Econbrowser

As shown in the graph on the right, the yield on 10-year Treasury bonds has declined 50 basis points since last March. Fed Chairman Alan Greenspan offered this analysis last month of the role that foreign purchases of U.S. bonds might have played in the decline in long-term yields:

The heavy accumulation of U.S. Treasury obligations by foreign monetary authorities is yet another hypothesis that has been offered. And, doubtless, those purchases have lowered long-term U.S. Treasury rates. But, given the depth of the market for long-term Treasury instruments, the Federal Reserve Board staff estimates that the effect of foreign official purchases has been modest. Furthermore, such purchases seem an implausible explanation of why yields on long-term non-U.S. sovereign debt instruments are so low.

Since the foreign purchases theory continues to have advocates (most recently, Stefan Karlsson at the von Mises Institute and some of the very well-informed readers commenting on my earlier post), I decided to take another look at this explanation.

Section 5 of the Treasury Bulletin records the volume of net foreign purchases of long-term U.S. Treasury bonds and notes. These went from $277 billion in 2003 to $352 billion in 2004, and $101 billion already in the first quarter of 2005. In a thin market with significant risk and private information, one certainly might expect a change in the volume of purchases to alter the yield relative to other assets. Of course, the Treasury market is a very dense, low-risk, public market, so it's quite possible that huge volume changes could be absorbed with little change in yield. Whether the volume of foreign purchases that we've seen could be sufficient to account for a significant part of the change in the spread between long- and short-term Treasuries would seem to be an empirical question.

If increased foreign purchase of long-term Treasuries is the explanation, I would have expected to see its effects also show up in a couple of other spreads. The bottom blue line on the graph on the left tracks the difference in yields between Aaa-rated corporate debt and 10-year Treasuries. This spread has barely inched up over the last four months, meaning that if one were to blame the lower yield on long-term Treasuries on foreign investors, one would have to hypothesize that they bought Aaa corporate debt in similar quantities. On the other hand, the magenta line shows that the Baa-Treasury spread has risen about 35 basis points since March. So if you said foreigners were buying long-term Treasuries and Aaa corporate but not Baa corporate, you might claim to have explained more than half of the decline in the Treasury yield over the last four months.

Source: Municipal Market Advisors

Another place one could look for evidence that the declining yield on long-term Treasuries might be attributed to foreign purchases is the spread between Treasuries and municipal bonds, since the latter yield tax benefits to U.S. residents but not foreigners. The maroon curve in the graph at the right confirms that 10-year Treasuries have indeed fallen about 20 basis points since March relative to munis of the same maturity.

A final spread that seems worth consulting is that between 10-year Treasuries and the 10-year Treasury Inflation-Protected Security maturing in 2013, which spread is given by the yellow line in the left-hand graph above. This spread, which corresponds to investors' expectations of inflation, has dropped nearly 50 basis points since March.

There thus seem to be plenty of attractive explanations for the decline in long-term Treasury yields over the last four months. One might be able to attribute up to half of the decline to foreign purchases, or try to pin the whole thing on decreasing inflation expectations.

So why aren't Greenspan and others convinced? The problem is that when you look at the behavior of each of the above graphs for the past year as a whole (July 2004 to July 2005) rather than just since March, everything goes in the wrong direction. In particular, compared to a year ago, the 10-year Treasury yield has still fallen by 50 basis points, but this is almost the same as the fall in the TIPS yield since July 2004, meaning you can't attribute any of the year-to-year decline in the nominal 10-year yield to lower inflation expectations. The Baa yield has fallen even more than Treasuries since a year ago, so if the decline in Treasuries is to be attributed to foreign buying, foreigners must have been purchasing even more of the corporate debt. Same story from the munis-- very little year-to-year decline in the spread between 10-year Treasuries and municipal bonds.

So if, on a year-to-year basis, neither the inflation terrier nor the foreign-purchasing poodle are yapping, what are we left with? It seems to me you then have to attribute much of the year-to-year decline in long-term yields to a decline in the term premium itself. And as my research with Dong Heon Kim has shown, this is exactly the sort of thing that can often be an early signal of an economic downturn.

Calculated Risk has noted that, once you subtract recent inflation, the current fed funds rate is near zero in real terms when the CPI is used to measure inflation, though it would be about 1% if you used core inflation. And Macroblog argues that, if the Fed does not raise rates much further it might be able to replicate the 1994-95 soft landing. Both observations have merit, but neither line of reasoning would give me much comfort if the Fed were to push rates up by more than another 50 basis points. I'm therefore disturbed that option prices suggest the market believes there's a 10% probability the Fed could go all the way to a 4.0% fed funds rate by October. Unless the long yield rises significantly before then, I would strongly caution against that. My recommendation continues to be that Mr. Greenspan should think of an inverted yield curve as a big red stop sign. If he just speeds past it, some dogs might really start howling.