Global: The Term-Premium Case for Higher Yields
Richard Berner (New York) and David Miles (New York)
The US yield curve has continued to move towards inversion along its length, with short-term rates this week temporarily moving above 10-year yields for the first time since just before the recession of 2001, and the inversion of the Eurodollar curve between June 2006 and June 2007 grew to 20 bp. But this inversion is different: Nominal and real yields are lower now than when the Fed started tightening in mid-2004, not just in the United States, but globally. If the bond market isn’t signaling a recession or at least much weaker growth ahead, then what are the sources of this cyclical ‘conundrum,’ and what are its implications for growth and monetary policy?
In our view, a decline in the “term premium” to near zero has been the major factor holding yields down over the past 20 months. We define the term premium as the excess of the yield on a longer dated (say, 10-year) note over the weighted average of expected short-term rates. On average, it is the extra return an investor earns by holding a ten-year note instead of rolling over a sequence of short-term deposits to the same maturity. In other words, the term premium compensates investors for the risks in holding the longer-duration security. The sharp decline in distant-horizon forward rates (such as 9-year forward 1-year rates) relative to 10-year yields suggests that term premiums have declined.
In the US, comparing surveys of expected future short-term rates with long-term rates points in the same direction. For example, Brian Sack of Macroeconomic Advisers points out that longer-term expectations of the path of monetary policy have been relatively stable over the past few years, while 10-year yields have declined steadily. Likewise, a variety of empirical models (augmented by survey data) that extract the change in the slope, level, and curvature of the yield curve suggest that it is a plunge in term premiums, and not a radically lower path for future monetary policy, that has depressed yields (see, for example, Don Kim and Jonathan Wright, "An Arbitrage-Free Three-Factor Term Structure Model and the Recent Behavior of Long-Term Yields and Distant-Horizon Forward Rates," FEDS Paper 2005-33).
None of this explains why term premiums have declined. In our view, that decline reflects the Fed’s and other central banks’ unusually forward-looking guidance about monetary policy. Other factors — such as swings in saving-investment balances, increased faith in central banks’ anti-inflation credibility and thus reduced inflation risks, the ‘Great Moderation’ implying reduced economic volatility, and incipient demand for long-duration debt by pension funds seeking to reduce asset-liability mismatches — may also have contributed to the decline in yields over a longer time frame. But their recent, cyclical role has in, our view, been limited; instead, these are important secular forces, and none has changed much in the past two years. We do expect that changes in pension accounting and funding rules will escalate future demand to satisfy pensions’ duration-matching needs, but so far, the increase has been small and slow in coming.
The idea that a recent surge in world savings has driven term premiums lower is, for example, not easily reconciled with the latest IMF data on estimated global savings rates. The latest IMF estimates of the global saving-GDP ratio show the recent levels now slightly below the average since 1980, and they have barely increased in recent years.
Of course, it is possible that savings have nonetheless risen relative to the desire to invest globally. But that seems more like an ex-post rationalization of falling yields rather than something for which there is independent evidence. Indeed, the savings glut story is consistent with almost any movement in actual savings and in that sense something of an empty theory. If world savings/investment is higher, the advocates can say it is because the supply of savings schedule has gone up relative to the demand for investment and driven real rates down. If world savings/investment goes down, they can say it is because the investment demand curve has fallen relative to supply of savings, and this is why real rates go down. In other words, whatever happens to actual levels of investment and saving is consistent with some version of the story.
What needs to be done to make the savings glut story really persuasive is to convincingly identify the world supply of saving and separately identify world demand for capital schedules and show that supply has moved up relative to demand. But our judgment is that there is no clear evidence to support that.
Figuring out what has driven term premia and yields down is not just an academic exercise. At issue is how much of the term premium compression is permanent and how much is temporary. By promoting price stability, by damping macroeconomic volatility, and by increasing what officials call the “transparency” of their communication to the public, the Fed has contributed to the permanent decline in the term premium, we think (see “Risk Premiums and the Fed,” Global Economic Forum, July 25, 2005). More scientifically, our rate strategy team confirms quantitatively and convincingly that Fed words and tactics in the current tightening cycle have been the primary factor behind an additional decline in the premium that may or may not last (see Graig Fantuzzi and Nilsson Kocher, “Resolving the Conundrum,” November 14, 2005). By splitting a regression of the yield curve on the real Federal funds rate into two distinct sub-periods —1990-2003 and 2004 to the present — they show two very different patterns; using the results of the earlier period suggests that the curve from 2-year to 10-year yields would be substantially steeper. The Fed’s language clearly telegraphed the pace of monetary tightening, and with ambiguity over pace gone, investors could profitably reach for yield.
