GDP Numbers: What’s Wrong With How India Measures Manufacturing Output Data

With the release of the GDP figures for the quarter ending September, the government machinery has been in full swing advancing the narrative that economic growth is indeed back on track.

However, sorely missing from these narratives is the inconvenient factoid on the currently dismal state of consumption, especially rural consumption, in the country.

PFCE or private final consumption expenditure, an indicator of the consumption and consumer demand from households, is still far off from the pre-pandemic levels. The sentiment emanating from the ground doesn’t seem to forebode optimism either. The Consumer Confidence Survey conducted by the Reserve Bank of India for the month of November indicates that the outlook vis-à-vis the economic situation, employment, income, spending and inflation continues to be pessimistic and in the negative territory.

In such a scenario, a claim of an 8.4% real GDP growth rate has little relevance even as rural India battles plummeting wage levels, depleted incomes and widespread unemployment.

Earlier, The Wire highlighted the flaw embedded in the GDP computation methodology which uses data from the organised sector as a proxy for the unorganised sector. A consequence of this computation method is an inaccurate and inflated estimation of the strength of the unorganised sector as it is mistakenly equated with the organised sector. The imputed parity between the organised and the unorganised sector is naturally flawed considering that the domestic unorganised sector has been through a disastrous phase since 2014 as economic brinkmanship in the forms of demonetisation, the Goods and Service Tax (GST) reform, and then the lockdown have irreparably dented the micro, small and medium enterprises (MSME) segment in India.

Meanwhile, there are other grounds on which the veracity of the GDP estimates has been assailed apart from the erroneous usage of organised sector data as proxy for the unorganised sector. A section of economists, statisticians and academics armed with statistical evidence has been interrogating the integrity of national accounts under the new 2011-12 series, only for their objections to fall on government’s deaf ears.

To believe or not to believe?

A cloud hangs over the manufacturing sector output data released by the government. The total contribution of the private corporate sector (PCS)’s contribution to the GDP and the share of states and industries in PCS growth has also been called into question leading to a situation where the credibility of the GDP/GVA figures as a macroeconomic indicator has taken a considerable toll.

R. Nagaraj, currently a visiting professor at the Centre for Development Studies in Thiruvananthapuram, is part of a small spirited group that has been calling on the Ministry of Statistics and Programme Implementation to heed to the inconsistencies in the national accounts statistics.

He told The Wire that in his interaction with professionals from the corporate sector as well as those from the private equity segment, he keeps coming across individuals who avowedly proclaim that they have stopped looking at GDP numbers as reliable.

“They tell me that they look at high-frequency indicators or World Bank’s night lights data or some other indicators. Everyone has worked around the problem in their own private way but nobody is placing their trust in the government numbers. The doubts and the queries we raised persist.”

The Wire also spoke to Sebastian Morris, senior professor at the Goa Institute of Management and former professor at the Indian Institute of Management, Ahmedabad.

“There could be an error of as much as 1-1.5% in terms of the growth rate. It can’t be more than that. In a period when the GDP fluctuates, that is when you have a problem (in the correct estimation of GDP data). One can see that there is an increasing difference between the quick estimates of the GDP and the revised estimates, which is indicative of a bias.”

In July 2021, Business Standard carried out an analysis that seems to buttress the point made by Morris. In India, the GDP results declaration cycle runs for three years, unlike the US where three iterations of GDP are wrapped up in a space of three months. While the first advanced estimates are released in January, the second advanced estimates are released in February and then the provisional estimates in May. But this isn’t where the cycle ends. This is followed by three revised estimates every year. Effectively, the accurate data for one FY comes through after a lag of three years.

Coming back to the point made by Morris, Business Standard found that in FY18, FY19 and FY20, the percentage difference between the first advanced estimates and the final revised estimates has been as large as 13.2%. The analysis also found that during periods of low growth, the government inflated its value-added estimates and revised them downwards subsequently.

A labourer sleeps on sacks as traffic moves past him in a wholesale market in the old quarters of Delhi, January 7, 2020. (Photo by: Anushree Fadnavis Reuters)

The roots of the crisis

The crisis first took shape in 2015 when the Central Statistical Office, now National Statistical Office, presented a new National Account Statistics series with 2011-12 as the base year. The new series replaced the old base year of 2004-05. The new series, modelled on the lines of an international template, namely the United Nations System of National Accounts, adopted Gross Value Added (GVA) at basic prices. Adding net taxes on products to GVA at basic prices provides one with the GDP figures under the new methodology.

