Modeling long-term nominal interest rates

The Pure Expectations Hypothesis has long served as the preeminent benchmark model of the relationship between the yields on bonds of different maturities. When coupled with rational expectations, however, most empirical renderings of the model fail miserably. This paper explores the possibility that failure to account for changes in market participants' assessment of the monetary policy regime, including changes in the target rate of inflation and the response to inflation and output, may explain much of the failure of the PEH. Estimating explicit expectations for changing monetary policy regimes in conjunction with the PEH model goes a long way toward rescuing the PEH model. The long rate implied by the PEH for a stationary short rate process tracks the observed long rate. The predicted spread between the long and the short rate is highly correlated with the actual spread. The standard deviation of the theoretical spread is nearly identical to that of the actual spread. Overall, these results cast suspicion on the use of spread regressions to test the PEH.