The nagging 'formula' that is giving India's politicians and statisticians sleepless nights

"We are trying to compile a historical series and hope to provide it by the end of December," he said.

Until then, the questions may keep on coming - a state of confusion that is persisting to the point where it risks wrong-footing financial markets and even policymakers.The statistics office has set up an internal panel to audit the data and invited an International Monetary Fund team to review it. The IMF review starts on April 22.

Last week, the Reserve Bank of India (RBI) pointed out several gaps in the new figures, which it said clouded an accurate assessment of the economy and could lead to poor policy.

Others have been more scathing in their criticism. Morgan Stanley's Ruchir Sharma, for example, has called the new numbers a "bad joke" aimed at a "wholesale rewriting of history".

The statistics office says the new calculation, which measures GDP by market prices instead of factor costs, is more in line with global practise and helps to better understand the structural changes taking place in the economy.

The RBI and finance ministry agree, and aren't questioning the decision to change the methodology. But they can't reconcile the data with other indicators showing less vibrant growth.

For example, the new, robust expansion doesn't correspond to subdued corporate earnings and weak industrial production - and statisticians disagree over the way a new data set is used.

The statistics office attributes the divergence to the use of an improved database of hundreds of thousands of private companies. Previously, a small sample of large firms was used.

But R. Nagaraj, an independent member of a government-appointed panel that first suggested using the companies database to calculate GDP, says the mismatch reflects a decision to extrapolate the new data instead of taking it as a straight reading of output.

As a result, Nagaraj said, gross value addition in manufacturing is more than double the figure his panel's suggested methodology would have produced.

Kumar, the head of the statistics office, said the rationale was to make the samples comparable over a period of time.

"There has to be some benchmarking so that you compare the comparable," Kumar said. "We haven't come across any defect in our methodology."

"We don't have any political agenda," Kumar said. "Whatever questions are coming, we are answering them."