Working Paper: CEPR ID: DP430
Authors: John Driffill
Abstract: This paper examines data on interest rates in the United Kingdom information on changes in policy regime and their credibility in order to discover the period from 1959-87 using quarterly data. A stochastic regime switching model used by Hamilton, based on an AR(4) model for short rates, and the corresponding model for long rates, does not adequately represent the UK data. Yields on long-term UK government debt behave consistently with the expectations model of the term structure, on a number of basic tests. Their relationship with yields on treasury bills, however, is not consistent with the theory unless an autoregressive risk premium is introduced into the holding period yield on long bonds. The only evidence of a change in the time-series behaviour of long bond yields in these data occurs at the end of 1974. There is no evidence of a policy change in 1979 or 1980. The hypothesis that these interest rates contain unit roots cannot be rejected. Therefore, tests of the expectations model devised by Campbell and Shiller to take account of unit roots in the data were undertaken, but they revealed no evidence of departures from the expectations model.
Keywords: term structure; credibility; structural stability; monetary policy
JEL Codes: 023; 311
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| short-term interest rate changes (E43) | long-term interest rate changes (E43) |
| increase in short-term rates (expected to be short-lived) (E43) | slight increase in long rates (E43) |
| permanent increase in short-term rates (E43) | commensurate increase in long rates (E43) |
| autoregressive risk premium introduced (C22) | observed behavior of long rates reconciled with expectations model (E43) |
| no significant evidence of change in timeseries properties of interest rates (E43) | implications for monetary policy framework (E61) |