Barry Mann. assistant vice president-finance at First Federal of Richmond. Ind., has analyzed the proposed interest rate risk regulations by the Federal Reserve Board and the Office of Thrift Supervision. In tis commentary prepared for the Mortgage Marketplace, he suggests that both proposals have flaws and suggests a compromise. (See related story on page 3.)
Recently 1 participated on a committee of Savings and Community Bankers of America member institutions that compared the Office of Thrift Supervision and Federal Reserve Board interest rate risk components. About 70% of the Institutions that participated had a lower component under the Federal Reserve model.
The consensus of the committee was that the more traditional thrifts did better under the OTS model while those thrifts which are following community banks' business plans did better using the Fed.
As of June 30. my $180 million thrift's IRR component was $700,000 under the Fed method and $1.7 million under the OTS model. I also performed the two calculations for another Indiana institution that is a traditional thrift with a 50/50 split between fixed- and adjustable-rate mortgages. This thrift would have a 830,000 capital requirement under both.
My institution's criticism of the methodology of the OTS market value model lies in how it handles our nontraditional assets. primarily real estate mortgage investment conduits and nonmortgage loans.
The OTS is in the process of revising its approach to how it estimates market value for Remics. Its initial approach was to lump Remics into eight categories and pick a particular Remic tranche to represent all securities in that category. This approach was doomed because as Remics became more popular. the complexity of Remics grew.
Currently. If rates rise 200 basis points. the OTS model says that our Remic portfolio's market value would fall by 1096 while the fixed-rate mortgage (FRM) portfolio would fall by Just 696. This is strange since the weighed average life of the FRM portfolio is roughly twice the length of the Remic portfolio.
The OTS recognized its initial approach did not work and has dumped the task of calculating market values of mortgage derivatives back onto institutions and. ultimately. the brokers and Bloomberg. I feel more comfortable with them doing the calculations than OTS.
The Fed takes a simpler approach and one that could be advantageous for thrifts. If the mortgage derivative passes the Federal Financial Institutions Examination Counsel stability test. then it is placed into the time banks based on its current average life and not the final maturity. The current weighted average life of our Remic portfolio is Just three years while the weighted average final maturity is in excess of 20 years. Thus. a Remic 18 treated the same as a car loan. If a Remic does not pass the FFIEC test. then the institution would get a market price in the current and up 100bp case from Bloomberg like the OTS requires for all Remics.
Another disappointment of the OTS market value model concerns that way it treats nonmortgage loans. Consumer and commercial loans show surprising IRR volatility in market value under OTS approach. As of June 30. the OTS model estimated that our consumer loans lost 496 of market value and commercial loans 5% if rates rise 200bp. Yet. the weighted average life of the commercial loan portfolio is roughly two years while the WAL of the consumer loan portfolio is slightly longer.
Compared to the 696 loss in market value with FRMs mentioned earlier. I feel that the OTS does not recognize the IRR benefits of nonmortage loans. These intermediate-term loans act as a natural hedge against long-term FRMs. in fact. I like nonmortgage loans better than ARMs to hedge my rate risk because they make sense in all rate scenarios. ARM prepayments can hurt an institution in falling rate scenarios as borrowers switch to fixed-rate mortgages.
Consumer loans are discounted by the OTS based on monthly amortized cashflows with no prepayment rate assumption. Using a conservative 1096 prepayment assumption on a five-year car loan reduces the WAL from three years to two years and lowers IRR exposure estimates. The Ors admits that consumer loans prepay but uses a zero percent rate because they feel no reliable numbers exist. I suggest that the growing market of asset-backed securities offer a good source for prepayment assumptions.
Commercial loans are treated more harshly by the OTS since these loans are assumed not to amortize or prepay. The OTS believes that interest-only commercial loans are prevalent in the market place. While inventory loans are short-term and nonamortizing. fixed-rate loans which mature in more than a year definitely amortize. Thus a five-year lease is treated like Treasury instead a loan with a three-year WAL. Amortization can do little to distort the market value of short-term loan that doesn't amortize. but the lack d amortization distorts the market value of an asset that does.
Perhaps the biggest problem with the OTS proposal is not the methodology of the model, but the methodology of the component. Both the OTS and the Fed fail to realize what is truly at risk for changes In Interest rates. Capital calculated under generally accepted accounting principles is what is at risk. not market value.
Given the present low-rate environment. I find it strange that my thrift has to hold any capital against IRR when our present paper gains well exceed any potential loss from risIng rates. First Federal has $20 million in capital and current market value of $30 million. If rates rise 100bp. the market value falls to $27 million. If rates rise 200bp. then the market value falls to $24 million. Still, no GAAP capital of First Federal is exposed to rising rates.
If I were to revise the OTS proposal. I would like to see the OTS adopt a carrot-and-stick approach. The carrot should be given to those Institutions that pose no threat to the Savings Association Insurance Fund from their IRR exposure. These are primarily small. well-capitalized Institutions. Capital will always be the best protection against interest rate risk. These thrifts should be allowed to calculate their IRR exposure usIng the simpler Fed model first. If the results Indicate that no capital is needed. then the Institution would not fill out the OTS form.
If the Instutution would have to hold capital under the Fed model, it could elect to fill out the OTS forms and calculate the component using book value instead of current market value as the measuring poInt. If the market value of the Institution after rates rise 200bp is still above its GAAP book value. the Institution would not have to hold any capital against its IRR.
The stick would be reserved for larger thrifts and less capitalized thrifts. They would have to fill out the OTS forms and debate the results with the regulators. Again, I would only require these Institutions to hold capital based on GAAP capital exposed to rates rising 200bp and not paper market gains.