The rupee and a BMW both move fast. So will a rise in the rupee make the BMW move even faster?

Just-released data on WPI indicate that for the week ended May 26, year-on-year inflation had dropped to 4.85 per cent, a significant decline over the peak 6.3 per cent rate observed in March 2007.

Before that peak, I had written about the possibility that the inflation story was over: (Inflation, An Ex-Problem? BS, March 3): "There is considerable evidence to suggest that the Indian inflation problem is a has-been problem, a problem that has ceased to be�.so look for inflation to fall and for the government to take credit for its anti-inflation policies e.g. banning wheat and rice futures, telling cement makers to behave or else, tinkering here and a band-aid there�When that happens, please remember that the decline in inflation started well before the budget of 2008, and somewhat before the RBI started to be aggressive in raising interest rates and restricting credit."

A few weeks later, the rupee started to move and faster than you could say a speeding bullet, and faster than a BMW M6 accelerating to 60 mph in less than 5 seconds, the rupee had appreciated by 10 per cent, making it the strongest currency in the world.

This was a very non-RBI type of policy, and it generated some debate. Just as the critics were winning out (export growth even before the rupee appreciation was down to 7 per cent an annum, a subject to be examined later), we have now been subjected to the somewhat questionable argument that there have been instantaneous benefits to this rupee appreciation (so forget criticising the strong, stronger, strongest currency policy); in particular, that the scourge of high inflation has been vanquished by the strong rupee and the aam aadmi can now rest in peace. This, as the quote above substantiates, is exactly what I had predicted, but more importantly, feared.

C Rangarajan, former governor of the RBI, claims that the recent decline in inflation has been helped by the rise of the rupee. Ajay Shah, former adviser to the ministry of finance, and a columnist for this paper, goes several BMW steps further (BS, June 6): "An examination of the rupee-dollar exchange rate shows that the old exchange rate regime broke on March 15.... The WPI release of [inflation on] March 17 was the peak". To make sure no one misses the import of this mother of all casual empiricisms, Shah also produces a graph showing y-o-y inflation and a vertical line emanating from the peak and marked the ides of March. There is some valid discussion of how there is a cause and effect between appreciation and a decline in inflation, but the innovation in Shah's analysis is to produce an instantaneous causal effect of rupee appreciation on year on year inflation. To reinforce this instantaneous conclusion, Shah states: "[M]y calculations also suggest that the impact of rupee appreciation to Rs  40.5 upon inflation has largely played itself out". If only all this were true.

Let us consider some basic facts. Year-on-year inflation in primary articles is well above the average, and is currently at 9.5%. For fuel, which accounts for 14 per cent weight, inflation is close to zero; manufacturing inflation (64 per cent weight) is at 5 per cent, identically equal to overall inflation. Looks like food shortages led to a spike in primary articles' prices, and caused inflation to be higher than average. Cheaper food imports (via a stronger rupee) can definitely cause a dent in the price of food, but so small a dent that it cannot be seen even on the super fancy BMW. India may import 3 million tonnes of wheat this year; domestic production is close to 75 million tonnes. How can a small tail (even powered by the rupee) wag such a big dog?

The channel of influence for the "strong rupee is good for inflation" crowd is straightforward. Exchange rate affects traded goods, and if the rupee appreciates, prices in rupees go down, and lower inflation is observed. No one can argue against this simple, perhaps too simple, logic. A first test should be the observation of the opposite: when a currency depreciates then inflation should go up. While intuitive, it is not even obvious that an appreciation of the exchange rate will lead to lower inflation (or vice-versa). Consider the fact that until just a few years ago, currency depreciation and lower inflation were the norm for India. A year after the Asian crisis, inflation fell in the East Asian economies, despite a nominal, and real, depreciation of over 30 per cent.

But Shah (and perhaps Rangarajan) are making a different point -- that one does not know for about a year or so, but for a horizon of a few months, the effect is present and profound. If so, then one should observe that over the last three months, when year-on-year inflation has moved down from 6.3 to 5 per cent, prices with a large share in our imports and exports should have come down a lot. Prices of primary articles have increased by 2.6 per cent in the last three months, or an increase of over 10 per cent annualised.

Manufactured prices have risen by over 4 per cent annualised since February. Fuel prices, a heavily traded category of items, have not changed at all. Wheat prices (wheat is hardly traded)  have fallen by 6 per cent. One could go on, and it would be both wrong and pointless to do so.

Supply shocks, which are unaffected by monetary or exchange rate policy, have a large effect on inflation. Non-recognition of supply shocks in the analysis of inflation can lead to grievous errors, and all too apparent absurdities. The inflation rate in India has come down by 1.5 percentage points largely due to practically zero inflation in primary articles over the last three months, and a zero increase in fuel prices over the last year.

The excess inflation of 1.5 to 2 per cent that we observed was due to the supply shock rise in the prices of these same goods. Neither the way up in inflation, nor the way down, had anything to do with the value of the rupee. To argue so is like arguing that walking and a ride in a BMW M6 is the same because they both get you from point A to point B.