Wednesday, 11 August 2010

Bank Debt Spiral

The zero interest rate policy (ZIRP) will kill the banks. Falling interest rates help banks by increasing the value of their bond and loan portfolios. This is the well understood inverse relationship between discount rate and present value of a future sum. But you see keeping interest rates at zero does virtually nothing for the banks as rates cannot fall further. There is a short window where ZIRP is a positive but an "extended period" (in Fedspeak) is just slow death for the banks.

During that short period, the banks are still collecting on portfolios constructed when rates were higher but as those higher-yielding assets mature, there is nothing comparable to replace them. We hear constantly how banks can just borrow at zero and invest in Treasuries - pocketing the difference. That would be fine if yields on Treasury debt were not low and falling along with everything else. The other problem is that this simplistic formula assumes that banks' operating expenses are negligible. Both unstated assumptions fail any sort of reality check.

Back in the real world, T-bills yield virtually nothing. The 2-year note is now at 50 basis points as of today. The 5-year is at 1.43% and the 10-year at 2.68%. Assuming zero borrowing cost (which is overly generous), net interest is equal to gross interest. Large banks generally require a net interest spread of more than 2% to cover their expenses, so they will lose money even buying 5-year Treasuries. If they invest their entire portfolio in 10-year notes, they'll make about a 50 basis point spread on assets pre-tax. But the 10 years is a lot of risk in terms of time for rates to change and also a long time to tie up the money. And banks care BARELY eke out a profit by taking this extreme level of maturity risk. There is a reason why you never see loan portfolios with 10 year average maturities.

For those advocates who think banks can rebuild their balance sheets by buying Treasuries, you might ultimately be correct but there are so many things that can go wrong with that scenario. First consider the size of the hole in bank balance sheets. Recent activity at the FDIC suggests that many troubled banks are overstating the value of their assets by 30% or more - that is the average size of the hit when the FDIC takes them over. At a rebuild rate of 50 basis points annually (with a lot of risk) it would take a literal lifetime to repair the balance sheets via this strategy. It was much easier in the early 1990s when rates for the 10-year started at 9% and never went below 5.5%. There was plenty of room to generate capital gains on bank bond portfolios wit falling rates and still leave a reasonable current yield at the end. Anybody using that era as a template for bank recovery is going to be sorely disappointed. Does anybody still wonder why Japan is trapped despite 20 years of ZIRP?

All of this assumes that ZIRP is sustainable over decades and that the financial system is sufficiently stable to endure the pressure over the long term. Neither one is proven and the ability to fund the debt implied by ZIRP is particularly shaky. If it works, it will take 60 years As one one of our favorite bloggers Karl Denninger says "the math is never wrong."