Is ‘cherry picking’ of data is what it’s all about?

By Ravi Dutta Mishra

New Delhi, June 30 (IANS) RBI Governor Shaktikanta Das recently cautioned against "cherry-picking" of data, subtly arguing against the former Chief Economic Adviser's paper, but neither was Arvind Subramanian's paper the only one on GDP over-estimation nor his methodology.

A recent IIM Ahmedabad research paper titled 'Over estimation in the growth rate of National income' by Sebastian Morris and Tejshwi Kumari used similar methods to conclude that India over-estimated its growth rate.

"Ever since the Central Statistical Organisation (CSO) changed the method of computation of national output (income), from what was centred around final goods to a dominance of the use of estimates of value added across productive entities, the growth rate from the latter series did not seem to reflect the reality, " the research paper said.

While Subramanian's research paper showed that India's GDP growth had likely been overestimated by about 2.5 per cent points per year from 2011-12 to 2016-17, Morris and Kumari's paper said the growth was most likely under 6 per cent and closer to 5-5.5 per cent.

Subramanian's paper caused a huge controversy and a fiery response from the Economic Advisory Council to the PM.

The council gave a "point to point" rebuttal to his paper. Among the several points raised by the council was usage of selective indicators to derive a conclusion.

Besides, the council had said India has frequently revised its GDP calculation methodology and the latest one by Central Statistics Office (CSO) that changed the base year to 2011-12 along with other methodological changes, (as many countries do regularly) was no different.

"Trying to find suitable proxies or approximation of GDP growth rate, when the official estimate is faulty is a legitimate exercise," Professor R. Nagraj of Indira Gandhi Institute of Development Research (IGIDR), Mumbai told IANS.

"In fact, Subramanian has said that he was inspired to do the study by a paper on Chinese GDP growth overestimation published in Brookings Papers earlier this year, " Nagraj added.

Nagraj also said there is nothing unusual about Subramanian's effort to come up a proxy for official GDP growth rates using co-related variables.

Much of the criticism of GDP growth rates was that the official estimates were quite at variance with other macro correlates such as growth in bank credit, capacity utilization on manufacturing, investment growth, Nagraj said.

However, noted statistician Pronab Sen said that the most of such papers are produced considering the old method of GDP calculation as the correct one, which might not be the case.

Sen said that the new methodology was brought it to correct the flaws of the old methods, "Otherwise why would one use new methods?"

Several doubts arose after India brought changes in the way it calculates its growth rate, chiefly because it showed that India grew at a faster rate under NDA 2 than under the previous UPA 2, particularly when economists across the board had said the demonetization exercise will take a toll on the growth rate.

(Ravi Dutta Mishra can be contacted at ravidutta.m@ians.in)