This past week the market traded counter to the reflationary trend that had been defined by QE positioning. The yield curve flattened, inflation breakeven spreads narrowed, the dollar rallied, oil and stocks both pulled back. The bond contract traded right back to the influential 150-00 pivot level I have been watching, which provided stiff resistance despite a slew of weaker than expected economic data. Since the Fed announced QE the volatility in the bond market has been intense, and I think this is indicative of how it will trade for months to come. Last week I discussed the collapse of implied volatility in the bond market. This week I will discuss the rising risk of actual volatility.
When I wrote Analyzing the QE Meltdown, Ex Ante
I had sent the link in an email to a few friends who are in the business of investing capital with the following comment:
Now the key going forward is the tradeoff between inflation, employment, and interest rates. The market has sounded the alarm and now wants to know this question. What level of inflation is Ben Bernanke willing to accept to bring down unemployment? Until that is defined, the bond market , or more specifically, the long end of the curve, is left to fend for itself. That should continue to foster a volatile environment that will be tricky to navigate.
When Bernanke left QE open-ended with no defined economic objective, he left it to the market to interpret what data trends would produce more or less easing going forward. In fact it’s possible that each tier-one data point could now be traded as if it was an unpredictable FOMC announcement. The initial reaction to open-ended QE has been a reduction in implied volatility, but an unintended consequence is that it may trigger higher actual volatility. There is no more important economic statistic than the monthly employment report, and perhaps the second most important release is the ISM manufacturing report. This week we will get both numbers which are the first since the Fed launched QE III. There are two important questions investors are facing that the market will be answering with each economic release.
What is the current bond market discount for how much QE is forthcoming?
How will the incoming data adjust this discount?
By launching an open-ended QE, Bernanke made a very big change in how monetary policy would be conducted from his previous regime. Bernanke has been a big advocate of opening up the communication from the Fed in an attempt to make monetary policy more effective. In doing this he has lulled the markets into being used to him telegraphing every move. However with QE III being open-ended, the market will be returning to an era where the future path of monetary policy had to be interpreted from each data release. If you could define Bernanke’s tenure, it would be in the certainty of monetary policy. But with open-ended QE, that era has ended.
For example, what is the level of non-farm payrolls that will elicit a reduction in the unlimited QE discount? What if private payrolls exceed 150k for successive months? What happens when jobless claims drop below 350k? What about ADP? Then there are employment leading indicators where you have numerous data points that could be extrapolated into employment trends. You have ISM manufacturing, housing and retail sales, and within these statistics you have the ISM inventory to sales ratio, construction permits, and consumer spending that will all be indicative of employment trends. The market will calibrate the QE discount on each of these releases, and just as the recovery has been volatile, so too will be the reaction to the data. For years Bernanke made economic data largely irrelevant, but with open-ended QE III, he now makes the data extremely significant. In addition to the volatility that each data point will bring, investors will also be faced with the conflicting implications open-ended QE will have on inflation discounts. Softer data will, in theory, bring about more QE, yet it will also increase inflation discounts in the market. Weak data would then be net/net bearish for the curve and thus long term duration while the opposite would be true for stronger data. I think you saw some of that in Thursday’s response to the weak data. The supply demand dynamics in the mortgage market are thus now subject to inflationary discounts in not only the slope of the yield curve but also the currency and commodity markets.
This interest rate volatility will only be magnified by the ever present reflexive feedback relationship between Treasuries and MBS. As I explained in Explaining Irrational Behavior: It’s a Reflexive Process:
Because the borrower has the right to prepay his mortgage at any time the mortgage holder is effectively short a call option. Because you are short this call, mortgage securities exhibit what is known as negative convexity. For this reason when you are long an MBS you are short interest rate volatility. Convexity and its derivative cousin gamma have exponential effects on market price and is a major contributor to extreme market movements.
To hedge against this negative convexity, mortgage holders will calibrate this changing duration risk with Treasuries (typically 10YRs). When yields fall they buy treasuries to add duration and sell when yields rise to reduce duration. You can see that in extreme situations (like record low Treasury yields) mortgage holders can exacerbate movements in the market as they increase demand when prices rise further pushing them higher and vice versa. They can get caught in a feedback loop so to speak.
This reflexive relationship is intensified by the Fed’s influence. Consider the fact that due to the amortizing of principal in MBS, the more the Fed buys, the lower rates fall and the faster mortgages will prepay. Therefore in order to maintain the size of the balance sheet, the Fed will have to buy more MBS. Lower interest rates or higher prices will create more demand from the Fed, which is the essence of reflexivity. It also cuts both ways. If rates start to rise, the MBS in the Fed’s portfolio will pay principal slower, which will reduce the need to buy more MBS to maintain the size of the balance sheet, which may drive rates higher. When Bernanke kicked off the QE II discount at Jackson Hole in August 2010 the ensuing reflation correlation whereby the dollar fell and both stocks and commodities rallied was the easiest trade in history as the market was able to ride this trend for a specified period of time. Bernanke in fact bragged
about the positive effects QE II was having on stock prices even though it was only a game of speculators. What he didn’t brag about was the massive imbedded short volatility position by virtue of the implied USD carry trade. Due to the negative impact that rising commodity prices was having on consumer spending, the Fed decided not to extend the program, and when they turned off the printing press, the easiest trade in history was no more and the reflation correlation they engineered blew up in their faces.
Now in an MBS market that is also short volatility, investors are facing this uncertainty at every turn. No one, not even Bernanke, knows how many bonds the Fed will buy or how long the program will last. But the market will be forced to make that determination upon each incremental data point.
The last element of increased volatility is the fact that the market is loaded with very low long duration coupons. When the Fed actually launched QE II in November 2010 the Street was locked and loaded to hit Bernanke’s bid and the 10YR yield rose dramatically as the curve steepened on the back of a rising inflation discount. The 10YR coupon during those months was in the 2.625% range. Today the 10YR coupon is 1.625% adding to the effective duration and making it more sensitive to a move in interest rates. If we get a comparable 100bps rise in the long end of the curve, you are talking about a massacre in market prices. So you have a market that is short volatility facing a potentially very volatile environment. There is the nature of the volatility in the incoming data and how the market calibrates the discount of the ultimate size of the QE program. You have the changing inflation implications in the data’s effect on the size of QE and that impact on the yield curve. Then there is the reflexive relationship of the negatively convex MBS market and the added inherently volatile low coupon long duration assets. Taking all of these ingredients together and Bernanke is cooking up a witches’ brew of market volatility.
With current coupon MBS yields currently trading on top of 10YR yields you have to believe there is a considerable amount of purchasing already priced into market prices. However I believe that how much this discount changes with incoming data trends will be the source of an increase in volatility risk in the bond market which could translate into increased volatility risk in all markets.
No positions in stocks mentioned.