Debate over India’s GDP Numbers

The Central Statistics Office (CSO) introduced a new series of National Account Statistics in 2015. The important changes made in the new series included the change in base year from 2004-05 to 2011-12, employed a new methodology to estimate India’s gross domestic product (GDP) and used new data sets to arrive at the GDP.

Why did the GDP methodology get into controversy?

Even though revising base years, improving methodologies and opting for better databases are part of normal practice in national income accounting, the debate on the authenticity intensified in 2018 when the statistical establishment released two back-series (that is, recalibrating the GDP data for past years based on the new methodology).

The first back-series presented by the National Statistical Commission (NSC) in July 2018 found that the average economic growth between 2005-06 and 2011-12 was 8.6% instead of the 8.3% according to the old series. The second back-series, calculated by CSO and published in November 2018, found this average to be just 7%.

The accusation that the new method overestimates GDP received a boost after Arvind Subramanian, India’s Chief Economic Adviser between 2014 and 2018, argued that the new series overestimated.

What did Arvind Subramanian Say?

In a research paper “India’s GDP Mis-estimation: Likelihood, Magnitudes, Mechanisms, and Implications” former chief economic advisor Arvind Subramanian has claimed that India’s economy grew at a much slower pace between 2011-12 and 2016-17 (at about 4.5%) than the official estimate of 7 per cent.

Observations made in the Research Paper

• India’s GDP growth was overstated by about 2.5 percentage points per year in the post-2011 period.
• The former CEC arrived at the conclusion by the means of an index created by based on the data from 17 indicators that are strongly correlated with GDP growth and compared it with the official GDP data
• The indicators like electricity consumption, vehicle sales, airline passenger traffic, foreign tourist arrivals and manufacturing were chosen since they are produced independently of the CSO, which compiles the GDP data.
• His observations state that the indicators moved closely with the GDP between 2001-02 and 2011-12 and showed a wide divergence from 2011-12 to 2016-17.
• He also states that India was compared with 71 other countries on how a set of indicators (credit, exports, imports and electricity) is related to GDP growth where India proved to be an outlier here.

Counter Arguments by the EAC

The Prime Minister’s Economic Advisory Council (EAC) refuted the assertion by the former chief economic advisor (CEA) Arvind Subramanian that India’s GDP was overestimated between FY12 and FY17. The EAC countered by saying that:

• Arvind Subramanian’s analysis of economic growth is incorrect as it ignores the indicators related to the bulk of economic activity, such as those measuring the agriculture, services, and informal sectors.
• There has been a hurried attempt to conclude India’s complex economy and its evolution, relying more on private agencies such as Centre for Monitoring Indian Economy (CMIE) while raising doubts about Central Statistics Office (CSO) data.
• India’s GDP estimation methodology was on par with its global standing as a major and responsible economy. Arvind Subramanian had used 17 high-frequency indicators but ignored the representation of the services sector which contributes about 60% to GDP and the agriculture sector (18%) in the analysis.
• Further, the tax data was overlooked in the assertion of Arvind Subramanian by reasoning that major changes in direct and indirect taxes in the post-2011 period would render the tax to GDP relationship different and unstable and hence make the indicators unreliable proxies for GDP growth.

EAC white paper concludes that attempting to approximate GDP of such a country based on some correlations and four variables using simplistic econometric techniques and challenging the existing edifice of data collection is not only demoralising to that dedicated personnel but also technically inappropriate.

Clarification over Back Series Data

NR Bhanumurthy, the chair of the NSC sub-committee that presented the first back-series countered some of Subramanian’s fundamental assertions with the following observations:

• The nominal GDP growth rate, which is the only observable variable, has not changed under the old and new series.
• There was no consolidated CPI before 2011. Hence the argument that the gap between CPI and GDP deflator was low between 2002 and 2011, and wide between 2011 and 2016, is unfounded.
• If one applies the Production Shift approach wherein new activities (that are added in the new series) are given progressively more weight (or less weight) as one move ahead in time (or moves back in time), the two series do not show any discordant breaks either for GDP or for gross value added.

NR Bhanumurthy justifying new methodology concludes that given the wide differences between the two data methodologies and data sets, the whole debate is more of like “apples to oranges” comparison. One can predict the direction of GDP growth using leading indicators such as car sales, but one cannot ‘estimate’ GDP growth rate for a year gone by based on such indicators.