In a couple of days from now, farmers around the country will begin to look skywards for signs of the first rains. So will the officials of the India Meteorological Department (IMD). For even though the IMD has forecast a normal monsoon this year-its 18th consecutive "normal" forecast-the past three years' record has cast a shadow over the credibility of its predictions.

FORECASTING FOIBLES What the Meteorological Department assured and what we got

IMD has forecast normal monsoon for 2005. But excessive snowfall, cooler than normal summers are worrying. Forecast was off the mark in two of the past three years. Normal rainfall was predicted for 2002 and 2004-both were drought years. IMD changed its forecasting model in 2002. This year too it is experimenting with a new model.

In 2002 and 2004, despite the Met Department's prediction of normal monsoons, the rains turned out to be both deficient and delayed in many regions. This year officials have been cautious. They have already predicted a week's delay in the arrival of monsoons in Kerala (June 7 instead of June 1). But the overall monsoons are predicted to be normal. To be precise, IMD estimates the rainfall between June and September 2005 to be 98 per cent of the long-term average. Rainfall above 90 per cent of the long-term average is considered to be normal.

Can we believe the IMD? Already some weather disturbances threaten the onset of the rains. Excessive snowfall last winter coupled with cooler than normal summers till recently have prevented sufficient heating over the Indian subcontinent. That is the reason for a week's delay in the arrival of monsoons in Kerala. "Monsoons form in response to the heating that takes place over the subcontinent. If the heating is not strong enough then they get delayed," says S.R. Kalsi, additional director-general, Meteorology with IMD. If apart from arriving late, the rainfall also ends up being deficient, it will be the third year in the past four that the country will get inadequate rains. In 2004, 44 per cent of the districts in the country had deficient rainfall. In 2002, 61 per cent of districts received less than normal rains.

To be fair, weather forecasting is one of the riskiest of all forecasting. The build-up of the monsoon winds is influenced by interactions among a number of local, regional and planetary responses. On its part, notwithstanding the past forecasting foibles, the Met Department has been upgrading its forecasting model. The first error in its judgement-the 2002 forecast-had coincided with a change in its forecasting technique. The IMD relies on empirical models where the impact of several parameters influencing the monsoons is captured in mathematical equations to arrive at forecasts. Since 1988, the Met Department had been using a 16-parameter model. In 2002 the number of parameters was pruned to eight. This year a new model is being tested at IMD's Pune office which will track monsoon through just 5-6 key parameters. But the forecast for 2005 has been arrived at by using the eight-parameter model.

The experience of the past few years shows that even if monsoons are statistically normal-that is, rainfall eventually does turn out to be over 90 per cent of the long-term average-its skewed spread over time could play havoc. For instance, in 2002 overall monsoons were deficient by only 19 per cent, but the deficiency in the critical month of July (the month of sowing the kharif crop) was 49 per cent. Along with timing, geography plays a critical part too. Uttar Pradesh, Punjab, Haryana, Rajasthan and Madhya Pradesh account for 86 per cent of India's wheat output. Lower than normal rainfall in some of these states will severely curtail the wheat crop, even if the national average of monsoon is normal.

This is why economists and corporate executives aggregate rainfall to be less significant than the timing and the spread of monsoon. "A deficient monsoon is not necessarily correlated with deceleration in agriculture growth. Deficiency has to be widespread and during the sowing season. For instance, low rainfall in June may not matter much because farmers can postpone sowing to July," says Subir Gokarn, chief economist with CRISIL. In 2004, the monsoon was 13 per cent less than normal, yet agriculture growth was positive. Shiv Sivakumar, chief executive, Agri Businesses, ITC, explains that a 100 per cent rainfall which is an average of 120 per cent (excessive) rains in certain regions and 80 per cent (deficient) in others is less desirable than a uniformly spread out 90 per cent rainfall.

Beyond all this, the monsoon's influence on agriculture in particular and economy in general is also on the wane. As Sivakumar points out, Indian agriculture in getting more diversified and the share of rabi season crops (October to March) in total agriculture production is rising. Unlike kharif season crops (June to October) which rely heavily on the monsoon, rabi crops depend more on residual soil moisture and temperature. An Agriculture Ministry estimate shows that the share of rabi crop in total food grain production has risen from 36 per cent in 1970-71 to 48 per cent in 2003-4, with a corresponding fall in the share of kharif crop. Trends in the past 15 years of agriculture production show that a good monsoon doesn't always result in good agriculture, though a year of bad rainfall does stymie it. Over the years the degree of damage is also declining.

If the monsoons matter less for agriculture today, they matter even less for the economy. To begin with, agriculture's share in the country's gross domestic product (GDP: monetary value of all the goods and services produced in a year) has shrunk to less than 20 per cent. Till the early 1990s, agriculture contributed 30 per cent of the GDP. Thanks to a huge build-up of foodgrains stocks with the government, a falling agriculture production doesn't translate into a spurt in food prices any more (see graphic: Fading Impact of Monsoon).

Another critical link between agriculture and the economy is rural demand. To the extent a poor crop reduces the farmers' income, it also reduces rural demand for non-agricultural products. That, if it does happen this year, could take some steam out of the rip-roaring industrial growth. But here too, things may not be as bad as they have been in the past. Says Siddhartha Roy, chief economist, Tata Group: "Rural demand comprises agriculture as well as non-agriculture income. The share of non-agriculture income in total rural income has been rising. So, don't expect demand for consumer goods to come down significantly in case rains are inadequate."

All this proves is that the monsoon's hold over the nation's economic life is declining. But it is not dismal-not as yet. If and when the rainfalls do turn out to be less than normal, expect sentiments to start getting shaky all the way from the stock markets to state capitals. The chorus of demands for relief and subsidies will stretch government expenditure and fiscal deficit. And even a minor upward pressure on prices, at a time when oil prices threaten to stoke inflation every now and then, would become difficult to manage.

The Government can, of course, do a lot to further dilute the impact of the monsoons on the economy. In addition to the long-term factors like increase in area under irrigation, the Government can speed up crop diversification through a pricing policy that is not biased in favour of rice and wheat cultivation. Removing restrictions under mandi laws and opening up inter-state trade are measures that will increase private investment in agriculture and increase the corporate sector's stake in Indian agriculture and farmers. Short of all these steps, let's hope that IMD does not let the economy-and itself-down for the third time in four years.

Index