Do errors in forecasting inflation lead to errors in forecasting interest rates?

In the first of three related, and consecutive, papers we showed that forecasts for short-term policy interest rates in NZ and UK deteriorated over the first six months to a point when they became useless, after the first two quarters. Moreover they were ex post biased, underestimating future interest rates during upturns and the reverse during downturns. Both NZ and UK have been inflation targeters during our data period. In this second paper we ask, first whether inflation forecasts exhibit the same syndrome as the related interest rate forecasts, and whether errors in the inflation forecast may help to explain errors in the interest rate forecast. We find that the pattern of inflation forecast errors is qualitatively much the same as those for interest rates, but that the inflation forecasts are quantitatively better, both in terms of prediction error and of bias. The evidence on the relationship between inflation forecast errors and interest rate forecast errors is mixed. Over the whole time period, both in NZ and UK, there is no such relationship. But if one should strip out certain short periods, when domestic interest rates appear to have been affected by external factors, then there does seem to be such a relationship, with under (over) estimates of future inflation associated with under (over) estimates of future policy interest rates.

Download