As discussed in our response to Question 80, we have little confidence that the community metric multipliers are properly calibrated. As discussed in our response to Question 78, we also suspect the validity of the home mortgage borrower benchmark. Unless these concerns can be addressed, we strongly advise against using them as binding elements in performance measurement, such as in combination with the market benchmarks.
The best way to avoid market underservice is to motivate better performance, particularly by encouraging a race to the top. This can be best accomplished by a combination of elements, as we described at the outset of this comment letter.
The market metric for Outstanding performance is too high. Banks should see that they have a reasonable chance of attaining Outstanding performance. Setting the Outstanding market metric threshold at 125 percent of industry performance – a level the Agencies estimate that no bank with assets exceeding $50 billion would achieve and that we estimate would include banks with only 2 percent of banking system assets – would clearly signal to banks that an Outstanding rating is beyond reasonable reach. Instead, banks would rationally resign themselves to an overall Satisfactory rating, along with perhaps 80 percent of all banks. Moreover, since Retail Lending accounts for 45 percent of the overall rating, and since the market metric is highly likely to be binding in most cases, it is highly unlikely that a bank could achieve an overall Outstanding rating unless it is Outstanding in Retail Lending. If the great majority of banks expect to receive the same Satisfactory rating, there will be little motivation within a given bank to keep up with high performing competitors, and the bank will find no discomfort in the middle of the pack. The NPR’s approach could also make CD performance immaterial to most banks’ overall CRA rating, as we describe elsewhere, which would be a very serious concern.
If Outstanding is beyond reach a bank’s reach, its objective might rationally be only to avoid a Needs to Improve rating, suggesting the bank should be comfortable performing below industry average. Should these dynamics take hold, they could easily feed a decline in lending. We take little solace in the fact that the Agencies would also publish numerical performance scores in addition to ratings. The headline rating will be what matters. This prospect would be a highly counter-productive result for communities.
The Agencies should also reconsider the market metric for High Satisfactory retail lending performance. A High Satisfactory market threshold set at 110 percent of industry performance is problematic because it locks in the conclusion that an estimated 60 percent of the industry is destined to perform weakly. This is the case because when industry performance rises or falls, every percentage of that performance will also rise or fall. We infer that the Agencies propose to set high standards in hopes that banks will seek to exceed them, thereby serving their communities. We doubt this strategy will succeed.
A better way – indeed, a necessary way, in our view – to incent High Satisfactory performance would be to differentiate much more between the points awarded for High Satisfactory and for Low Satisfactory performance. As proposed, the difference is literally minimal: only one point. This signals there is little difference between banks performing at the 15th and 85th (or even 95th) percentiles of the industry. If, however, the points differential is wide enough, it will signal to banks that a High Satisfactory puts them on the road to Outstanding, while a Low Satisfactory warns they could slip to Needs to Improve. 39 Of course, performance conclusions and ratings will aggregate numerous metrics, especially for banks with multiple AAs. Most banks will receive a mix of performance scores. Because the long road to an overall rating starts with the performance scores for each metric, getting the metrics, multipliers, and assigned points right is essential to achieve the best outcomes for LMI people and communities.