Former CEA Arvind Subramanian says GDP growth inflated, govt disagrees

In a research paper that analyses data on 17 different economic indicators in the real sector, former chief economic advisor of India Arvind Subramanian has said that India’s real gross domestic product (GDP) growth “was more likely to have been between 3.5-5.5 per cent” in the 2011-2016 period, rather than 6.9 per cent as estimated by the National Statistical Office (NSO). He said that the new GDP series overestimated real GDP growth by 2.5 percentage points.

This has reignited the debate on the accuracy of India’s GDP numbers in particular, and has once again highlighted the inefficiencies in the estimation of national accounts in general.

But the government stood by its own numbers, and clarified that the estimation adheres to global norms and is an output of a complex exercise.

“It is important to note that a comparison of the old and new series are not amenable to simplistic macro-econometric modelling,” the statistics ministry said in a press release. 

The new series of national accounts with base year 2011-12 was formulated in 2015, and is currently used to compute quarterly and annual GDP, and its methodology has been under scrutiny from economists and statisticians since then. 

"With structural changes taking place in the economy, it is necessary to revise the base year of macroeconomic indicators...to ensure that indicators remain relevant and reflect the structural changes more realistically. Such revisions not only use latest data from censuses and surveys, they also incorporate information from administrative data that have become more robust over time" Mospi said. 

Subramanian analysed output volume-based data, such as the index of industrial production, exports and imports, footfall of tourists, auto sales, airline passenger traffic, and electricity and petroleum consumption among others, to make his claim.

He concluded his paper by saying, “the entire methodology and implementation for GDP estimation must be revisited by an independent task force”.

This came as a surprise to some bureaucrats who have worked in the finance ministry in the period when Subramanian was the CEA. A retired official said that Subramanian never raised the issue of inconsistencies in the GDP data in the Economic Survey or Budget meetings.

Economists that Business Standard spoke to, however, raised a few crucial points on the research paper.

Former chief statistician Pronab Sen said that the paper estimates the difference between the estimate of GDP using volume-based indicators (as Subramanian did), and that estimate calculated using value-based indicators such as the Ministry of Corporate Affairs (MCA) data (as the new methodology does). This very difference represents “productivity” or value added per worker in the economy, he said.

“The paper’s finding means that there has been an equivalent improvement in productivity,” he told Business Standard. “But this would also mean that the older series underestimated growth in the years prior to 2011-12,” he added, since it was more robustly based on volume indicators.

R Nagaraj, professor of economics at the Mumbai-based Indira Gandhi Institute of Development Research, who was a member of the sub-committee on private corporate sector set up by the advisory committee on National Accounts Statistics, reiterated that MCA data estimates the value added by the private corporate sector (PCS) inaccurately.

“This is the root cause of enlargement of its share in GDP and its faster growth,” he told Business Standard. PCS forms a substantial part of the manufacturing sector in the economy.

He said that there is no clear estimate on how many of the "active" companies from the MCA data considered for GDP estimation are really producing goods and services on a regular basis. The recent technical report from the NSO on the services sector underlined this, he added.

He also pointed out that the sample of companies who file their financials regularly with the MCA is a “self-selected sample” and not a “scientifically drawn random sample.”

“We have a sampling procedure of companies in which the sample size, sampling fraction and universe from which the sample is drawn are all indeterminate,” Nagaraj said.

Sen highlighted that the MCA data does not help measure contribution of various sectors from manufacturing and services to the economy accurately. “We do not know how many of the companies still produce those goods or supply those services under which their registration in the MCA happened. There is no way to factor in the changes in companies’ nature over time in the current methodology,” he told Business Standard.

Subramanian, in his paper, showed that IIP, which is based on volume of production/output of companies in the formal sector, is used as a proxy to compute the contribution of the informal sector to the GDP. Higher growth in formal sector represented by IIP “inflated” growth in the informal sector in recent years, especially after demonetisation and goods and services tax implementation, he wrote.

In addition, he said that lower oil prices also pushed up growth figures. Fall in oil prices reduced the input costs for many industries in the manufacturing sector. The use of a single deflator to calculate real value added from nominal value added under this scenario “artificially inflate growth figures by significant amounts”, he wrote.