Inflation: The Hot Button Policy Issue

Inflation has become the immediate challenge for the UPA government. That was not the perception through the first three months, as issues of apparently greater moment ? basic questions of economic reform and management of differences with the Left parties supporting the government from outside, took centre stage. Even as headline (wholesale price index) climbed past 5, then 6%, the feeling was perhaps that things were under control and it was only a matter of time that as friendly rains quenched the thirst of the parched Indian soil, things would settle down to the comfort zone of 4 to 5%. The spike to 7.5% for the week ending July 24 stirred things up. Government announced that it would take appropriate fiscal measures and the RBI that it would look at suitable responses on the monetary side.

But should the spike in inflation have been so unexpected? We have a serious debility with the manner in which we compute inflation. We measure inflation this week, compared to what prices were in the corresponding week of the previous year. Why do we not compare it with the immediately previous week? Because prices have a seasonal component ? everybody who buys fruit and vegetables knows that ? and the easiest way of normalising for the seasonal variation in prices, is to make the comparison with the same point in the season of last year. Of course, all prices do not reflect seasonal variations ? and over the course of years they occupy an increasingly larger weight in the commodity basket. Is there a way to knock off the seasonal element in the data series? Yes, there are well-established statistical techniques and most official agencies the world over publish de-seasonalised data series.

The principal shortcoming of measuring prices (or for that matter any other parameter) with that in the corresponding week or month or quarter of the previous year, is that we do not know whether the change that we have measured transpired in the beginning of the intervening period or in the middle or recently. Clearly if inflation is high, there is considerable merit in determining whether it was high because prices rose sharply 10 months ago, or over the last couple of months. There are some drawbacks to statistical elimination of the seasonal element, because these are estimates and accordingly can introduce non-transparent estimation errors. Moreover as time goes by, the estimators for excluding the seasonal element does change, leading to revisions in the past data ? and that too is non-transparent because it is indigestible to an intuitive approach. The best thing is of course to look at both the untreated and de-seasonalised data.

But anyway, as already mentioned, a large part of the commodity basket is not intuitively hugely subject to seasonal factors. By and large, the prices of manufactured goods as a general category do not have a significantly large seasonal element. That is also true for petroleum, coal and electricity which is influenced by external factors (crude prices) and domestic political considerations. Anyway, I have computed the 3-month lagging annual seasonal rate for the WPI-manufactured goods data series beginning the end of September 2002. That is, the price level in December 2002 is compared to that in September 2002 and the increase compounded to obtain the annual rate.

You could do this with a 1-month lag, but there is then too much volatility (noise). From this table, it is easy to identify the months when inflation went uphill and when it eased. The uphill tracts are March, April and May 2003, but you could have spotted it earlier in January and February. The next stretch is August, September and October 2003 ? but this time you could have spotted it only in August. The most recent bout is in July 2004 ? evident in the numbers for June. All this, one could summarise, as a perhaps unscientific, but useful approach to track inflation.

To this one can superimpose what we know to be the inevitable that would arise from changes in the world crude oil prices. If we were to present the table for WPI for manufactured good and commercial energy, we would find a big spike of 9.5% in the combined data for June 2004, as prices come off a trough in April and May 2004. The trough was in part an outcome of the electoral process, since the previous government did not want a price hike in petroleum products till polls were over, hoping undoubtedly like most people across the world that crude oil prices would begin to soften ? a prayer that went unanswered.

It is difficult to figure out what the impact of the promised policy package on the course of inflation would be. However, selective tariff cuts could shave off some of the edge of inflation from one side. Some monetary tightening ? the capital market seems to have factored in a 50 basis point increase in the near future ? should take care of the other leg, in areas where tariff changes cannot reach. There are of course many other reasons ? India-specific and globally contextual ? for the RBI to opt for a rate hike, but in conjunction with the policy imperative to contain inflation, a tightening, it seems, is to be more of a sure thing.

The author is economic advisor to ICRA