We had data up to a point & then made projections: T C A Anant

Interview with Secretary & chief statistician, Ministry of Statistics and Programme Implementation

Advance estimates released by the Central Statistical Office (CSO) have drawn flak from many quarters. Yesterday, Planning Commission Deputy Chairman Montek Singh Ahluwalia raised questions over the methodology. Today, the finance ministry joined the issue. Chief Statistician of India T C A Anant tells Indivjal Dhasmana that advance estimates are based on actual data up to November or December and projections thereafter. He agrees that projections could miss a changing trend. Excerpts:

The finance ministry says the advance data overestimate economic growth when the economy is in upturn and suddenly change track and underestimate when it is in downturn and alters gears. Do you agree?
Let me explain the methodology. We make advance estimates on the basis of availability of data till a point, typically up to November or December in some sectors. Then, we make a projection, using data for the last year.

The criticism made against advance estimates is that you project it in a linear fashion and, hence, have ignored the growth momentum triggered by the government’s recent measures for 2012-13. How do you take that criticism?
It is entirely possible. Projections are based on regression analysis in a linear way. If non-linear trends emerge, if a U-turn is there, it may not be captured. But, I have no information on the non-linear trend. We have data up to a point and made projections thereafter.

Even in your revised estimates released on January 31, GDP growth numbers were changed sharply for some years. For 2010-11, those were revised by 0.9 percentage points to 9.3 per cent from 8.4 per cent estimated earlier. Why did we have such sharp revisions?
It is so because the revisions given on January 31, 2013 were based on the Annual Survey of Industries and the earlier numbers were based on the Index of Industrial Production.

Critics say such sharp revisions meant that policy makers were barking up the wrong tree. Do you agree?
Well, there is no great deal I can talk about. Data are constrained by a number of factors. For one, the Annual Survey of Industries takes time to come.

If you take financial year 2011-12, companies finalise accounts in September 2012. We get underlying accounts of their profit and loss etc. thereafter. It takes us seven-eight months to get the numbers and another five-six months to process those.

Typically, it will be by December 2013 when the Annual Survey of Industries will come.