As we have said before, asset growth at banks can cover up a lot of problems. Of course, if you believe that, then it goes to reason that the opposite must also be true - decreasing the size of the bank uncovers a lot of problems. The reality is that both statements are true and the concept is unknowingly causing banks problems. Specifically, due to capital constraints, many banks are attempting to deleverage and reduce asset size in order to bring capital ratios back in line. The problem is that a smaller balance sheet can also sometimes result in more required capital.
The issue at hand is one of adverse selection. When banks let assets run off through attrition or asset sales, there is a propensity to apply a strict set of rules across the balance sheet. For example, in an effort to offset losses, banks may sell securities with a certain gain in them. The problem is that these could also be securities that are providing positive risk-adjusted cash flow, so selling them may hurt future earnings (in fairness, however, selling mortgage securities at a high premium/gain may serve to reduce prepayment/yield erosion risk, so sometimes this can also be a good strategy).
Similarly, a bank may not want to renew loans coming up for maturity. This creates a problem. Generally speaking, while some loans of lower credit quality could randomly prepay, it may also be the loans that can be paid off by the borrower or refinanced elsewhere (the higher quality loans) that are the first to go. Letting lower risk/higher quality loans run off can leave a bank with a set of adversely selected loans, thereby requiring more reserves and/or capital to manage.
To see the extent of the damage, we modeled a community bank portfolio with a normal distribution of credit. We started the portfolio with an average probability of default (PD) for the industry of 7.32% (a startlingly high number for many banks to see in print) and created a Monte Carlo simulation that let 10% of randomly selected loans run off. The simulation gave loans that have below a 3.00% PD, a 20% higher incidence of running off. We then let the model run and set a time constraint at 1Y. The result was that the average bank under these constraints would see their average PD rise from 7.32% to 7.97% within 1Y, given no change in credit environment. This means that reserves would have to go up, not down.
Best practices for shrinking the balance sheet usually involve understanding your loan portfolio and utilizing both a credit stress analysis and risk-adjusted loan pricing. The trick is to incentivize riskier loans to go, away while retaining the better risk-adjusted return loans in the portfolio. It also means doing this in such a way that the balance sheet remains diversified and positively selected. Knowing the underlying risk of a loan and pricing gives banks a quantitative basis for selecting loans they want to keep and those they want to let run off.
One example of this is a loan on an industrial property. Back in 2006, the average floating rate loan on a warehouse produced about an 8.0% risk-adjusted ROE. Now, due to negative economics, lower cash flow coverage, a higher LTV, higher capital requirements and slimmer margins; that loan may produce a negative 21% ROE, or a loss of $21k each year per million. Here, a bank may want to offer customers a $40k cash incentive to refinance elsewhere.
Unfortunately, shrinking the balance sheet requires more thought than growing the balance sheet. Unfortunately, in deleveraging randomly, many banks may have exacerbated their capital problem by unknowingly retaining riskier loans and letting the stronger ones pay off. To prevent this, knowing the economics behind each loan and the portfolio is the best course, so a strong economic decision can be made.