Note prepared by the Specialist Adviser on uncovered interest rate parity (UIP)

The interest parity condition lies at the heart of most explanations of how interest rates affect the exchange rate. There is, however, much misunderstanding of how it operates. It is commonly used to predict how the exchange rate will respond to a change in interest rates, but considerable care needs to be taken to get this right. The use that appears to be made of UIP by the Monetary Policy Committee as reflected, for example, in the Minutes of May and June 1999, illustrates this.

The UIP condition asserts that domestic interest rates will equal foreign interest rates once their return has been converted to domestic currency, in other words, adjusted for the expected rate of change of the bi-lateral exchange rate over the period of the investment. This assumes that investors are risk neutral and there are no capital controls. Risk aversion will add a risk premium to this relation. UIP implies that the expected rate of depreciation of the exchange rate (measured as the domestic price of foreign exchange—ie pounds per dollar or euro) equals the interest differential between the domestic and the foreign rate measured as domestic minus foreign. Thus if the one year sterling interest rate exceeds the euro rate by one percentage point, then sterling is expected to depreciate over the next year by one percent unless interest rates change. In other words, it predicts how the exchange rate is expected to change in the future if there is no change in interest rates

The UIP condition does not say, as is sometimes thought, that a one percentage point increase in the sterling rate will cause a one percent depreciation of the exchange rate. In fact, in response to the higher return on sterling assets there would be a capital inflow and sterling would appreciate, not depreciate. Moreover, this would happen instantly. The explanation for this can also be found in the UIP condition.

A logical consequence of UIP is that the current exchange rate can be expressed as the sum of all expected future interest rate differentials, but measured now as foreign minus domestic. Thus an increase in the current domestic interest rate (or an expected increase in future rates) would cause sterling to appreciate. The key to reconciling this apparent conflict in the interpretation of UIP is that the increase in the interest rate causes both the current and the expected future exchange rate to appreciate, but the current appreciates more than the future causing an expected future depreciation. This is the so-called over-shooting implication of UIP in which an increase in the domestic interest rate causes an appreciation of the current exchange rate, and an expected depreciation in the future.

To illustrate, since its inception in January 1999 the sterling short interest rate has exceeded the euro interest rate implying an expected future depreciation of sterling against the euro, yet sterling has appreciated over this period. One possible explanation is that a widening of the interest differential has also made sterling assets more attractive.

The Minutes of the Monetary Policy Committee of May 1999 (para. 25) state that "The central projection assumed that sterling's effective exchange rate would depreciate in line with uncovered interest parity (i.e. interest differentials)." This would only be correct, however, if interest rates were held constant. This type of argument could not be used if the MPC changed the interest rate. The Minutes of June 1999 (para. 32) suggest an attempt to retrieve the situation in the statement "Some members of the Committee were uneasy, on empirical grounds, about the use of the Uncovered Interest Parity assumption in the central projection, and for them sterling's persistent strength was not a surprise." This view is consistent with the long-lasting interest differential between United Kingdom and euro interest rates. It would then, however, be at odds with the argument that the recent strength of sterling was not due to the MPC's interest rate policies.

It is a well-established fact that exchange rates are very difficult to predict. This includes the precise response of the exchange rate to a change in interest rates. It is clear from the second implication of the UIP condition why this is. If the exchange rate is the sum of current and all future interest differentials then the weight given to the current interest differential is likely to be quite small, and information (or rumours) likely to affect future interest differentials could be far more important. Arguably, uncertainty about future interest rates is the main reason for the large error in forecasting the exchange rate, and why a policy of exchange rate targeting is usually ineffective.

It follows that a temporary change in the interest differential may have little effect. To have a significant impact on the exchange rate, the current interest differential would need to be thought a good predictor of future differentials. In other words, the change in the current differential would need to be thought long-lasting. For example, if the interest rate change were expected to last a year, then the impact effect on the exchange rate, and hence inflation, in the first quarter would be four times larger than if it were thought to last only one quarter. And if the increase were sustained for the whole year, so that market expectations were correct, the cumulative effect after one year would be ten times larger. Even if the interest rate change were reversed after one quarter, the larger impact effect would remain in the system.