Wednesday, November 24, 2010

The negative convexity of MBS is kicking in

Yields on 10-yr Treasuries are up 50 bps from their October lows, which were the lowest yields since the panic of late 2008. We can safely assume that yields were driven to those extremely depressed levels by the prospect of another round of quantitative easing—the market was front-running the Fed's plan to purchase Treasuries. Since then the Fed has confirmed its intention to purchase up to $600 billion of Treasuries, but yields have nevertheless jumped by half a percentage point. That's the market's way of flashing a big thumbs down to QE2.

The bond market last summer was worried, as was the Fed, that the economy was in trouble and needed more help. But now things have changed, and the bond market has figured out that the economy is not in desperate need of more money. The economy is not on the verge of a double-dip recession—it is growing and it's likely that growth in the current quarter will be stronger than the third quarter. Moreover, dumping more money into the economy can only increase the risk of higher inflation, at a time when virtually all commodities have been rising for almost two years, gold prices have been in a decade-long rally, and the dollar is at or near its weakest levels ever. A strengthening economy and the prospect of higher inflation are the bond market's worst nightmare, especially when yields are trading at generational low levels. As I said before, no amount of QE2 can keep Treasury note and bond yields from rising when the fundamentals are pointing to higher yields.

We now see that the bond market vigilantes are beginning to marshall additional forces. As a result, yields on current-coupon FMNA securities have jumped 67 bps since their October lows. Plus, the spread between MBS and 10-yr Treasuries has widened about by 40 bps since hitting a record low late last July. MBS spreads appear to be on their way to 120 bps (if not eventually much higher), which is the average level of the spread over the past 20 years. Maybe the Fed will be able to keep Treasury yields marginally depressed (i.e., distorted), but they can't keep MBS yields down forever. And maybe it's not just the bond market vigilantes at work here, maybe it's also the case that demand for mortgages is rising, as I pointed out in my earlier post today, which noted that new applications for mortgages are up 25% since last July. This is really turning into the bond market's worst nightmare, and we're likely still in the early stages of a big selloff.

Mortgage-backed securities have a unique characteristic which can fuel near-panic conditions when the economic and financial fundamentals change unexpectedly. It's called negative convexity: the effective duration (interest rate sensitivity) of MBS increases when yields rise, and decreases when yields fall. Back in July, mortgages had a duration of just under 1 year, because the market expected to see massive prepayments of mortgages due to the big decline in yields up to then. Mortgages now have a duration of a little over 3 years, which means they are three times more risky to own if yields continue to rise. When yields fall, rising prepayments effectively turn mortgages into very short-maturity bonds, but when yields rise, then declining prepayments can turn mortgages into very long-maturity bonds. The latter is what's happening now, and it's causing large institutional investors who are holding sizeable positions in MBS to scramble to hedge their increasing exposure to rising interest rates. The main way they hedge (i.e., shed unwanted duration) is to sell Treasuries. This augments the selling pressure that has already served to drive Treasury yields up by 50 bps. Higher yields will dictate even more selling of Treasuries to avoid bigger losses on MBS holdings.

The Fed is powerless to stop this higher-rate locomotive, especially since it's being driven by improving economic fundamentals and the perception that inflation risk is rising.

One way for this drama to end is for the Fed to call off QE2, and I suspect that will happen before too long. That will still leave in place the forces (e.g., an improving economy and rising inflation risk) that are driving yields higher, but higher yields on Treasuries are not going to derail the recovery. Given how low yields still are, rising yields are symptomatic of a stronger economy, not a threat to the economy.

Full disclosure: I am long TBT and long a 30-yr fixed rate mortgage at the time of this writing.


Benjamin Cole said...

Please G-d, please let Scott Grannis be right this time, and please let several solid years of inflation and growth begin.

Grannis has been promising inflation for two years, and instead all we get is deflation. Prices are lower now than in 2008, on the CPI. I fear we will see even lower prices in the years ahead.

David Beckworth has an excellent post today on global commodities prices being more tied to emerging market growth than US monetary policy.

Milton Friedman says QE is the way to go when you hit zero bound and you still have deflation.

Can someone channel MF and have him appear before a convention of "conservative" economists?

I fear my brethren are wayward and lost.

Calling MF, where are you? said...


Would/have you buy/bought longest dated out of the money calls? Or just straight purchase of TBT?

Would you similarly use PST?

Scott Grannis said...

Being retired, I have determined that I should avoid speculation via the purchase of option contracts. TBT is a way to bet on the direction of long-term interest rates that are largely beyond the control of the Fed. PST is a way to bet on medium-term interest rates that are heavily influenced by what the Fed does.

Steve Fulton said...

Interestingly enough the move in the 10yr since the announcement of QE2 has resulted in the MBS market adding just over $500 billion in 10yr equivalents of duration to the bond Market via negative convexity. Pretty much wiping out the $600B the Fed was going to spend. The market has a way of seeing through dopey ideas. said...


Yep. A case of 'buy on the rumor, sell on the news'.

Scott Grannis said...

Steve: Thanks for that very interesting little tidbit. It is amazing how the market works.

Charles said...

Nobody knows what QE2 will accomplish. Nobody knows how much they will do, how fast they will do it and when they will stop. Nobody knows what foreign governments will do in response. For that matter the markets do not know what the markets will do. Long term debt has become risky. So interest rates are up and this is one more hurdle for the economy.

The evidence suggests that the economy is accelerating to the point that the broadest measures of unemployment will stabilize. We have a little way to go before core inflation accelerates but there is no question that this administration's policies have raised the natural rate of unemployment significantly, both by reducing marginal incentives and through the erosion of human capital.