For banks that have investment leverage or looking to embark on an investment leverage plan, take heed - it is highly probable that you will end up hurting your shareholders instead of helping them. With little to get banks excited about in the market, broker-dealers are pitching leverage plans like crazy, many with no real understanding of what they are doing.
Leverage, of course, works both ways. Banks can make money, but they can also lose money at a faster than average rate. The reality of leverage is that almost any time a bank borrows on a wholesale basis and reinvests in marketable securities, it does so with a negative total return profile. In other words, there is no free lunch. When a bank adds a layer of leverage it takes on 100% of the risk for about 92% of the expected return - due to capital charges, transaction costs and maintenance expenses. To overcome this financial friction, broker-dealers will cloak their leverage program pitches with a combination of credit, liquidity and optionality risk.
For example, without these risks, bankers can see something approaching the true cost of the program. Let's strip out risk and do a 5Y leverage program, taking an FHLB advance and reinvesting into 5Y Treasuries. The 5Y FHLB advance cost 1.73% this morning and the 5Y Treasury goes for a 1.12% for a 61bp annual loss. Well, that is stupid you say, as I can't make any money investing in Treasuries! That is correct, however that is our point, as this is an apples-to-apples comparison of risk. In fact, out of the 50+ different leverage programs we looked at this morning using a combination of municipals, bullet agencies, callable agencies, corporates and all sorts of mortgages; a 61bp annual risk-adjusted loss was actually our BEST return.
If you want to add some agency credit risk, you can borrow at 1.73% and reinvest it at 1.18%. That is only a 55bp nominal annual loss. Unfortunately, agency credit risk is on the rise again as the explicit Gov't guarantee runs out presumably at the end of 2012. This uncertainty, combined with headline risk, equates to a risk-adjusted 65bp of expected annual loss. Adding optionality just makes matters worse, as now we are not only taking on prepayment risk, but dramatically more liquidity risk. A leveraged portfolio of 5% mortgage pass-thrus comes out to a ballpark 101bps expected annual loss, while the average CMO leverage portfolio (composed of sequentials) comes out to 125bp and ARMs comes out to more than 206bp of expected loss. It should be noted that these expected loss rates have rarely looked better, as optionality has been dampened due to fewer banks doing refinancing and banks are once again getting paid better for taking on liquidity risk (but that premium is quickly falling).
Why do these leverage programs continue to occur if the math is so bad? The answer is that a leverage program can be made to look great on paper. The more smoke of credit, interest rate, optionality (watch those caps) and liquidity risk is thrown in, the harder it is to tell what is going on. A couple of static interest rate shocks on the leveraged portfolio doesn't cut it, as they only capture a mere fraction of the risk that the bank is taking on. Unfortunately, most broker-dealers don't even understand themselves how to value the cap on an agency ARM or how the structure of PAC CMOs truly work.
Here is a rule of thumb on leverage - it is recommended any time you can get below-market funding (say, thru your core deposit base) or can invest in an instrument where you know the return is better than how the market prices (like, maybe a loan to a borrower you know well). Other than that, stay away, because once you go down the path of leverage, you leave the realm of investing and cross over into speculation.
Leverage programs may work in the short-term, but over time, most will tend to lose money on a risk-adjusted basis. If you are convinced that rates are going to remain flat for some time, go ahead and put on a leverage program, but understand the risk you are taking. Should inflation catch fire in a couple of years and rates dramatically increase, the leverage program that looked so attractive in 2010 will act as a major albatross to earnings. Unless you have a crystal ball, stay away from leverage, the risk these days is too great.