NEW DELHI—Former Chief Economic Adviser Dr. Arvind Subramanian’s working paper claiming an overestimation of the national Gross Domestic Product (GDP) by India’s statistics system between financial years 2011-12 and 2016-17 has set the cat among the pigeons.
The key reason for the controversy: the former CEA has claimed that India’s GDP between the financial years 11-12 and 16-17 was overestimated by 2.5% per year. Thus, the actual growth rate of the GDP was 4.5%. This means India was not the fastest growing economy of its size that was so far believed widely. That Dr Subramanian was himself the CEA during some of the years cited in the study lends weight to his claim and analysis, and ammunition to critics who question why he didn’t speak out earlier.
These were also the years during which India changed both the base year (from 2004-05 to 2011-12) it uses to calculate its GDP and the methodology used for doing so. In January 2015 and in late 2018, the Modi government implemented changes twice to India’s past and present GDP numbers. They have since been politically controversial as the performance of the UPA 2 government looks worse in comparison to the NDA 2 government.
The controversy is far from settled. If anything, Dr. Subramanian’s paper only adds to it by claiming that the overestimation was a result of the methodological changes implemented during NDA 2, though their spadework began during the UPA 2 years.
The former CEA also offers a solution to the controversy. As he mentioned in his piece for The Indian Express, Dr. Subramanian favours the revisiting of India’s GDP estimation by an independent task force comprising both national and international statisticians, macroeconomists and other experts.
HuffPost India explains the claim made by Dr Subramanian in his paper and the official defense by the Modi government to it, in addition to speaking with some of India’s reputed economists and statisticians to understand what they have to say about the claim made in the paper.
A) What is Dr Arvind Subramanian’s key claim and what evidence is it based on?
In his paper, Dr. Subramanian claims that India’s GDP growth has been overstated by about 2.5 percentage points per year after 2011 — implying the economy actually grew at about 4.5% each year, rather than the official GDP growth figure of 7%. He concedes that the data available are not granular enough to give per year estimated GDP for the financial years after 2011 till 2016-17.
The basis for his claim lies in two steps adopted by the ex-CEA in the paper for conducting his experiment. One, compilation of data for the growth of 17 indicators which reflect activity in an economy and have a direct correlation with India’s GDP growth. They are: electricity consumption, 2-wheeler sales, commercial vehicle sales, tractor sales, airline passenger traffic, foreign tourist arrivals, railway freight traffic, index of industrial production, index of industrial production (manufacturing), index of industrial production (consumer goods), petroleum consumption, cement, steel, overall real credit, real credit to industry, exports of goods and services, and imports of goods and services.
These 17 indicators were chosen also because they are produced independently of the CSO and their data are compared for the years from 2001 to 2017. What Dr. Subramanian found is that the rate of growth of the 17 indicators is in sync with the GDP growth before 2011. But in the years post 2011, the rate of growth of these indicators appears out of sync with the GDP. For instance, export (goods and services) growth is 14.5 percent before 2011 and a much less 3.4 percent in the years thereafter. This dip in exports is reflected even though the economy continued to grow in the same range as in the years before 2011. There are corresponding figures for the other 16 indicators in the paper which follow the same pattern.
This, according to the paper, is evidence for mis-measurement. The ex-CEA then uses statistical models comparing data from other countries and observes that the GDP growth in those countries was much lesser than India. Ploughing through some complex data, the paper then estimates GDP overestimation by 2.5 percentage points by India’s statisticians.
b) What was the Modi government’s response?
In a detailed statement, the Ministry of Statistics and Programme Implementation (MoSPI) said, “The GDP estimates released by the Ministry are based on accepted procedures, methodologies and available data and objectively measure the contribution of various sectors in the economy.”
It explained that the changes to the base year and methodology for calculating GDP were made as per the latest international statistical standards decided through the System of National Accounts (SNA) in the year 2008 evolved through discussions by member states in the United Nations. The United Nations Statistical Commission adopted the SNA 2008 in 2009.
The SNA 2008 is the basis for the changes made to the base year (from 2004-05 to 2011-12) for calculating the national GDP in January 2015.
“With any Base Revision, as new and more regular data sources become available, it is important to note that a comparison of the old and new series are not amenable to simplistic macro-econometric modelling,” it said. The statement also noted that the GDP growth projections brought out by various national and international agencies are “broadly in line with the estimates released by MoSPI”.
If he had just said it is unexplained, it would have been alright. He has overreached by saying that this is an overestimation: Former Chief Statistician of India Dr Pronab Sen
c) How are economists and statisticians responding to the claim?
There isn’t a singular view among India’s statisticians and economists about the Dr. Subramanian paper, even though there has been a consensus that the national GDP figures require to be better explained. Here is what economists and statisticians who spoke with HuffPost India said.
Dr Pronab Sen, former Chief Statistician of India
He has gone too far in his interpretation. Because what he could have absolutely correctly said is that if we had used the old method of estimating the GDP, then the GDP growth would have been 4.5%. So 2.5% of growth is left unexplained.
Now to say that this 2.5% is simply overestimation is wrong. It is unexplained, yes. But the explanation could be that there is overestimation or it could be that there are other drivers of growth which are doing it.
And the two big other drivers of growth are improvements in productivity and improvements in product quality. You think about Maruti selling 800s and they switch their production to Dzire. The value of Dzire is two and a half times the value of the 800. That pushes up the GDP. And productivity is obvious. That is you produce the same quantity at lesser cost. These two can be very strong drivers of growth. In fact, in the developed countries, most growth comes from these two factors. It doesn’t come from volumes.
So if he had just said it is unexplained, it would have been alright. He has overreached by saying that this is an overestimation.
Radhika Pandey, NIPFP
An attempt to compare the GDP numbers from 2004-05 base year and 2011-12 base year is challenging. In contrast to the earlier episodes, the shift to the new 2011-12 base year was accompanied by comprehensive changes to the methodology and data sources. It is by now widely documented that there are issues with the 2011-12 base year. At the same time, it should be noted that the 2004-05 base year also had measurement issues. There was a discussion about overstatement of services sector in the 2004-05 series. Therefore to consider the 2004-05 base year series as sacrosanct as the paper by the Ex CEA considers on the ground that it had high correlations with some of the real sector variables like IIP, credit may not be appropriate.
Estimation of GDP through a bunch of regression specifications can be used to develop a coincident indicator that can serve as a proxy to assess the state of the economy. It may be too simplistic to present an estimate of GDP through regression specification. There are a number of databases used by the statistical agency which are outside the researchers’ ambit. One of the plausible reasons for over-estimation the paper attributes to is the move to value-based approach from volume-based. The move was in line with international best practises. So while there are challenges in estimation, those need to be addressed rather than attributing the overestimation to the shift to the new methodology. There is a need for a comprehensive debate about sources, methods and deflators used in the 2011-12 series. In parallel, there is merit in developing techniques to assess the state of the economy independent of GDP