The Term Premium: the Economics Version of the Flux Capacitor

This is the economy in recovery: US 3Q GDP printed this morning at 3.9%, up from the advance estimate of 3.5%, and well above the consensus forecast of 3.3%. Although some portion of 2Q GDP of 4.6% was clearly due to a weather-related bounce back from the -2.1% of 1Q, it now seems obvious to perhaps all but the most stubborn bears that the US economy has been on a strengthening economic trajectory since at least the middle of 2013. 

The so-called conundrum? Long-term US Treasury yields continue to decline. In 2013 as the economy began to show signs of persistent strength and the Fed indicated it was finally preparing to end QE, 10y Treasuries sold off by ~120 bp, yields touched just over 3.00% in January 2014, and the yield curve steepened. Instead of long rates "following through" on that momentum as the economic recovery gained steam - as most* financial analysts expected (myself included) at the beginning of 2014 - long rates have reversed course and steadily rallied throughout the year, and the yield curve has flattened. Why? for the love of corn muffins, why?

*Jim Vogel and Chris Low of FTN Financial were one of the few, if not only, high profile strategists to predict that the 10y Treasury yield would fall substantially in 2014. Huge kudos to them for getting the direction right, particularly against an intensely tight consensus the other way. I don't have access to their research so I don't know what reasons they gave for the call, but unless they were projecting that rates would fall due to a renewed recession and QE4, that's just darn impressive. 

The level of interest rates, the shape of the yield curve and the business cycle. As all students of finance and economics know, interest rates tend to rise and the yield curve tends to flatten as the economy improves, inflation rises and the business cycle peaks. Conversely, interest rates tend to fall and the curve steepens again as economic strength declines and the business cycle heads to a trough. The primary driver of the flattening and steepening has traditionally been the Fed's conduction of monetary policy: with the Fed changing the level of short-term interest rates while long-term rates adjust to economic and market conditions. In normal business cycles, the Fed raises short-term interest rates - to prevent the economy overheating and inflation from rising too fast - through the peak of the cycle, which typically results in a flattening of the yield curve. Eventually the Fed lowers short-term rates - which effectively steepens the yield curve - to ease credit conditions as the economy weakens, in order to prevent or numb a recession. 

Short-term rates are pricing in a tightening cycle... The Fed has yet to raise short-term interest rates, but the front end of the yield curve has modestly sold off in anticipation. Interest rate futures contracts also show that the market is pricing in eventual rate hikes, beginning roughly in the middle of 2015. 

...Amidst a secular decline in long-term interest rates. For more than 20 years, these cyclical variations in interest rates and re-shapings of the yield curve have been taking place within the context of a secular decline in long-term yields. From What Caused the Decline in Long-term Yields? by San Francisco Fed economists Bauer and Rudebusch, July 2013 (edited for brevity):

During the past year, yields on long-term U.S. Treasury securities have fallen to historically low levels. ... These low yields partly reflect cyclical factors, including the slow pace of economic recovery, modest inflation rates, and accommodative monetary policy. However, in addition to cyclical variation, yields have registered a longer-run decline. Indeed, from 1990 to 2012, 10-year U.S. Treasury yields fell fairly steadily from just over 8% to around 2%.

Identifying the sources of this long-run decline in interest rates is of great interest to monetary policymakers, bond investors, and other financial market participants. But it is quite difficult to do. 

Many economists have attributed much, or in some cases most, of the secular decline in long term rates to a collapse in the term premium. Also, one way that the Fed purported to measure the effectiveness of QE was by estimating how much of the cumulative drop in yields over the various purchase programs was due to compression of the term premium.

The term what? I hear you. I will try to make this as simple as possible, but honestly it's such a convoluted concept - and subject to almost comically wide variation in estimation methods - that at times it's difficult for me to take it seriously (shhh...don't tell any economists). 

Defining the term premium. Let's return to Bauer and Rudebusch (edited slightly for clarity, formatting added):

Long-term interest rates can be separated into two components: expectations of average future short-term interest rates and a term or risk premium that investors require for bearing the risk of a long-term bond investment.

• The expectations component is driven by inflation expectations and expectations of future real rates of return, which depend on future economic growth.

• The risk or term premium component is determined by the amount of uncertainty about these future developments and by the degree to which investors are risk- averse and require compensation for a given amount of variability.

These two components of interest rates are not readily observed and must be inferred indirectly. The usual method involves the estimation of a dynamic term structure model, a statistical model that describes the interaction of short-term and long-term interest rates and their evolution over time. With estimates from such a model, a researcher has the necessary tools to identify the expectations and risk premium components of interest rates.

Editor's aside: May I just point out that there is no mention of supply and demand anywhere when estimating long term rates? So if, say, China were to enter the market and buy a few hundred billion of Treasury securities in order to deflate their currency, and that were to have a meaningful impact on long term rates, these models would instead be trying to account for that as perhaps a change in either inflation expectations or a change in risk aversion? I strongly suspect this is a limit in my own understanding, not necessarily in the definitions or models.

Economics is not physics. There is a body of very sophisticated, statistical finance literature that is dedicated to estimating changes in the term premia of interest rates, most often the 10-year Treasury rate. This research is done by people who are exceptionally serious, talented and dedicated. The problem is that there are no fundamental laws of economics. 

There are at least five different classes of econometric models that have been used to produce term premium estimates, with at least four producing estimates going back to 1961. There is a wonderful chart of these term premium estimates over time, and more detailed explanation of the various models, available on Simon Taylor's blog  here. 

He provides an outstanding high level overview and I strongly recommend you read his work. The synopsis is that by 2007, four of these models estimated widely divergent term premium  from -2.00% to +2.00% when the yield on the 10yT was ~5.00% . 

The Fed - either in its infinite wisdom or infinite hedging - chooses to use a term premium model that produces estimates roughly in the middle of the range of the other four. To my knowledge - and I could be wrong - modeling and estimating the term premium has not become less controversial or more robust over time, and I don't know that the variation in the estimates has meaningfully decreased. In my opinion that tends to undermine the practicality of using the term premium to explain a significant fluctuation in long term interest rates, or assigning any importance to a projection made based on a change in the estimate of the term premium.