Further parsing the term premium into nominal and real components, as have Kim and Wright, suggests that two-thirds of the decline in the term premium over the period since June 2004 has been real and only 30 basis points has reflected a decline in inflation premiums. We’d be the first to admit that this evidence is highly tentative, as have the authors. But pricing in the TIPS market approximately backs up that conclusion, unless one concludes that the market is seriously mispriced; ten-year break-even inflation is running about 230-250 bp, while US real yields stand at 2% and in the UK they are at 1%. If inflation risks have indeed dropped significantly, the cost of inflation insurance should be lower and TIPS yields should be significantly higher than they are. In fact, yields on inflation-proof bonds have fallen by roughly as much as, if not more than, yields on conventional bonds in most of the major markets.
In both the US and the UK, nominal bond yields on long dated conventional government bonds have fallen fairly substantially over the past three years or so. In the US, long dated Treasuries now yield about 4.5%. The average over 2002 was a bit under 5.5%. In the UK, long dated conventional bonds yield a little under 4% now and were at around 4.75% in 2002 (on average). In both countries, the fall in (real) yields on long dated inflation-proof bonds over this period has been greater than the fall in long dated conventional yields.
So the gap between nominal and real yields over this period -— which is some combination of expected inflation plus an inflation-risk premium — has in both countries increased. This makes the story that the fall in long yields is largely driven by much lower expected inflation and inflation volatility a bit hard to swallow. Another way to put this is to say that the data suggest that the fall in nominal yields is roughly accounted for by a fall in real yields, leaving limited room for the inflation factors to play a role. Falling real yields on inflation-proof French debt tell the same story.
Yields on long dated TIPS and French OATiS have come down from about 3.5% in 2001 and the first part of 2002 to about 2.5% in 2004 and to about 2% now (and a bit below that for OATiS). That's a fall of around 1.5% over the period since early 2002. Real yields on long-dated index-linked gilts over same period also fell by about 1.5% (from about 2.5 to under 1%). So the decline in real yields is comparable across these markets and in all cases is more than the fall in yields on conventional (nominal) bonds.
For market participants, this seemingly theoretical debate about the sources of lower yields is actually of critical importance. Those temporary factors compressing term premiums may be ending. Now that monetary policy is completely dependent on the inflation and growth outlook, and thus is inherently less certain, term premiums and long-term yields are likely to rise. Indeed, there is an ironic twist in the yield curve debate. Many fear that the flattening in the US yield curve is an omen of much slower US and thus global growth or even recession. They point to the tendency of the Fed to overdo tightening, the yield curve’s business-cycle forecasting prowess, the asset-driven nature of the US consumer, and evidence of a housing slowdown (see “Debating the Yield Curve,” Global Economic Forum, November 23, 2005).
But if a declining term premium — rather than the weighted sum of expected future changes in short-term interest rates — has recently contributed both to lower long-term yields and to a flatter yield curve, both imply faster, not slower, future growth, exactly the opposite of the traditional interpretation. And if the term premium has declined permanently for whatever reason, making the current structure of interest rates more stimulative than otherwise, then the Fed will have to raise short-term rates by more than otherwise to achieve its goals (see “Yield Curve Angst,” Investment Perspectives, November 23, 2005).
Despite our confidence about the validity of each of these assertions, recent experience of being wrong about yields has made us humble about the risks to this call. Among them: Secular disinflationary forces may overwhelm the cyclical factors boosting inflation. Term premiums may stay low for some time, reflecting the Fed’s measured approach.
But investors would do well to be mindful of risks in the other direction. A weaker dollar and the elimination of slack in the economy could combine to push inflation up more rapidly. Term premiums could unwind as rapidly as they have declined. And for market participants, three years of crying wolf about yields has bred market complacency about upside risks to interest rates. |
