Conditions of banks in the Ninth Federal Reserve District have improved materially over the past couple of years by most balance sheet measures. A summary measure of conditions as assessed by bank supervisors’ bank composite ratings shows less improvement. Why this difference in improvement, and how long might this gap continue?
First, a few facts. Accounting measures have improved in the aggregate. Consider asset quality for the median district bank: A standard measure I prefer (the ratio of noncurrent plus delinquent loans to capital plus the allowance) hit its peak three years ago in the first quarter of 2009 at 22 percent. This same figure as of first quarter 2012 was about 13 percent. Even measures that are not near pre-crisis levels show gains. Year-over-year loan growth for the median district bank is very close to zero as of first quarter 2012; loan growth was about -4 percent in mid-2011 at its nadir.
As of first quarter 2012, about 30 percent of district banks had ratings in the “3” or below range (“less than satisfactory”). About 11 percent of banks are “4” or “5” rated, which puts a bank into “problem” status. There has been some improvement since the third quarter of 2010 when 33 percent of District banks were in less than satisfactory condition. But the total improvement seems small relative to the accounting gains.
What drives the difference?
Ratings lagged the weakening in conditions on the way down. The asset quality data just referenced had the worst point about 18 months before the low point for the supervisory measure. It is not surprising that ratings are lagging the transition on the way up.
I think two other features of supervisory ratings explain the lag.
First, bank supervisors tend to avoid short-term volatility in ratings. Ratings, particularly weak ones, have consequences. Problem banks, for example, will almost always face a formal and public legal action from their supervisor. A bank supervisor will want to make sure that the factors driving such an action are persistent before downgrading, guaranteeing that supervisory ratings will lag accounting data during downturns. A preference for persistence means that the numbers will likely improve a bit before supervisors improve their ratings.
Second, supervisory ratings reflect more than “the numbers.” A bank, for example, could face a less-than-satisfactory rating if its credit administration falls below acceptable standards even if accounting data are not that bad.
Based on the data I have seen from Federal Reserve exams and conversations with other bank supervisors, I expect the overall improvement in district bank ratings to be more noticeable by the end of 2012. In 2011, about 20 percent of the roughly 70 Federal Reserve Bank of Minneapolis exams led to upgrades in the composite rating. That level of upgrades should continue in 2012, which would reduce the number of problem banks and banks in less-than-satisfactory condition. I see this trend as consistent with recent Federal Reserve guidance on upgrades for banks with assets less than $10 billion (SR 12-4). This guidance notes that upgrades can occur when firms demonstrate improvements in condition and risk management that are likely to continue. There is no fixed time period before an upgrade can occur.