RDP 2000-07: The Effect of Uncertainty on Monetary Policy: How Good are the Brakes? 6. Conclusion

Policy changes in practice tend to produce a smooth path for interest rates while the path of policy interest rates generated by models or policy rules is often considerably more variable. This paper has investigated whether the inclusion of uncertainty can help reconcile the theory to the practice. It has shown that, in general, parameter uncertainty does not induce much smoothness when its effects are directly incorporated in the model. However, particular forms of parameter uncertainty may have some impact.

The main finding of the paper is that mean parameter uncertainty about the interest sensitivity of output can reduce the aggressiveness of optimal policy in the model. Thus, if it is the case that the effectiveness of monetary policy is greater than that suggested by the estimated model and the policy-maker knows that, policy is likely to be less aggressive. In particular, the policy-makers' beliefs about the effect of their actions on the economy are important determinants of the size of policy moves. However, even allowing for these forms of uncertainty, the path of policy interest rates generated is still considerably more variable than that observed in practice. This suggests that the main explanation for the smooth path of interest rates observed in practice lies elsewhere.

An issue that this paper has not addressed is whether there are any losses from a smooth path of interest rates. Lowe and Ellis (1997) tentatively conclude that smoother policy does not generate much increase in the volatility of inflation and output. Thus, even if the policy approach that has been adopted in most industrial countries has not been completely optimal because it has been excessively smooth, the costs of that have not been great. Moreover, there may be costs to increased variability in interest rates that are not captured by the model, which would further reinforce that conclusion.

Finally, the results in this paper are unable to explain the relative infrequency of reversals in the direction of policy, as opposed to continuations, that is observed in practice. The forms of uncertainty discussed in this paper are unlikely to provide an explanation. A more likely explanation might involve the potential adverse effects on the credibility of the central bank of frequent reversals in the direction of policy.