RBI raised the repo and reverse repo rate by 25 basis points last Friday, signalling its concern about inflation. India?s headline inflation for February zoomed towards the double-digit mark at 9.89%; the manufacturing component also inched to 7.42%. Figure 1 shows the recent inflationary scenario in the food and manufacturing components of Wholesale Price Index (WPI).
On the food inflation front, the rise to 17.83% of the year-on-year growth rate of food prices in February is worrisome. High food inflation is raising a storm, but what appears to be a far bigger problem are the factors pushing up the manufacturing WPI. The high food inflation is usually explained as a result of supply-side constraints that are beyond the domain of short-term policy intervention, but spilling over of inflationary pressures on non-food items is a cause for concern. The year-on-year growth of sugar prices has risen to 56% in February. It is to be noted that sugar is classified in the manufacturing component of WPI. The year-on-year fuel prices have also touched double digits at 10.1% in February. The fuel price rise feeds into manufacturing inflation via rise in production costs. Robust growth in the manufacturing sector, exceptional growth in the capital goods sector and the pick up in bank credit are the potential factors on the demand side pushing up non-food inflation.
The shift of inflationary pressures from food to non-food items becomes more evident if we capture the short-term dynamics in prices using month-on-month inflation measures. Figure 2 shows the month-on-month growth rates for seasonally adjusted WPI-food and WPI-manufacturing.
While the month-on-month WPI-manufacturing is hovering around 6-7%, the month-on-month numbers for WPI-food show a sharp fall to around 1% in February, but the former feeds an upward pressure on overall WPI, forcing it towards the double-digit mark.
The key question that arises in the given scenario is, do we see any signs of inflation easing in the coming months?
We attempt to answer this through a simple approach of forecasting that is based on understanding the time series structure of the series under consideration. The approach is based on forecasting using Auto Regressive Moving Average (ARMA) models. The forecasting methodology does not reflect the impact of economic considerations like changes in the monetary policy stance, exchange rate pressures and oil price shock. It solely reflects the autocorrelation structure of the series under consideration.
Our analysis shows that the year-on-year WPI inflation is likely to touch the double-digit mark in March. It will then cool down moderately but is likely to remain in the range of 7-8% until August. Figure 3 shows the forecast for WPI year-on-year for six months. A similar exercise is performed for the manufacturing component of WPI. We find that the year-on-year numbers are likely to remain in the range of 8-9% until August.
There is an inherent flaw in the institution of inflation measurement in India.
Inflation measurement is based on year-on-year growth, which is a lagged indicator of the price situation. It captures the developments of the last twelve months. In order to track the current price situation, it is better to look at the month-on-month numbers. The month-on-month numbers are affected by seasonal fluctuations, hence growth rates of seasonally adjusted series are analysed.
We perform the forecasting exercise on the seasonally adjusted month-on-month WPI. The calculations reported here are based on the seasonally adjusted data published and updated every Monday at http://www.mayin.org/cycle.in. The predictions reveal that seasonally adjusted WPI measured on a month-on-month basis will decline to the comfortable below-5% zone by July.
A similar exercise is performed for WPI-manufacturing. Since the series does not show an identifiable seasonal pattern, we analyse the month-on-month numbers without seasonal adjustment. We find that the month-on-month WPI-manufacturing is likely to remain in the range of 4-5% for the next six months.
The prediction of food price inflation is not attempted because of its volatile nature. The time series model of such a volatile series renders the estimates unstable, leading to unreliable forecasts. In general, the month-on-month indicators give a more timely signal and the year-on-year indicators tend to follow the month-on-month numbers in a sluggish manner. This is evident from the accompanying table that shows our forecasts for WPI inflation based on both year-on-year and month-on-month numbers. The month-on-month forecasts show a moderation of inflation from July onwards, while the year-on-year forecasts do not show any sign of moderation before November.
?The authors are economists at the National Institute of Public Finance and Policy