Krugman Vs Leijonhufvud Or Theory Vs Reality

Paul Krugman criticizes the post by Axel Leijonhufvud that I previously linked to. His argument is based on the expectations theory of long-terms yield in which long-term yields are identical to future average short-term interest rates. He illustrates this with the below graph:

This argument is partly correct. It is clearly the case that long-term interest rates are strongly influenced by expectations of future short-term interest rates. 

However, it is also the case that largely because of central bank intervention, actual short-term rates tend to be lower on a long-term basis than long-term rates. Only briefly during cyclical peaks are short-term rates higher, meaning that borrowing short and lending long is a long term winning strategy. Here is a chart comparing the 5-year Treasury security yield with the Fed funds rate during the latest 10 years. As you can see, the Fed funds rate was almost always below the 5-year yield. The same thing holds true if you compare it to the 10-year yield, except that the gap was usually even greater.

Furthermore, the greater the extent to which the Fed chooses to surprise on the downside compared to expectations, the greater will the subsidy be.