The IMD has come out with its first forecast of the southwest Monsoon. There was a time they didn?t make it public. I am told the argument not to make it public came from a revered economist who was an economic adviser in the Finance Ministry arguing that a bad forecast would lead to speculation.
I argued that the gentleman was wrong. Dantwala?s classic on cotton marketing and Madhoo Pavaskar?s work on futures showed how savvy the trader was and he really did not need the MET to bolster his information sources. In fact, the forecast would, other things remaining the same, stabilize the market. Now they give us one forecast now, and the distribution at a later stage.
The model they use is not causal in the sense that it doesn?t tell us why the rainfall level will be what it is expected to be. It is a statistical model in which associations of eight independent variables have been examined for over a hundred years of data with the rainfall in the southwest Monsoon. These are El Nino in July- September last year, Eurasian snow cover in December, European pressure gradient in January, North West Europe temperature in January, Arabian Sea surface temperature in January and February, wind pressure pattern in January and February, East Asian pressure in February and March and South Indian Ocean surface temperature in March. The model is the best we have and it is good to go by it until we do better.
I say so since there are a lot of wise guys who love to run down our scientists and statisticians. I used to love this part of my work when I was a Minister since I am by training an econometrician and the economy, medicine and the MET are all stochastic processes. It was great fun to defend them and say how we are trying to do better.
Of course, a better forecast does not mean more rain. We don?t still know the details of the determinants of the forecast, only that the El Nino off the Pacific coast is benign. With a ninety five percent probability it can be said that the SW monsoon rainfall will be 96 percent of the Long Period Average of 89 cms. It would be nice if the MET would also release an Excel sheet on the Net for the curious. Of course, we have had a series of great monsoons and it has been more than 96 percent since 2004.
The more interesting aspect will come later because even ninety six percent of the average can hide a lot of variation within the country. Normally we end up in fairly serious trouble if around a fifth of the MET regions are deficient in rainfall. The probability of this happening in a 96 percent year is low, but is possible. Anyway, the big swings in cereal production and oilseeds and cotton take place in the unirrigated areas in Rajasthan, in Gujarat and parts of Maharashtra and the Deccan plateau. A swing of six million tonnes of grain and different news in oilseeds can be on the anvil.
An interesting aspect of the Indian economy is that its growth performance does not depend on the monsoon. In the 1990s, I had shown that in the period till the mid 1970s, in half the years GDP growth was over four percent and in the remaining very low or negative and since the mid seventies we had only three years with growth less than four percent.
This, of course, does not mean that the monsoon does not matter. A bad monsoon can cause a lot of misery to a lot of people in terms of work, income and drinking water, but we are better placed to take it on.
And please support the science establishment. We need to give them the serious challenge to explain the monsoon in causal and not just statistical associational terms. The monsoon is important and if India wont do it no one else will. It will be another big mission science project of the kind that Rajiv Gandhi had conjured?-he did go for the super computer to sponsor the medium term weather forecasting project at the agro climatic met zones. Big countries have to think big.
?The author is a former union minister. He can be reached at yalagh@gmail.com