The trouble with the recent brouhaha over inflation is that it seems to be based on impressions rather than facts. The local media, for one, which has become remarkably vigilant over the price rise issue, seems to be working on two premises. The first is that global food prices are exceptionally high, which is correct. The second premise is that high global prices and local food shortages are driving domestic inflation to levels close to 8 per cent - this is by no means entirely correct.

Since weekly inflation numbers, for what they are worth, seem to be guiding our industrial, export and monetary policy, it is important to get a handle on what they tell us. Let's look at the data for April 19, released on May 2. The aggregate inflation rate printed at 7.57 per cent for the week.

• Inflation, the silent killer

The latest data release for April 26 incidentally shows a similar pattern. Food articles in the "primary products" category, which include stuff like rice, recorded a year-on-year inflation rate of 5.14 per cent. Contrast this with the iron steel category, which saw an inflation of rate, hold your breath, of 35.2 per cent. In fact, if the iron and steel category is taken out of the inflation basket, the aggregate inflation rate drops all the way to 6.3 per cent.

What are the other big contributors to inflation? There are price pressures for agricultural products but these pertain largely to non-food articles like oilseeds and raw cotton. The inflation rate for the week of April 19 in non-food primary articles was 11.2 per cent.

If we were to exclude the four key contributors - iron and steel, edible oils, raw cotton and oilseeds - the rate for the week would drop to 5.8 per cent. There is an important caveat, though. These prices are wholesale prices, not retail. If retail margins tend to fluctuate, wholesale price movements need not necessarily reflect retail movements. So the exercise of trying to show a neat correspondence between wholesale price inflation data and changes in household spending (again a popular media fad) would be somewhat futile.

Instead of listing more data, let me try to point out the implications of the emerging pattern. First, while I am not trying to undermine the gravity of the global food crisis, it is important to recognise that food prices have played a secondary role in this round of high inflation. Thus it is somewhat misleading to reduce the analysis of current inflation to a debate on food security.

Second, if iron and steel prices have been the biggest driver, a decision to reduce prices that steel producers have taken is likely to reduce inflation pressures. I am not going into the issue of whether "implicit" price controls of this kind can be debilitating for industry. However, if the objective is to bring inflation down as quickly as possible, this could just do the trick.

Third, the categories in which the price hikes have been the sharpest are also those that have large imports and exports. The corollary is that their prices closely track international prices. It does not take rocket science to figure out that excess depreciation of the currency can exacerbate these price pressures.

Unless the government's strategy in the future is to continue to use price controls aggressively to cap inflation, it might be a good idea for the central bank to step in and prevent further fall in the rupee. This would become particularly critical for managing the incipient pressures from oil prices. With international prices showing no signs of abating, the only way to limit oil subsidies and reduce the repressed inflation in oil prices is to keep the rupee in check.

Capping prices and subsidising producers cannot be a permanent solution. Subsidies need to be funded and as the supply of oil bonds (to finance the subsidies) goes up, so will interest rates. Growth will suffer at some stage. There is no free lunch.

Let me turn to the good news. Wholesale price inflation measured year on year (that is comparing today's prices with those last year) often hides more than it reveals. I have often found that comparing this week's prices with the previous week's makes much more sense.

Over some time, the rate of inflation measured week to week has been falling. While the average price index level continues to rise, the rate of increase has been declining. If this trend continues, the price index could level off in a couple of weeks and then, perhaps, even start declining. This should also help bring the conventional year-on-year inflation down. Thus by the end of May, year-on-year inflation could move closer to the 7 per cent mark.

Finally, I think the idea of releasing inflation on a monthly rather than weekly basis is a very good idea. Inflation is reported monthly the world over and I do not see any virtue in our trying to do this every week. Weekly data lead to excessive volatility in financial markets, particularly the bond markets.

Weekly figures are also likely to be more susceptible to large error. The real challenge as far as data go is to improve the quality and make them as representative of actual changes in a household's cost of living as possible. There is also a need to make it as comprehensible as possible - thus along with a monthly inflation series, more emphasis on changes from month to month (instead of just year-on-year variations) is imperative.

The author is chief economist, HDFC Bank [Get Quote]. The views here are personal