THINKING OF LOAN LOSS RESERVES

We would like to propose another New Year's resolution for some banks – stop thinking about loan loss reserves as an income statement item. How many times have you heard a banker say, "It is an 8% loan, but we have to take 1% in reserves, so it is really 7%"? For starters, the math doesn't work when you consider the amortization of the reserve. Second, if you don't realize a 1% loss, don't you get your reserve back into earnings? You do, and as a result, many banks underestimate a loan's return, forcing them to make suboptimal decisions. In our opinion, reserves are nothing more than glorified capital. We say "glorified" because they have all the properties of capital, without hurting ratios. In a perfect world, reserves should be set to cover the expected loss of a loan. That is to say they should cover any losses after recoveries. Unfortunately, many banks do not keep track of actual losses (and the cost of workouts) for a static set of credit underwriting standards, nor do banks generally look at loans on a portfolio basis over a significant period of time. As a result, many can only "ballpark" what reserves should be and largely rely on setting allowances based on peers. The problem with this is that peer levels may not be the most accurate for a particular bank and tend to be overly cautious. In support of this point, we track more than 450k independent bank loans throughout the country in our Credit Shock and Loan Pricing Models. If you lump them all together in one aggregate portfolio, look at seasoning, cross-correlations, historical performance and expected performance, the average loan loss reserve works out to be almost 78bp. As of Sept. 2006, independent banks held loan loss allowances at 1.28% of total loans. Admittedly, this is an apples-to-oranges comparison. Reserving at the expected loss is the same as looking at the statistical mean or average. In other words, half the time bankers may be taking losses in excess of reserves. Since loan losses can be volatile and CEOs hate reporting decreased earnings, most banks instinctively reserve to a "worst case" scenario. While this is a disservice to share price performance (as current accounting earnings are being decreased at a multiple, because banks forego current earnings for a future "worst case" that may never come), it is a perfectly logical conclusion. For an apples-to-apples comparison, if you take 2 standard deviations from the mean loan loss for independent banks, reserves should be around 1.36%. This level should cover losses about 95% of the time for an average bank without loan allocation concentrations. At 1.28% reserves, the average independent bank is protecting themselves in about 90% of the cases. We are not trying to quibble with bankers over where they set their reserves. Our point here is that if bankers talk about their return on loans after the allowance, then they are underestimating the return about 90% of the time. Loans should be originated on an economic basis, not accounting or regulatory. To risk-adjust the loan, look at what the annual expected loan loss will be (net of recoveries) and then decrease the yield by that amount (in order to compare the potential loan with other investment alternatives). This gives a much better representation of the loan's return and keeps the calculation of loan loss allowance separate and where it should be – on the balance sheet.