We are seeing more and more of what the street calls “structured repo” for funding purposes. These transactions are much like Federal Home Loan advances in that they provide funding, often at artificially low rates, which are collateralized with treasuries, agencies, mortgage-backed securities or investment grade corporate bonds. Usually, the haircuts on these deals are in the 90 percent to 99 percent range. That is, for a 90 percent haircut, the dealer will lend 90 cents on the $1. During the term of the repo, the dealer marks the position to market and may require additional margin to collateralize the repo if the borrower is seriously underwater due to changes in market rates.
A typical example of a structured repo is a five-year, no-call two-year with monthly or quarterly Bermudan options owned by the dealer after the two-year no-call period. A very low rate is usually charged due to the sale of the Bermudan options. For example, a recent rate was 2.71 percent. Nothing is free, of course. The reason for the low rate is that the dealer, after the two-year no-call period, will call the repo if rates are above 2.71 percent and keep it if rates are lower than 2.71 percent. The borrower bank “loses” in either case; that was the cost of the very-low rate, two-year funding.
These strategies nearly always involve shorting volatility. That is, because options are sold in the repo to obtain a low rate, and options are sold to obtain a high rate in the bond asset (typically agency-callables or mortgage-backed securities), the best outcome is for rates to move only a little. In that case, the large spread (which results from the option “sales”) is collected for the longest possible time. If, on the other hand, rates fall a lot, the bond prepays and the “cheap” liability is not so cheap anymore. If rates rise a lot, the bond extends but the repo gets called and the then underwater bond must be funded at a higher rate.
These deals have a place in a community bank’s overall funding strategy, but with the following reservations:
1. Unlike the FHLB, these repos are always collateralized with high-quality, liquid collateral. A liquidity issue could be presented if the borrower does not have enough collateral to collateralize the repo. This is particularly true in the margin call situation.
2. What exactly is the liquidity position in these call situations: Two years? Seven years? Somewhere in between?
3. All of these repos, in order to provide cheap funding, require short option positions in derivatives that are “embedded” in the repo. The accounting for these could change and losses marked through income if fair value accounting becomes the new accounting regime.
4. A FHLB advance has a somewhat different liquidity profile because the FHLB will take a larger collateral pool (i.e. commercial loans, albeit at a 50 percent haircut) and the FHLB probably has a greater loyalty to its members than a strictly arms-length deal with the street.
5. Any strategy that goes awry could subject the bank to excess interest rate risk (e.g. failure of the FISA test), liquidity risk and regulatory risk.
Any institution that deals in these transactions should thoroughly review its liquidity policy and have contingent plans for liquidity in stress situations (i.e., margin calls on collateral that may or may not be available). Limits are prudent also and should be a part of any liquidity policy review.
Clearly the most likely use of these transactions will be in leverage deals with bonds of some sort as the asset purchased with the proceeds of the repo. A thorough review and simulation should be undertaken for each contemplated transaction, both to analyze gains and losses and also the liquidity position in all situations. Any simulation should take into account extreme moves in rates — 300 basis points, for example — because of the risks outlined above in the case of large moves in rates up or down.
Always keep at least one eye on the accounting for the repo and the accompanying bonds. Small changes in the accounting rules can have a large impact in a fair value accounting world.
In short, these tools can be valuable but should be used with caution and with a full understanding of how they work in all situations.
Tom McKernan is in the Capital Markets Group of Country Club Bank, Kansas City.
Copyright © March 2008 BankNews Publications