In 2015 India has changed its data sources and methodology for estimation of GDP. Since then, many have raised questions about the veracity of our GDP estimates. Economists have expressed their concern over the MCA-21 database used in GDP calculations. Over it, Center’s deferring the release of NSSO’s labor force survey that portrayed disturbing scenario has indeed weakened faith in the official data. As many as 108 economists issued a joint-statement in March calling for the government to restore independence of India’s statistical institutions. This is further worsened when a newspaper reported that NSSO, having found gaps in the MCA-21 data that it sampled, had to drop its two planned surveys on the service sector.
However, it is worth recalling here that the new series data that has gone into estimation of GDP growth chose to measure output by taking into account data on corporate profits as against the earlier practice of using volume-based data. And theoretically, this was considered a major improvement over using output surveys since it was felt that as corporate data was audited and also as it formed the very basis for their tax payment, there was hardly any incentive for misrepresenting data. But with NSSO reporting that over a third - 36 percent - of the companies considered active firms under MCA-21 database could not be traced, critics have once again raised their voice fearing that the MCA-21 database would inflate GDP estimates. Though most of these critics believed that there is an overestimation, no one has attempted to quantify it.
Against this backdrop, Arvind Subramanian, former Chief Economic Adviser to Ministry of Finance, has published a working paper from the Center for International Development at Harvard University, claiming that India’s GDP growth from 2011-12 to 2016-17 is likely to have been overestimated by 2.5 percentage points per year. This has obviously reignited the uproar in the media.
Let us first examine what Subramanian did: He has taken 17 real indicators and tested the correlation between their annual growth and our GDP growth. Based on the year in which the change in methodology of calculation of GDP was introduced by the government, he had divided his study period into two parts: pre-modification period (2001-02 to 2010-11) and post-modification period (2011-12 to 2016-17). His correlation study revealed that the indicators were positively correlated with the GDP growth for the pre-modification period. They, however, broke down for the post-modification period since growth of these indicators was “substantially lower in the post 2011-period.” This negative correlations for post-modification period, according to him, mean that GDP was substantially overestimated.
He then did a cross-country regression that related GDP growth of 70-plus countries with four indicators, viz., credit, electricity, exports and imports, and drew a trend line. It revealed that India’s GDP growth followed the trend line till 2001-02 to 2010-11 but subsequently, i.e., during post-modification period, it got deviated from the trend line significantly. He then estimated the likely GDP growth had India remained on the trend line. Based on these findings, Subramanian states that the “methodology changes introduced for GDP estimates resulted in an overestimation of GDP growth for the period from 2011-2016 by about 2.5 percentage points per year, with a 95 percent confidence band of 1 percentage point.”
He then made that striking statement: the actual growth during 2011-16 was more likely to have been 3.5 to 5.5 percent as against the reported average of 6.9 percent. This obviously casts doubt on our stellar growth story. And such doubts, critics believe, are more likely to ebb-out inward flow of foreign capital. Obviously, this has raised much uproar in the media.
The Economic Advisory Council of Prime Minister has, of course, countered Subramanian’s claim with an 8-point rebuttal. The Confederation of Indian Industry too criticized the paper, drawing attention to the failure of paper in taking into account sizeable agricultural sector, parts of fast-growing IT industry and services sector. Some have dubbed his research as bad econometrics for, according to them he has used “70 countries as control and only one country (India) as treatment” and hence conclude that Subramanian’s “2.5 percent GDP overestimation and its confidence interval are highly suspect”.
Nevertheless, as the very rebuttal of The Economic Advisory Council of Prime Minister states, “This certainly does not mean that the paper should not be taken seriously”, one cannot simply dust off former Chief Economic Adviser’s findings. Further, it is necessary here to bear in mind that Subramanian has adopted a similar method that economists in the West often tried out, i.e., to estimate the real GDP of countries when their official data smelled dubious by analysing top economic indicators. Further, in support of his research, Subramanian argues that no country has grown over 7 per cent with exports below 5 per cent.
That said, it must also be admitted that with the kind of FDI inflows and the capital market activity witnessed, it is too difficult to accept his findings that India’s GDP grew @ 3.5 percent during the period 2011-12 to 2016-17. Nevertheless, we must honour his research and its conclusion that India’s GDP is overestimated by 2.5 person per annum post-modification as an econometric guesstimate. And this acceptance incidentally casts a doubt on the official datasets and the methodology adopted to estimate GDP. Even Pronab Sen, former Chief Statistician of India, said that growth after 2016-17 “is being overestimated. But by how much, I have no call.” It therefore calls for a healthy debate from the intellectuals and a relook from the government at the whole process of GDP calculation, for it has tremendous impact on policy making. For, the need of the hour is: transparency in economic data.