Wanted: A professional statistical set-up

It should be oriented towards capturing the facts and making them public in a timely manner

Errors, both of omission and commission, by the government of India’s statistics-economics fraternity has led to a grave situation. The net result is that long-term policy formulation has now largely gone blind. At the micro level, this situation has arisen where both growth and employment don’t have historically comparable data. All that we at best know is whether we are doing better or worse than the past, and the robustness of even those claims is debateable. Moreover, we know little on the components and the character of the massive economic change that is occurring in India. As a consequence, policy can now no longer be assessed or formulated in an informed manner, and so what we are left with are uninformed shouting matches.

Quite inexplicably the large and small sample surveys on employment and consumption were not carried out by the National Sample Survey Organisation since 2011-12. A large sample survey occurred in 2009 but the powers that be did not like its results so another one was held in 2011-12. It would have been great if a large sample employment survey were carried out in 2014-15 or at least a small sample one. But it was not, and today in 2018 we don’t know how the employment and its character have changed this decade. Has employment gone up in the country? We don’t know? Are the educated more likely to find a job? We don’t know. Are women who had withdrawn from the workforce in the 2000s, returning in 2010s? We don’t know. Is IT employment rising more rapidly or slowly? We don’t know. Are security guards still among the most rapidly growing employment options as they were in the 2000s? We don’t know.

And then the worst omission of them all. A new GDP series was made public with a new and improved methodology or so it was claimed. But behind excuses of being understaffed and underfunded, no studies that went into the details on comparability were shared with the public. A document was put in the public domain that essentially said that the GDP of the past and the new GDP are not comparable. In other words, sequence of a GDP growth series that start from the middle of the last century suddenly stops at 2011-12. And what is claimed to be a new and improved one is being forced on us, with no proof on why it is improved, apart from some paper claims.

What is the problem with the new GDP series? We don’t know, and we don’t even know if there is a problem! It is simply something different. Any half decent statistical entity would have first done a comparison of how things may change. It would have shared these with the public, ascertained the pros and cons. But most important, it would have estimated a back series before going ahead with implementing a new methodology. The back series would have helped compare the old and the new. CSO says we don’t have enough data for a back series. If data paucity was such an issue, why not construct a forward series of the old productionbased method. Even though that is not the first best option, at least we could have had some comparability.

It is now widely accepted that the period 2004-2011 saw low employment growth despite high economic growth. We can compare GDP and employment growth in that period. But we cannot do so for the following period as employment figures don’t exist and GDP is not comparable. Economic policy formulation therefore has gone blind. Mr Jaitley can make a claim, and Mr Gandhi can make a counter claim. Who can tell who is right? Not the serious economist or the statistician for sure.

What is the way forward now? First, focus on the National Statistical Commission, Central Statistical Organisation and the National Sample Survey Organisation. Clearly these organisations need to be overhauled, which means greater independence, greater funding, and much greater answerability. Second, a superior quality back series of the 2011 base year GDP series must be created irrespective of the costs or difficulties. If the data are not there, a series of micro or dipstick studies could be used. Or a new 2011 series GDP data using the old production method could be created. Admittedly these are not first best solutions, but at least the new and the old can then be compared. Whatever route they take, ensuring comparability must be a precondition for all future senior hiring in India’s statistical set-up. Third, ensure the results from the detailed employment survey are put in the public domain speedily irrespective of which state or sector does well or poorly.

Finally, India needs an independent economic set-up which is not about judging performance of any government, but simply oriented towards capturing the facts and making them public in a timely manner. These facts may be convenient or inconvenient, to the government or the Opposition, to the bureaucrats or economists that advise it, but they must be brought out. This is the only responsibility of a professional statistical set-up.