At the heart of the GDP estimation debate is the replacement of the Annual Survey of Industries database with the MCA21 database for estimating the contribution of the private corporate sector.

Red flags were naturally raised – not just by economists but amateurish observers as well – once the first revised estimates for 2016-17 were released which stated that the economy grew at a phenomenal 8.2% rate during demonetisation. The situation was complicated further by the release of two back series that presented a contradictory picture of the economy under the UPA. The first back series released in August 2018 by a committee set up under the National Statistical Commission highlighted that the UPA-I and UPA-II years were marked with higher economic growth rates compared to NDA years. This obviously put the BJP in an embarrassing spot, effectively ceding to the Congress gloating rights for piloting India towards a stronger economy. The government, on the defensive, took down the report and reuploaded it with a caveat that the figures were not final and should not be cited anywhere.

As a face-saving act, in November 2018, a new back series was released by the CSO which pulled down the average growth rate under UPA years to 6.7% while the growth rate under NDA was bumped up at 7.3%. The release of the two back series – so starkly contradictory to each other – only served to amplify the distrust of the GDP figures.

By May 2019, troubles grew for the CSO when the National Sample Survey Office in its 74th round conducted a survey of service sector enterprises mirroring the Annual Survey of Industries. This survey report posed new questions on the reliability of the MCA21 database used in the estimation of the GVA. The survey’s finding revealed that out of a universe of 35,000 ‘active’ companies sampled by the NSS, 45.5% did not respond to the survey, while 21.3% of the companies were misclassified and 24.2% were closed or non-traceable.

Naturally, critics argued that many of these companies were ghost or shell companies that were merely vessels for channelling profits or inactive. As per the 2019 MCA annual report, there are over 11.34 lakh ‘active’ companies – essentially companies that have filed financial returns once in the last three years. However, the data used for GVA estimation is drawn from a small set of only 3 lakh companies. For the new series, the CSO uses a blow-up or a scale-up methodology to calculate the GVA of non-filing companies. This is problematic given that in the absence of financial returns of a sizeable proportion of companies, any estimation drawn after blowing-up will convey a distorted picture of private sector growth and consequently, aggregate growth.

“The new series has several problems and inconsistencies and they were pointed out by many scholars. Not just that there were alternate estimates prepared by people like Arvind Subramanian and  Sebastian Morris, who tried to put a number to the extent of over-estimation. The last serious discussion on this critical issue took place in 2019 at the National Council of Applied Economic Research, and after that, there has been silence from the government’s side. Unfortunately, the government has stonewalled all our criticism. It is refusing to respond to the criticisms except for the fact that it has been rhetorically dismissing the criticisms instead of engaging with them.” Nagaraj told The Wire.

Opacity over  manufacturing sector output

In a research paper published in the Economic and Political Weekly, Nagaraj cites evidence to show that the new series overestimates average annual growth rates for the manufacturing sector (at constant prices). Nagaraj takes two periods; the first one from 2004-05 to 2011-12 and the second one from 2011-12 to 2018-19.

In the first period, the growth rate for the manufacturing sector clocks in at 9% by the IIP, 10.7% under the ASI and 9.5% as per the old series. Meanwhile, in the second period, the growth rate for the same sector comes to 3.8% by IIP; 5.2% by ASI and 7.4% under the new series.

There are greater divergences under the new series. Nagaraj blames the shifting of computation of GDP estimates from ASI to the MCA21 database and the changed methodology for the inflated estimates.

“We are trying to recover to where we were before the pandemic. Things are not right on the ground. Compare the absolute GDP numbers of 2019-2020 and 2020-2021. There is a negative growth rate in 2021 of 7.3% going by official numbers. If you want to get back to the same pre-pandemic level, you will need a growth rate of anywhere between 8.5-9.5% in 2021-2022,” he said.

“When the government talks about getting back to the growth rate of 8.5-9.5%, it means that if that rate is attained, it will get back to the pre-pandemic level. We have two years of complete loss. One year has been of negative growth and one year has been of recovery. To me, it isn’t growth. We are recovering to where we were before the pandemic, and that’s it,” Nagaraj added.