CECL’s January compliance deadline for larger US banks is fast approaching. As everyone expected, the states of readiness vary widely among financial institutions. The CECL team leader at Citi, a charter member of FASB’s Transition Resource Group for Credit Losses, is on record as to how aggressive and committed and successful the bank was in its preparation and how unwelcoming it would be of any further delay or reconfiguration. But other banks continue to besiege their congressmen and the SEC in the vain hope that the cup will pass.
Legislation lingers in both the Senate and the House that would push all CECL deadlines back another year or more – to allow for further impact study – but apparently without the necessary support from Congressional leaders. Major consulting firms that have been laser-focused for the last few years on directing CECL implementation for clients have suggested to RapidRatings that much of the industry’s initial compliance will nevertheless be nominal, slipshod, and in need a major remediation later in the New Year.
Anxiety over the looming deadline extends far beyond reporting entities and their service providers. CECL watchers will recall the thoughtful commentaries that two major institutional investors, Capital Group and Franklin Templeton, submitted to FASB Chairman Golden a year ago. They each made the case for how much harder it was going to be to compare bank performances against one another, if reporting was no longer reflective of incurred losses but rather of losses estimated according to each bank’s own individual loss-modeling and macroeconomic expectations. Among the equity research community, the UBS team of Saul Martinez and Antonio Chapa was first to articulate the problem in published materials, while others have followed with similar cautioning.
None of these observers were likely surprised by the disparate estimations of CECL-related loan-loss reserving that have attended this October's bank earnings calls, increases running anywhere from 5% to 50%.
Stunning confirmation of the analytical challenge also comes this October as the market learns the results of the CECL Benchmarking Study conducted in 1H2019 by the research team of Accenture, GCD and IIF. Eleven major banks took part in the Study – banks sufficiently confident of their modeling by then to conduct the required exercises. They were given identical hypothetical portfolios across asset classes, along with identical macroeconomic scenarios, and asked to apply their CECL modeling in order to generate the loss estimations they would post.
Researchers were shocked by the results, as was the CECL team at RapidRatings. There was no semblance of consensus on any asset class. For example, in the hypothetical portfolio of BB-rated, 5-year unsecured C&I loans, Estimated Probabilities of Default (EPDs) ran from near-zero to nearly 4%.
RapidRatings cannot think of a clearer case for rigorous independent challenger PD modeling under CECL than these wildly divergent numbers. Such modeling will help overly pessimistic entities avoid unnecessary increases to ALLL, while also helping overly optimistic entities avoid the underestimations of loss with which they would otherwise discredit themselves over time.
Excerpts from the Benchmarking Study are expected to appear in a major industry publication before year end, after which the researchers will publish it. RapidRatings will alert blog readers appropriately.
However, RapidRatings would also like to urge the larger industry community – the community beyond the original 11 respondents – to take advantage of the opportunity that researchers are now providing in Round Two of the Study, announced last month. Presumably, numerous banks now have the grounding they did not have several months earlier to enable their participation in these estimated-loss exercises. Newly participating banks will acquire a clear picture of where their models stand in relation to others in the industry. Many interested parties – the industry as a whole, the audit community, regulators and especially the investing public – will get, in turn, an even-better view of the substantial complexities embedded in the pending, widespread compliance with the CECL Standard.
Round Two is described here: