With weekly inflation levels crossing 8 per cent, most in the government are talking of a reduction in inflation levels only after the monsoon.

In the interim, a battle rages between those who believe curb on export and other supplies are the way to tackle things and those who believe these reduce supplies.

A fresh study by the IMF's India office, however, suggests the battle against inflation may already have been won as far back as last month.

In which case, supply-side measures by the government or monetary measures such as the increase in the cash reserve ratio by the Reserve Bank of India [Get Quote] would have been unnecessary.

Official inflation is measured by the change in the price index (wholesale or consumer) over that a month ago (year-on-year). What the IMF has done is to measure the monthly change in this index, and then removed the impact of the change that takes place due to seasonal factors such as the crop coming in.

The reason some economists prefer this is that the year-on-year index is heavily influenced by single events (a sharp rise in prices last year will artificially lower inflation the next year, for instance) while the seasonal adjustment eliminates this. The monthly data, seasonally adjusted, show that inflation started declining in April.

The IMF paper also does a seasonal adjustment on the weekly data, calculates the average inflation for the past five weeks and multiplies this by 52 to arrive at the likely inflation for the year. It says this is just 4.7 per cent, a figure that shows inflation is already within the RBI's comfort zone of 5 to 5.5 per cent.

Yet, despite inflation being comfortable, the IMF points out, the year-on-year changes would still show (erroneously) inflation levels to be very high. If inflation continues at the average of the past five weeks, the IMF says, the year-on-year inflation rate would still rise to 9 per cent till June, making headline news -- it would fall below 6 per cent only in February.

To show how misleading the year-on-year analysis is, the IMF constructs a scenario where, beginning now, inflation falls to zero (that is, prices simply do not rise over current levels). Under this, the year-on-year method would still show inflation rising even next month and it would fall below 6 per cent only in November.

Not surprisingly, the study has as many supporters as detractors. Credit rating firm ICRA's economic advisor and member of the Prime Minister's Economic Advisory Council Saumitra Chaudhuri is dismissive of the seasonal adjustment of monthly data and says it is very sensitive and can end up giving incorrect results.

Interestingly, JPMorgan does similar seasonal adjustment on a monthly basis and while its numbers are similar to IMF's from January 2006 onwards, they differ from March/April 2008 -- for April, the inflation levels are 1.02 per cent against the IMF's 0.2 per cent.

Also, since over half the items in the Wholesale Price Index haven't been revised for the last one month and around 40 per cent for the past three months, it is not certain whether the IMF analysis factors this in.

A sudden revision of data, as happened in March when prices of several steel items were raised after 27 to 53 weeks for instance, saw the index jump to a 10-month high. 

DK Joshi, director and principal economist of credit rating firm Crisil points out that while 4.7 per cent inflation may be within the "comfort zone", this mustn't be confused with the RBI's comfort zone since the latter is defined in terms of year-on-year inflation, not the month-on-month number.

Oxus Investments chief Surjit Bhalla, however, also does such seasonal adjustment and concurs with the IMF results. Pronab Sen, the country's chief statistician, hasn't examined the IMF study but concurs that inflation growth has probably peaked.

In which case, the next task Sen's National Statistical Commission needs to do is to study if the traditional year-on-year method of calculating inflation is giving us incorrect results, leading to incorrect policy actions.

Inflation, the silent killer