Why RBI’s Inflation Projections Have Been Missing The Mark
Retail inflation fell sharply last year as food prices softened unexpectedly.
That drop in food inflation also meant that the Reserve Bank of India’s inflation projections went wrong, leading to questions about the need for the two interest rate hikes announced in 2018. Those rate hikes have been since reversed by the two consecutive rate cuts announced by the MPC in February and April 2019.
Inflation projections are particularly important in an inflation-targeting regime where the central bank uses the projection as an intermediate target and acts accordingly. As such, a projection which is off the mark leads to a greater risk of policy errors.
Addressing these concerns, deputy governor Viral Acharya said in the February 2019 monetary policy meeting that the central bank was introspective of the “lament” over its inflation forecasting ability. Citing an internal study, he said central banks tend to make large errors in economies where food is a bigger component of the basket.
A study published as part of the RBI’s Mint Street Memo series on Thursday put up the evidence behind Acharya’s comments.
Food And Demonetisation
The study pointed out that the large share of food in India’s consumer price index and the shock imparted by demonetisation had a role to play in forecasting errors. When adjusted for these, projections by the RBI compared well with estimates by other central banks taken in the sample set, the study added.
From April 2015 to December 2018, according to the study, actual inflation outcomes deviated significantly from the RBI’s projections for a quarter ahead twice—during the third and fourth quarters of 2016-17, and the second quarter of 2018-19.
The deviations were led by large and unanticipated shocks originating primarily from vegetables and pulses, according to the study. Vegetable prices rose sharply in April-June 2016 and collapsed after demonetisation. Prices of pulses remained elevated from January 2015 through July 2016, before starting to moderate.
The correlation between the monthly changes in overall food inflation and vegetables prices has been as high as 0.9. As a result, food prices began to deflate after rising by as much as 8 percent in July 2016. In the second quarter of 2018-19, the seasonal uptick in the prices of vegetable in July-September 2018 was the lowest since the release of the latest CPI series and the prices didn’t pick up even after that.
Both episodes were associated with unanticipated sharp fall in food inflation particularly in case of perishables. The first episode was in the context of demonetisation and the second related to the unusually low summer season vegetables price rise, followed by significant winter season correction due to unanticipated excess supply conditions.RBI’s Mint Street Memo
Among non-food items, volatility stemmed from global crude oil prices.
RBI vs Other Central Banks
In an analysis across select countries, the RBI found a strong correlation of 0.9 between the share of food in CPI and the volatility in headline inflation. At 45.9 percent, the share of food in CPI in India, is much higher than the 10-20 percent in other countries.
The mean error, when calculated over all four time horizons that the RBI forecasts, stood at 70 basis points compared to 0-40 basis points for the other countries in the study.
However, this changes when the time period of demonetisation is excluded.
Excluding the demonetisation period from October 2016 to June 2017, mean error for India stood at 30 basis points, with forecasts biased towards the upside. This, the RBI said, is within the range of forecasting errors across other central banks.
Some central banks the RBI benchmarked its inflation forecast to, such as those of the U.K. and Sweden, have been successful in forecasting inflation with zero mean error for the duration studied. Other central banks included in the study were New Zealand, Hungary, South Africa and Czech Republic.