Behind the government's massive data revision

India’s macroeconomic statistics have come under intense scrutiny in recent years. Matters reached a head a few months ago when the International Monetary Fund’s (IMF’s) Article IV Statement gave the national accounts a ‘C’ rating for the second consecutive year. The price data got a ‘B’ rating.

This comes even as the country’s data systems are undergoing a massive overhaul after a gap of a decade. Estimates of retail inflation, national accounts and factory output will be based on a new series and base years with an updated basket of goods, weights and data sources.

These will be released gradually, starting with the revised consumer price index (CPI), which will be published on February 12. That will be followed by release of national account estimates or second advance estimates of gross domestic product (GDP) for 2025-26 on February 27 on a new base year. The index of industrial production (IIP) is also being updated; the new series is likely to be released on May 28.

This mammoth exercise began a couple of years ago and required many surveys to get fresh data. Expert groups were formed to work on the methodology and data sources. The Ministry of Statistics and Programme Implementation (MoSPI) also held several rounds of discussions with a team from the World Bank, experts from the IMF and the United Nations on improving and compiling the methodology in the new series, says N.K. Santoshi, Director General, Central Statistics Office.

Saurabh Garg, Secretary of MoSPI, points out that this is far from a routine statistical update. “It requires a comprehensive review and refinement of hundreds of datasets, incorporation of new and emerging data sources, and meticulous methodological work to ensure the revised figures truly reflect the current structure of the economy,” he says.

After all, better data will aid government policymaking and private sector decision-making. As Reserve Bank of India Governor Sanjay Malhotra noted recently, the exercise would help in making more calibrated policies and aid the central bank in sustaining price stability and improving economic growth.

The concerns

The MoSPI is hopeful that the updated indices would take care of concerns around the old base year, double deflation, as well as high statistical discrepancies, issues that have cropped up time and again in GDP estimates. The IMF has also flagged these issues. Double deflation is a statistical method to calculate the real value added by removing intermediate inputs at constant prices.

“National accounts data are available at adequate frequency and provide broadly adequate granularity. However, some methodological weaknesses somewhat hamper surveillance and warrant an overall sectoral rating for the national accounts of C,” the IMF had noted. The methodological issues it flagged related to coverage, including an outdated base year; use of wholesale price index (WPI) as data source for deflators in the absence of a producer price index; and excessive use of single deflation, which may introduce cyclical biases.

Besides, the IMF said that at times there were sizable discrepancies between production and expenditure approaches for measuring GDP. To address that, the coverage of expenditure approach and informal sector would need to be enhanced.

On granularity, the IMF noted that a further breakdown of Gross Fixed Capital Formation (GFCF) by the institutional sector (published with a significant lag) and disaggregation of quarterly production and expenditure approach estimates would allow for a more detailed analysis of economic trends. GFCF is a measure of investments.

“On price statistics, CPI is available at adequate frequency and timeliness. The rating granted to the coverage reflects outdated CPI base year, items basket, and weights (2011/12), implying that the CPI basket likely fails to accurately represent current spending habits,” the IMF noted.

Several economists have also flagged concerns that growth may have been overstated in recent quarters due to issues in compilation of national accounts and use of a single deflator. The divergences between GDP and gross valued added and measurement of the informal economy have also been pointed out as problems. For instance, economists highlighted discrepancies of 3.3% between GDP calculated by production and expenditure approaches in the second quarter of FY26, when the economy grew at a higher-than-anticipated 8.2% year-on-year. Discrepancies were 2.3% in the first quarter of the fiscal and have been flagged by economists time and again.

When the Q2FY26 GDP estimates were released at the end of November, a report by HSBC Global Investment Research noted that the Central Statistics Office’s practice of using single deflation, instead of double, in calculating GVA data leads to an overestimation of growth in periods when commodity prices are falling. It also noted that India’s services sector GDP deflator aligns more with goods-oriented WPI inflation than with CPI services inflation, adding that this is particularly a problem when manufacturing inflation is falling because of softer commodity prices. This ends up deflating services inadequately, leading to exaggerated real growth. As per its calculations, CPI services inflation for the quarter was 3.4%, which was higher than the 1.9% GDP deflator used for services. “If we switch to the former, real services growth falls by 1.5 percentage points, and GDP growth falls by about 0.8 percentage points,” it said.

Nomura, too, noted that July-September 2025 was the second consecutive quarter when the market not only underestimated the quantum of GDP growth, but also its direction, forecasting a slowdown instead of a surge. “Really low GDP deflators have added to the unpredictability of the series. Consequently, there seems to be a divergence between GDP statistics and evidence from high-frequency activity indicators, making it difficult to gauge whether the GDP data are indeed reliably reflecting the state of the economy,” it said.

These are not stray concerns but reflect oft-repeated problems, although Chief Economic Adviser V. Anantha Nageswaran recently refuted worries about GDP overestimation, saying such questions were not asked when growth was on the lower side.

However, economists point out that the nature of the economy has changed over the past decade and the official data is unable to capture that fully.

Ranen Banerjee, Partner and Leader, Economic Advisory Services Government Sector Leader at PwC India, says the current base of the GDP series is 15 years old. Since then, the structure of the economy has changed with the emergence of the gig sector and e-commerce, which employ many people. “The nature of logistics has changed. Many women’s self-help groups and farmer-producer organisations have also come up during the intervening period. The data does not capture this. There have also been structural breaks in the series due to the introduction of the goods and services tax (GST) and the outbreak of the Covid-19 pandemic,” he says.

