Gareth Morgan, Director of Gareth Morgan Investments
Now that the RB has averted the fall in the TWI below its tolerance level of 53, albeit at a higher level of interest rates, it’s appropriate for them to let us know what their new forecast of inflation is. After all, while their policy reaction has avoided the inflationary impact they fear from the lower currency, it has generated, courtesy of higher interest rates, lower inflationary pressures from the domestic economy which will have to suffer these higher rates. Presumably the 0.8% forecast has now been lowered.
Herein lies a conundrum for RB exchange rate policy. Their policy reaction changes the underlying determinants of inflation and so only if they are revising their inflation forecast with the frequency that they’re changing interest rates to maintain a predetermined exchange rate condition, is their fine tuning of the currency a transparent process. Indeed to aid that transparency it would help if each inflation forecast was accompanied by the forecast contribution from the individual inflation determinants – exchange rate, overseas prices, wage costs, domestic margins, etc. Only then can we assess just how significant and credible the RB views of the contribution from the exchange rate, or from wage pressures or anywhere else, are.
Obvious from recent events is the reality now that the RB forecasts of inflation are all important in any market assessment of the future course of the exchange rate. Full disclosure of those forecasts and their composition is a necessary adjunct to that assessment. Apart from disclosure of the deterministic model of inflation forecasts the RB issues, the second issue that this fine tuning of the exchange rate raises is just how precisely the Bank is able to forecast inflation anyway. Are their forecasts sufficiently accurate to give meaning to the type of fine tuning we have seen? Exchange rate movements of 0.8 on a 53.8 TWI arguably imply CPI affects of at most 0.5%. Whether this degree of movement warrants a policy reaction depends upon how accurate RB inflation forecasting
Based on their forecasting performance since the RB Act was introduced, we estimate the standard deviation of RB one year ahead forecasts as 1.7%. That is, the actual outcome of inflation, when forecast at 1% has a 67% likelihood of being between -0.7% and 2.7%. This translates to a greater than 50% chance of being outside the 0-2% range. With such imprecision it is perhaps overly ambitious to assert that a postulated 0.5% inflation shock warrants a policy reaction. After all we can be no more confident that the shock is lifting inflation from 1.7% to 2.2% than we are that it’s lifting it from -0.7% to -0.2%, given the imprecision of RB inflation forecasting.
The two issues of forecast transparency and forecast accuracy raise doubt on the efficacy of exchange rate fine tuning. Whether the state of the art of inflation forecasting is sufficiently robust to enable the current RB Targets Agreement to be met remains debatable. Irrespective of the final outcome of the recent currency/inflation forecast imbroglio, we can’t be confident that the RB Act has yet found its final form.