How not to revise GDP base year to bring back credibility of data

Also read: India GDP to grow 6.5 pc in 2026, but there's urgent need for jobs: World Bank

Abandoned options

Few know that the Centre had planned to introduce a new GDP series with 2017-18 as the base, but abandoned it because economists questioned the choice.

Pronab Sen, chief of the Standing Committee on Statistics (SCoS) during 2019-24, had called 2017-18 (FY18) “abnormal” and “not the right year for the new base” because of the twin shock of demonetisation (November 2016) and GST (July 2017). Both severely damaged the informal sector.

Earlier, the Centre had abandoned rebasing GDP on 2009-10, which had seen the Great Recession and one of the worst droughts to hit India.

Fresh questions

Given that the National Statistical Commission (NSC) prefers rebasing GDP every five years (to capture rapid economic changes) while the current move comes 11 years later (between 2011-12 and 2022-23), this may not face stringent objections.

Nonetheless, questions would be raised for three reasons:

(a) 2022-23 (FY23) counts as a pandemic recovery fiscal

(b) GDP growth for FY23 at 7.6 per cent is hemmed in by two very high growths: 9.7 per cent in FY22 and 9.2 per cent in FY24

(c) absence of Census 2021.

Is FY23 right GDP base year?

The FY23 growth of 7.2 per cent appears an outlier. In FY25, GDP growth was 6.5 per cent, which is better aligned with the long-term average growth of 6.6 per cent during pre-pandemic FY13-FY20 and 6.1 per cent during the entire 2011-12 series (FY13-FY25, including FY21 and FY22).

For FY26, the IMF and World Bank project growth of 6.2 and 6.3 per cent, respectively.

The absence of Census 2021 matters because one of the critical inputs for base-year revision is the Household Consumption Expenditure Survey (HCES) of 2022-23, which is not based on truly representative samples of the population drawn from the old Census 2011. The HCES 2022-23 will also affect the CPI of 2024.

More important, however, is to address the credibility issues with the 2011-12 GDP series.

Also read: Under Modi Raj, economy has suffered structural retrogression, lost dynamism

Charges of growth overestimation

When the current GDP with 2011-12 base was introduced in January 2015, India’s growth (i) surpassed China’s to become the fastest growing major economy and (ii) became the fifth largest economy in the world.

(In Purchasing Power Parity, or PPP, terms, India became the third largest economy in 2009 and continues to be so for 16 years, but in per capita GDP, it ranks 144th among 196 countries in 2024, making Indians one of the poorest in the world, per the IMF.)

In 2019, Arvind Subramanian dropped two bombs. He was Chief Economic Advisor (CEA) for four years between 2014 and 2018 and the 2011-12 GDP series was introduced when he was in office. His first bomb came in June 2019.

'Bomb' number 1

Subramanian wrote a paper at the Harvard University, which said: “Official estimates place annual average GDP growth between 2011-12 and 2016-17 at about 7 per cent. We estimate that actual growth may have been about 4.5 per cent…”

“Growth overestimation varies from 2.5 percentage points per year (without electricity) to 3.7 percentage points per year (with electricity)," he added.

He had used 17 high-frequency indicators (electricity consumption, vehicle sales, passenger and freight traffic, IIP, petroleum consumption, cement, steel, credit outflow, trade data, etc.) for his estimates.

The Prime Minister’s Economic Advisory Council (PM-EAC) promptly rejected it, raising questions about the use of high-frequency indicators for alternate GDP growth estimate, not the original datasets used for GDP estimates. This was an unreasonable attack because both the Finance Ministry and the RBI (bulletin) use these very indicators to analyse the health of the economy every single month.

'Bomb' number 2

The second bomb came in July 2019 in response to this criticism.

He wrote another paper: “Since the underlying data (for GDP estimation) are not available publicly, nobody outside the CSO can ‘estimate GDP’.” The CSO is short for Central Statistical Office, which is the right authority to prepare GDP estimates. This was shocking indeed.

PC Mohanan, former acting chairman of the NSC, reacted by saying that such an opaque use of data for GDP base year revision had never happened in India earlier, as all datasets used were made public.

Mohanan had resigned months earlier over alleged data suppressions. The HCES of 2017-18 was junked for showing rising poverty and the Periodic Labour Force Survey (PLFS) of 2017-18 was withheld until after the 2019 General Elections for showing a record high unemployment rate, he alleged.

But both reports had found their way into newspapers. That added fuel to the public debate over the credibility of official data then raging.

Also read: Why manufacturing continues to fail even with unprecedented incentives

In March 2019, 108 economists and sociologists from across the world issued a public appeal to the Indian government not to “suppress uncomfortable data” and “restore access and integrity to public statistics”.

‘Politicisation of data’

The provocation was as much the suppression of the HCES and PLFS of 2017-18 as two contradictory back series of 2011-12 GDP and retrospective upward revision in GDP growth that “did not square with related macro-aggregates”.

The two back series GDP data were released in 2018 – one in July (NSC’s) and the other in November (CSO’s). The CSO one was adopted (NSC’s junked), which had cut down GDP growth during the earlier UPA years (more of it later).

The charge of “overestimation” of GDP growth continues till today.

Subramanian has raised it multiple times since then over subsequent GDP growth numbers – most notably in September 2023, March 2024, and July 2024, using dissonance with other key indicators like inflation, consumption, jobs, etc. to burst the myth of a “gangbuster growth”.

