At the Atlantic Economy Summit in Washington last month, Sheila Bair fielded a question about the just-released results of the latest bank stress tests. The former FDIC chief took pains to point out that they were an improvement over earlier iterations by virtue of keying off a truly dire economic scenario, but then ticked off a number of ways in which they fell short. One was in that they focused solely on credit risk, when historically, adverse interest rate moves have proven very effective in decimating the banking sector. Witness phase one of the savings and loan crisis, in which hasty deregulation and gimmickry in the early 1980s set up the crisis later in the decade, or the derivatives wipeout of 1994, in which an unexpected 25 basis point Fed funds increase created bigger losses than the 1987 crash, or the losses on US bond portfolios in 1997 and 1998, which among other things nearly wiped out Lehman.
The perils of interest rate risk have largely receded from memory since the US has been in a long-term disinflationary trend since 1983. But with rates at zero, we have nowhere to go but up from here.
Chris Whalen, in his latest newsletter, argues that this risk is even nastier than it might appear. One way of mitigating interest rate risk is by holding shorter-dated instruments. The reason is that the more back-weighted your payments are, the more exposure you have to changes in interest rates.
Investors measure this sensitivity via duration. There are somewhat different ways to measure it. One is the weighted average of payments. That gives you a result than under most circumstances is very close to the price sensitivity of a bond to interest rate changes. At “normal” interest rates, a current coupon bond of just under 10 years will have a 10% sensitivity to interest rate changes.
But, as Whalen points out, this all goes haywire when interest rates are super low. So much of the value is in the repayment of principal and so little in the intervening interest payments that it pushes duration out and increases interest rate risk disproportionately. That effect may be further compounded by the fact that banks are desperate for yield and with the yield curve flatish, they are likely to be extending the maturity of their assets.’And of course, anyone holding mortgages will see them extend, since refis will halt and home sales will probably be depressed, since higher interest rates mean more expensive mortgages, which all other things being equal, means lower home prices.
Whalen sets forth his concerns:
For a number of years now, US banks have been loading up on low yielding paper that is a function of the Fed’s efforts to reflate the US economy. Since many Americans are not able to refinance their home mortgage, the Fed’s efforts are not particularly effective, but we are not supposed to talk about housing. Banks and corporations have been able to refinance their debts, lowering interest expenses but also increasing the volatility – that is, the duration — of bonds and related swaps and options.
The trouble with low interest rates is that as coupons fall, the duration on a given bond lengthens exponentially. Whereas the duration on a Fannie Mae 5 or 6 can be measured and managed using traditional interest rate risk management tools, in a low or zero rate environments the effective duration on low or no coupon securities becomes so long and so difficult to manage that hedging becomes problematic.
Keep in mind that most banks today are seeking to maximize net interest margins on an interest rate book where the effective yield is falling. There is little incentive to lend cash or securities to other banks given that rates are zero, so banks simply place their excess cash into government and agency debt that has little cash flow yield and essentially infinite duration risk – risk that cannot be hedged.
“Of course, the banks also own a lot of duration…1.35 trillion of GSE MBS, notes one of the members of our mortgage finance discussion thread. “The Chinese had to sell to someone.” But such levity aside, the fact is that most banks are not even trying to hedge their interest rate books because cash flows on earning assets are so paltry. Thus as and when interest rates do rise, many larger banks that fund themselves in the markets will come under immediate pressure.
Now banks may be able to cover some of this up for a while. They may be able to put these low-yielding assets in a hold to maturity book (as they did in when similarly caught in the 1990s) which will spare them taking mark to market losses. But they’ll still lose money on an ongoing basis if they have assets that yield 3% that they are now funding at 5%.
Tom Adams added another cheery thought via e-mail:
There is also a coming problem for HELOC borrowers – these loans will reset from interest only payments to principal and interest payments in the not too distant future. The HELOC interest only period was typically around 10 years, so that would put the reset dates in 2-3 years for a big chunk of bank portfolio HELOCs. Wells Fargo, in particular, also did a huge amount of convertible HELOCs – in a rising environment, borrowers could convert to fixed rate HELOCs, at the current market rate. This would have the effect of making the par valued variable rate loans convert to below par fixed rate loans for Wells, if rates rise significantly. If it happens at the same time that many of the same loans are also experiencing payment shock due to a conversion to P&I payments, that could get really ugly for Wells with big step ups in both duration and credit risk. Something to look forward to!
Indeed. Of course, if we’ve zombified our economy as well as Japan has, this day of reckoning may be very long in the making. And given the consequences to banks of leaving ZIRP, perennial zombification may be a feature rather than a bug.
This post originally appeared at naked capitalism and is posted with permission.