Similarly, consumption patterns have changed significantly. For instance, monthly expenditure on food items for Indian households has dipped below 50% for the first time. Besides, per capita income has almost doubled since the CPI basket was last updated. “It is very important to reflect the correct consumption pattern in the CPI to accurately reflect real inflation for consumers,” says Banerjee.

Further, the share of the informal economy was estimated at about 45% in 2011-12, which is likely to have changed owing to progressive digitalisation and formalisation of the economy. “It is important to better understand the growth of this segment and not rely solely on extrapolations given several alternative datasets have become available,” he says.

N.R. Bhanumurthy, Director of the Madras School of Economics, says the GDP base revision was delayed for far too long, primarily because the Household Consumption Expenditure Survey (HCES) and other survey data were not available. The norm is to revise it every five years. “Even in the CPI revision, there has been so much discussion around the consumption basket and issues with it,” he says.

The revision of the base year will take care of questions over the divergence betweeen GDP and GVA growth rates as well as the high discrepancies that arise.

The revisions were last carried out in 2014-15 when 2011-12 was made the base year. It is being updated to 2022-23 in the new series for GDP and IIP and to 2024 for CPI. Apart from this, the sectors and the basket of items that are used for computation are also being updated with more data sources becoming available. These are expected to capture new economic activities such as gig economy and e-commerce as well as consumption and manufacturing patterns, providing a more accurate picture of the economy.

Additionally, the Ministry of Commerce and Industry is working on updating the WPI on new data sources, including the Producer Price Index and the Services Price Index.

The Great Reset

At present, more than 300 datasets are used for compiling India’s GDP estimates. These include administrative data, survey results, financial accounts, and sectoral data, says Garg. During a base revision, each of these datasets is reviewed, validated, and updated, he says, adding that new data sources are explored and incorporated as well.

The current GDP base year revision will incorporate the latest surveys such as the Annual Survey of Unincorporated Sector Enterprises (ASUSE), Periodic Labour Force Survey, Public Finance Management System as well as HCES and administrative data from GST, Ministry of Corporate Affairs, and E-Vahan. It will also reflect methodological improvements in estimation and deflation.

MoSPI also plans to use single extrapolation or double deflation, wherever feasible, to arrive at constant price estimates; use deflators at a more disaggregated level and update rates and ratios from different studies covering fodder, fisheries, milk and its products, construction, and road transport. It is hoping to do away with statistical discrepancies between production and expenditure estimates by publishing supply and use tables (SUTs).

The revision of the CPI base year to 2024 depended heavily on the HCES. The upcoming revision is expected to increase the number of items in the CPI basket and update weight structures.

Importantly, apart from data from physical outlets, prices will also be taken from e-commerce platforms in 12 cities with a population of over 2.5 million as per the Census of 2011. The MoSPI is also coordinating with the railway ministry for rail fares, petroleum and natural gas ministry for fuel prices and the department of posts for postal tariffs. Additionally, it plans to compile price data for airfare, telecom services and OTT platforms from online sources.

The revised IIP series will also be aligned with the national accounts framework, improving coherence between short-term industrial indicators and GDP estimates. At present, the IIP is compiled using production data from 14 sources, covering 407 items or item groups across sectors. The item basket, which is derived from the Annual Survey of Industries, is being updated. Outdated products will be removed, while new products will be included from emerging industrial innovations.

The revised IIP will also provide more granular data on sectors. For instance, the mining index will now cover fuel minerals, metallic and non-metallic minerals while electricity production will include both renewable and conventional sources.

It will also have a substitution mechanism to take into account any factories that may face operational disruptions, permanent closures or changes in production line. Interestingly, the MoSPI is also considering chain-linked indices, which would enable the IIP to better capture structural shifts, emerging industries and changes in production patterns by updating weights more frequently.

Way Forward

Former chief statistician T.C.A. Anant, who oversaw the last base revision, says the MoSPI is trying to incorporate as many new developments and data into the new GDP as possible and has addressed many existing problems.

However, he feels more could have been done, such as greater use of GST data. “But it is good work in progress, and one can never achieve perfection in any of these exercises. They should also plan for the next data revision exercise within the next five years to ensure that such a long delay does not recur,” he says.

While statistical discrepancies can be eliminated by publishing SUTs, which were earlier done after a slight gap but will now be more prompt, he says it would have been better to work harder on increasing the frequency of expenditure side data such as the HCES.

Another development in this space is the recent adoption by the United Nations Statistical Commission of the System of National Accounts 2025 (2025 SNA). Countries are likely to transition to this by 2030. The MoSPI has started work on different aspects of environmental accounts, forest, land, soil and ocean accounts, and has prepared a five-year road map so that by the time the 2025 SNA is implemented, the building blocks are already in place.

It also plans to undertake more frequent revisions of base years and conduct the HCES every three years.

More data sources will also be available. While the PLFS is now undertaken every month, the ASUSE provides quarterly data on the informal sector. An annual survey of service sector enterprises is also being conducted, while the household income survey and the survey on migration will be conducted in 2026.

The decadal Census, being carried out now after a delay of five years, will add to this sudden surfeit of data. Carried out by the Registrar General, it will provide more granular data on India and its citizens and MoSPI plans to follow that up with the Economic Census in 2027 that will enumerate all establishments in the country.

India’s so called “data deficit” is likely to be bridged through these efforts. But the country cannot afford any let-up in the momentum of updating data estimates. Hopefully, the next revision will be undertaken well before 2035! 

@surabhi_prasad