The term structure of interest rates has often been used to draw inferences concerning the expectations of market participants. If a policy change is believed to be permanent and its effects on interest rates are understood, this should be reflected in the implied theoretical relationship between short rates, which are observable, and long rates, which are not. Changes in monetary policy regime may affect the typical durability or `persistence' of changes in short rates. In such cases, when the change in policy regime is believed to be durable, long rates should change in accordance with changes in short rates. If the long rate does not behave as predicted by the theory, this may indicate that the policy regime is not credible. For example, if the policy stance is credible, a temporary increase in short- term interest rates with the stated intention of reducing inflation may, actually reduce long rates, as occurred in the early 1980s in the UK. If it is not credible e.g. if the rise in short rates is viewed as an adjustment to a persistently more inflationary environment then long rates may respond by rising rather than falling (as in the UK in early 1990).

In Discussion Paper No. 430, Research Fellow John Driffill uses data on interest rates in the UK to study the credibility of changes in macroeconomic policy regime. In his model of the term structure, the yield on long bonds is a weighted average of the current yield on three-month treasury bills and expected future yields, such that the yield on the treasury bill equals the sum of the coupon yield and the expected capital gain on the long bond. If an increase in short-term interest rates is expected to be of short duration, the long rate should increase only a little, whereas if it is expected to be permanent then an equal increase in long rates is predicted.

Driffill uses UK quarterly data on treasury bill yields and long government bond yields for the period 1959-87. He finds that bill yields follow a first-order autoregressive process, with a structural break around the end of 1974, while long bond yields only behave in accordance with the model's predictions when a risk premium following a first-order autoregressive process is added to the theoretically predicted yield. Again, there is a structural break at the end of 1974.

He finds no evidence to support the view that the introduction of the MTFS in 1980 affected the persistence of changes in short rates, nor that there was any significant change in the response of long rates to short rates then or since. These data do not point clearly to any change in policy in 1980. There is, for example, no indication that the Conservatives' low inflation stance reduced expectations of long-term inflation and thereby reduced long-term interest rates relative to short rates.

The Term Structure of Interest Rates: Structural Stability and Macroeconomic Policy Changes in the UK
John Driffill