The US two-year yield has risen nearly 57 bp so far this year. The Federal Reserve has hiked rates twice and is poised to raise the Fed funds target rate in a few weeks. The 10-year yield has fallen 11 bp. The decline in the 10-year yield seems counter-intuitive and many worry that if it continues, it may signal a recession. To be sure, the 2-10-year curve is still positively sloped by 58 bp, but it is the flattest since late 2007.
The short-end of the curve is anchored to a large extent by Fed policy and perceptions of its trajectory. Long-term rates have three components: long-term expectations for real rates, inflation expectations and the term premium, which includes everything else.
To begin explaining the decline in the 10-year yield this year, it is helpful to consider those three components. Let's start with the long-term expectation for the real rate – what economists have dubbed 'r-star' (r*). It is something not observable, so economists have devised proxies. A useful proxy is long-term forecasts for Fed funds, which are then adjusted for the Fed's inflation target.
Consider that last December the median forecast of the long-term equilibrium rate for Fed funds was 3.0%. Assuming it reaches its inflation target (2%), that would put the real rate nearer 1%. In the latest forecast (September), the median Fed forecast now stands at 2.75%. Nearly half the decline in the 10-year yield this year could be explained by the lower, long-term real rate expectation.
The second component of long-term interest rates is the inflation expectation. There are two broadly different ways to document long-term inflation expectations: market-based measures and surveys. The market-based measures, like the breakeven rates ( the difference between the yield of inflation-linked securities and conventional securities, or five-year forward forwards). The problem with these is that they are not clean and the implied yield is also a function of the relative supply and demand of these instruments.
There are also numerous surveys including quarterly ones of professional forecasts that the Fed conducts. The University of Michigan conducts a monthly survey as well. Surveys suggest little change in inflation expectations. The latest Fed survey found a median expectation of average inflation from 2017-2026 of 2.25%. Last December, the 10-year forecast stood at 2.22%. The University of Michigan survey for the 5-10 year inflation expectations has also been stable. It has averaged 2.5% this year, unchanged from last year's average, though down from an average of 2.7% in 2015. On balance, a shift in long-term inflation expectations does not appear to account for the decline in long-term interest rates.
That brings us to the third component of long-term interest rates: The term premium. It is a residual category that includes everything else that influences long-term rates besides the real long-term rate and inflation expectations. There are several factors that may have led to the decline in the term premium. For example, there are more than $10 trillion in negative yielding bonds (in Europe and Japan). This may incentivize some investors to take a lower pickup on long-term US bonds (over short-term notes) than they may have if rates abroad were not negative.
A year ago, many expected fiscal stimuli and infrastructure spending. The US 10-year yield rose a little more than 60 bp in the November-December 2016 period. The market has scaled back such expectations. Also, geopolitical uncertainty may be another factor that may have weighed on the term premium.
When Bernanke proposed his answer to the Greenspan conundrum of the flattening of the yield curve in 2005, he cited foreign demand for US Treasuries coming from primarily from Asia and oil exporters. This was a savings glut that reflected the shift in many emerging Asian countries that had swung from current account deficits to surplus after the Asian financial crisis, and the inability of other countries, like China and oil producers from absorbing their own savings. The savings glut of at the time was understood to lower the term premium.
The conclusion of this review is that the decline in the US 10-year yield and the flattening of the yield curve are understandable. It is a function of 1) the Fed raising short-term interest rates, 2) the decline in the understanding of the long-term equilibrium real rate (r*), and 3) the decline in the term premium, which could be a function of international developments and scaled-back expectations of US fiscal stimulus.
Going forward, investors may be best served by monitoring there three components of long-term interest rates. Our work with surplus capital as a general condition warns that the long-term equilibrium interest rate could decline further. Inflation expectations might turn out to recent prints rather than long-term structural factors. This warns that if inflation does rise next year that long-term interest rates could respond dramatically. The term premium could also change depending on international developments and long-term US investor behavior (talk of pension funds locking in duration to better manage future liabilities while they can get a larger tax benefit that might be available next year).