Many other economists, like Arun Kumar and Ashoka Mody, have also said the same. In October 2024, Sen confirmed what was long suspected: politicisation of data (“Yes. If the rumours about data being sent to NITI Aayog or the Cabinet Secretariat for vetting before release are true, then that is politicising data.”).

Questionable data

Many other flaws with the 2011-12 GDP series have been called out over the years. Here are a few key ones:

Unknown, untested and unaudited MCA-21 data: For the first time, as yet unknown (not in public domain) and self-populated by companies (untested and unaudited) database of the Ministry of Corporate Affairs – known as MCA-21 database – was used to estimate manufacturing and services contributions to GDP – the two sectors accounting for 75.9 per cent of GDP in FY12.

Also read: GDP growth likely to slow down due to global uncertainty and tariff war

In 2019, when the NSSO verified this MCA-21 database for the services sector alone (not manufacturing), it said “about 45 per cent of MCA units” listed “were found to be out-of-survey/casualty” (didn’t exist, didn’t operate or engaged in unrelated activities).

Dramatic rise in manufacturing GVA: Manufacturing Gross Value Added (GVA)’s share jumped from 14.7 per cent to 18.1 per cent in FY12 as some components of services were added to it, taking it closer to the ‘Make in India’ goal of 25 per cent without a sweat.

Dubious back series data: The back series data came four years late (in November 2018). It is usually released along with the base revision.

Back series data

More importantly, perhaps for the first time in India, three back series data were prepared to go with the 2011-12 base year revision: First one by the Central Statistical Organisation (CSO) in January 2015 (not made public), second by the NSC in July 2018 and the third by the CSO in November 2018 (latter two made public).

The first two were junked by the NITI Aayog (politicisation of data) – even when the Aayog has neither locus standi nor domain expertise; it is CSO’s task with the NSC’s supervision – for showing higher growth rates in the UPA years (Manmohan Singh government).

Also read: Why India is facing an investment winter despite booming GDP

That this was the case with the first one readied by the CSO in January 2015 (when the 2011-12 GDP series was introduced) came to public notice only when then NSC chairman Sen disclosed in December 2018 to a national daily: “NITI Aayog took just one look at the growth rates which were going higher after 2004-05 and said, ‘We can’t allow it’.” Arvind Panagariya was then the Aayog’s Vice Chairman (now chairman of the 16th Finance Commission).

Revision of UPA numbers

The same fate befell the second back series (NSC’s) of July 2018 – by then Rajiv Kumar had replaced Panagariya. His objections were particularly about to two double-digit growths in UPA years (10.2 per cent in FY08 and 10.8 per cent in FY11).

The third and final one, released in November 2018, that met with Rajiv Kumar’s approval, had brought down those GDP growth numbers to 7.7 per cent (FY08) and 8.5 per cent (FY11), respectively. The average GDP growth of 8.24 per cent during FY06-FY14 (UPA years, NSC estimates) was also lowered to 6.9 per cent.

There is yet another issue with the back series.

The CSO’s back series of November 2018 went back to 2004-05 – not the usual up to 1951-52.

Even tracing the back series to 2004-05 is questionable because one of the key inputs, the MCA-21 database, came into existence in 2007-08 and stabilised after 2011-12 – as the NSC’s report of July 2018 had revealed.

Also read: Bold reforms, enhanced capacities key to India's growth story: FM

Extension of back series

In yet another shocking development, the back series was extended to 1951-52 sometime in 2021 (without a date stamp) quietly and released without explaining how this miracle happened in absence the MCA-21 database going back that far in time.

Retrospective revisions in GDP: Retrospective revision in GDP are too frequent to list here. The last one was in February 2025 when GDP growth went up from 7.1 per cent to 7.6 per cent for FY23 and from 8.2 per cent to 9.2 per cent for FY24). This data fiddling is so pervasive that former RBI Governor YV Reddy once quipped (about the demonetisation data) in 2017: “In India not only the future is uncertain, even the past is uncertain. So, they keep revising the data.”

Poor deflators: Another long-standing issue with GDP estimate is the use of deflators (price indices like CPI and WPI). Developed economies switched to producer price index (PPI) in 1970s to capture price changes for producers (as CPI is for consumers); India was supposed to adopt PPI to replace WPI since the latter leaves out services sector which has the maximum share (55 per cent of GVA in FY25). Nothing more has been heard of it.

Extremely large ‘discrepancies’: Yet another concern is unusually large “discrepancies” added on the expenditure side GDP to match with production side data (GVA). From less than 1 per cent during FY12-FY16, it skyrocketed to (+) 3 per cent, (-) 3 and (-) 3.2 per cent of GDP in FY18, FY23, and FY24, respectively.

Also read: India to stay fastest-growing economy with FY26 GDP at 6.2–6.7 pc: International agencies

What Centre must keep in mind

Explaining the implication of one such high ‘discrepancies’ in Q1 FY24 data (discrepancies of (+) 2.8 per cent of GDP), Mody wrote : “It shows that while income from production increased at an annual 7.8 per cent rate in April-June, expenditure rose by only 1.4 per cent. Both measures clearly have many errors. The NSO nonetheless treats income as the right one and assumes (as implied by its “discrepancy” note) that expenditure must be identical to income earned. This is an obvious violation of international best practice.”

All the above details tell one what the Centre must avoid this time and what it must incorporate while revising the GDP base year to 2022-23 to bring back credibility.