethe University in Frankfurt, pointed out that banks' internal rating systems are a black box for the public and that external credit scoring is already used in banks' internal credit assessments. While he doubted that algorithm-based machine learning methods were at this stage playing a significant role, he pointed out, among others, the problem that AI-based scoring methods resist transparency requirements simply because they were not "explainable." It is to be expected that the actual objective of the regulation will not be achieved, he said, but that the compliance requirements for banks will increase and evasive shifts to the less regulated market players will occur. Thomas Hartmann-Wendels, Director of the Seminar for General Business Administration and Banking at the University of Cologne, focused on the possible methodologically founded alternatives and problems of using banks' internal risk parameters. He pointed out that  originators and regulators of internal ratings may be guided by misleading incentives. Beyond the data problems, these can lead to systematic underestimation of loss and default risks. Accordingly, capital adequacy is too tight. Carsten Burhop, holder of the Chair of Constitutional, Social, and Economic History at the University of Bonn addressed in his paper a central aspect of the research debate regarding the banking and financial crisis of 1931, namely the thesis of banking failure. His analysis was devoted to possible organizational deficits of banks concerning the practice of internal bank credit assessment, which he examined based on the Reichsbank's credit lists as well as by doing a case study on the systematic credit assessment of ADCA as one of the largest regional commercial banks during the interwar period. While a well-developed system of credit analysis existed, inadequacies emerged during the 1920s. Some were structural, others due to an increasingly lax approach to the applicable lending standards. In his commentary, Dieter Ziegler, holder of the Chair of Economic and Business History at the Ruhr University in Bochum, supplemented this picture with the example of an increasingly deficient practice of credit assessment at one of Berlin's major banks during the 1920s. This was partly due to structural problems and an increasingly lax approach to lending standards during the 1920s.