Last month the Federal Reserve released minutes from the March 13 FOMC meeting.  These minutes suggest that the members perceived the economy to be improving.  The knee jerk reaction of market observers was to conclude that Fed easing policy was closer to being over than they thought before the release of the minutes.  Treasury yields rose sharply, by more than 10 basis points in the next couple hours.  Why did Treasury yields rise?  Was it really due to the diminished likelihood of more quantitative easing (QE)?

To answer this question, it might be useful to ask another one:  Why are long-term interest rates so low in the first place?  To most people the obvious answer is that the Federal Reserve is keeping rates low.  According to this view, the Fed controls the short end of the curve by setting the fed funds rate, and the Fed controls the long end through a commitment to keep short rates low for an extended period of time and through purchases of long-term Treasury and mortgage bonds (QE).  An alternative answer is that rates are low because inflation is low and economic growth is meager.

According to the consensus view, when the Fed stops QE, long-term yields must rise.  It may well be true that termination of QE and rising long-term yields will coincide, but I think the causality story is wrong.  Rates fell in 2008 because nominal GDP growth fell off a cliff.  The Fed reacted to this by lowering the funds rate to zero and then implementing QE.  Had the Fed not eased monetary policy, the collapse in the economy would have been greater, and long yields would have fallen even more than they did.

Why does it matter which story is you believe?  The implication of the consensus story is that the Fed controls the level of long-term rates and can keep them low indefinitely by purchasing large quantities of Treasury bonds or Mortgage Backed Securities.  I think that is not correct.  Continued QE may well be a good thing, but not because it keeps long-term yields low.

Long-term rates will rise when growth improves and inflation rises.  The combination of inflation and real growth is growth in nominal GDP (NGDP).  NGDP has grown over the past year at a 3.75% pace and over the past five years at a 2.4% pace.  Historically, the level of long-term Treasury yields is strongly correlated with the rate of NGDP growth.  In fact, a smoothed measure of growth in NGDP provides an excellent tool for projecting the central tendency of Treasury bond yields.  I find that a regression of the ten year Treasury yield on a five year average NGDP growth yields a slope coefficient essentially equal to 1.0, an intercept insignificantly different from zero and an R2 of .62.  This equation does not capture all the volatility of interest rate changes, but it does provide a good rule for defining “fair value” for the ten year yield. 

                        Fair value yield = 1.0 times 5 year NGDP growth

Today, that model predicts fair value for the ten year yield of 2.4% compared to the actual level of 2.0%.

Suppose the Fed targeted the rate of NGDP growth at, say five percent and was prepared to increase its balance sheet until that objective was reached.  Eventually the Fed would succeed in reaching this objective, and bond yields would move sharply higher.  Thus, easy Fed policy ultimately leads to higher rates, not lower rates.