Duvvuri Subbarao is unlikely to forget 2012/13 in a hurry. The Reserve Bank of India governor had to spend the better part of the financial year trying to figure out just how much monetary policy could be eased and when. It proved to be a particularly difficult task largely because his team's gross domestic product GDP) forecasts , on which monetary policy decisions are taken, were constantly changed.

In a space of just 10 months in 2012/13, April to January, the RBI put out four forecasts of GDP. Beginning with 7.3 per cent in April, the number was thrice revised downwards. By end January Subbarao forecast the economy would grow just 5.5 per cent in the current fiscal year.

The perils of forecasts in turbulent times were once again brought into sharp focus on February 7 when the Ministry of Statistics and Programme Implementation came up with its advance estimate for 2012/13. If proved right, its projection of five per cent growth for the current fiscal year would make it the slowest growth year in a decade.

On the same day, the Finance Ministry pointed out the limitations of the Statistics Ministry's methodology while arriving at the advance estimates and reaffirmed that it would stick to its own forecast that the economy would grow a little faster. The projection by the Statistics Ministry was largely based on data for the first eight months of the fiscal year. The Finance Ministry mandarins argue that estimates based solely on past data can be inaccurate.

The Chief Statistician of India and head of the Statistics Ministry, T.C.A. Anant, is unfazed. "You can only capture what's there in the figures," he says. Economists who make forecasts have a different take. They claim the fuss is really about techniques of forecasting, But the war of words also encapsulates the difficulty in making projections in turbulent times. N.R. Bhanumurthy, professor at National Institute of Public Finance and Policy, works on GDP forecasts for the Link Project, one of India's early private initiatives in making forecasts. He classifies the Statistics Ministry's advance estimate as the projection of existing trends.

The others in the business make assumptions about everything from the price of a barrel of oil to what Finance Minister P.Chidambaram would do in the Budget, he says. "What we do is a conditional forecast and it depends on various assumptions," says Bhanumurthy, who last July forecast 5.9 per cent growth for 2012/13. The Link Project began in 1991 out of a collaboration between Delhi School of Economics and Institute of Economic Growth.

Anant's counter to the criticism about advance estimates is that his ministry is merely following its mandate. India's official statistical system uses the United Nations System of National Accounts as a template. A team of economists from outside the government, the Advisory Committee on National Accounts, also guides the official system. "It is a simple projection," says Anant. "It is based on the previous (financial) year's data as well as current year's."

What is implicit in Anant's explanation is that a simple projection will not take into account the two big policy announcements in the current quarter (January to March): RBI's decision to loosen monetary policy and the Budget, of which expectations are running high. Does that imply people who factor in these announcements are more likely to be accurate? Credit rating agency CRISIL has analysts tracking 70 industries, making for an exhaustive database from which GDP forecasts can be made.

Its Senior Director and Chief Economist, D.K. Joshi, who heads the team of economists at the agency, doesn't seem to have it easy. "We revise our estimates around four times (in a year), sometimes even five," says Joshi.

Finally, it all seems to boil down to the environment. "Forecasting is easier in a stable environment," says Joshi. Anant's "simple projections" are also likely to be more accurate in a stable environment.