By R. NAGARAJ

December 16, 2025

(R. Nagaraj is a Professor (retired) of the Indira Gandhi Institute of Development Research, Mumbai)

The Gross domestic product (GDP), or gross value added (GVA), is the sum of all the “unduplicated” value of goods and services produced in a country in a given period, usually measured over a year, or a quarter of a year.

“Unduplicated” or “final” means the value of goods and services net of intermediate inputs that are used in production, such as raw materials, electricity and services used to produce the final output.

Definition and Origins

GDP/GVA are market-based concepts, meaning they include all goods and services (hereafter only goods for short) bought and sold in the marketplace. Therefore, by definition, GDP/GVA excludes goods and services produced and consumed outside of markets, such as in households and non-market institutions. For instance, food cooked at home for family or services rendered to the elderly or infants at home (the care economy) are, by current definition, excluded from GDP/GVA estimation.

The concepts of GDP and , more generally, national accounts are underpinned by macroeconomic theory propounded by J.M. Keynes about 90 years ago. Economists such as Richard Stone and Simon Kuznets developed the empirical counterparts of macroeconomic theory. Empirical measures and yardsticks guiding the preparation of national accounts follow the templates standardised by the UN System of National Accounts (UNSNA) in the late 1940s, enabling the comparability of the size and structure of modern economies. The UNSNA is periodically revised to account for the changes in the global structures of production and its institutions. In 1949, under the chairmanship of PC Mahalanobis, India constituted the National Income Committee to estimate the official national income since independence.

Process of Estimation

In principle, estimating GDP/GVA is simple: the quantity of each good produced in, say, a year, multiplied by its prices, net of the value of all intermediate inputs, gives GDP. Though conceptually simple, it is an aggregation of innumerable commodities and services into a limited number of standardised sectors. These sectors are production sectors, such as agriculture, industry, etc., or institutional sectors such as the corporate sector or household sector. These estimations involve enormous computations. A considerable amount of the theory of probability and statistical practices are deployed to arrive at the aggregate GDP estimate. Hence, it is a statistical construct measured in currency value (Rs or $), unlike a physical measure like temperature or physical length, which are readily readable from an instrument.

‘Real’ GDP

Such an estimate would be GDP at “current prices” for the year, say “t”. Similarly, GDP estimates for years t+1, t+2 … can be replicated consistently following the same methods. From these, the annual growth or yearly percentage change can be easily computed. However, such a growth number would represent a combination of the actual or “real” increase in quantity of goods produced and the effect of price changes (inflation). Economists and governments are usually interested in knowing the changes in “real” GDP (output) as a measure of social welfare. Hence, to obtain the real GDP, the effect of price changes needs to be netted out. Such a netting out of price changes warrants that GDP be estimated for a “base-year”, and by netting out price changes for the subsequent years.

Put differently, to obtain a measure of “real” change in goods and services produced, any measure in a currency needs to be adjusted for price changes (value of the currency), since the purchasing power of currency varies with time as the prices of commodities/services keep changing. As is widely known, price levels change as demand and supply conditions change in a market or an economy. Moreover, in times of rapid technological progress and obsolescence, prices of many goods become cheaper – sometimes even in absolute terms (such as IT hardware or solar cells recently). Hence, to adjust for the purchasing power of a currency, GDP/GVA needs to be adjusted for the prices of underlying goods and services. When it is done, we get the measure of GDP/GVA in “real terms”, or, at “constant prices”. Base year selection

To obtain a constant price series, we need a “base-year”, against which price changes can be netted out to obtain the real changes in output. In principle, the base year should be a “normal” year, meaning output growth and inflation during the year should be the average of a several previous years. Years with an economic shock such as famine/drought or a health emergency (like the Covid pandemic) or hyperinflation are avoided. Why? An exceptional base year could distort growth or inflation in subsequent years.

Why revise the base-year?

As the different sectors of an economy expand or contract at different rates, production structures tend to change over time. Prices change at different rates; some goods become more expensive relative to others, hence the relative weights of goods and sectors in the aggregate measure of GDP also need to change. Thus, their “weights” in the economy also need to be adjusted in an aggregate economic measure such as GDP. These changes need to be factored into the price indices to obtain the “real” changes in the economy. Hence, the base year of the price index needs to be revised frequently to reflect the changes in “relative prices” and the economy’s production structure.

How frequently is the base year changed?

This depends on the nature and pace of change in an economy. For a rapidly growing economy with high inflation, the base year needs to be revised regularly and frequently. However, as mentioned earlier, the base-year revision is a statistically challenging task, requiring newer data sets for the base year and considerable technical resources. The UNSNA recommends “re-basing” every five years. However, for  pragmatic considerations, revision is roughly undertaken every ten years. For instance, India has rebased its national accounts eight times since 1948-49.

How is the base year revision done?

Simply put; to prepare the base year estimate of GDP and its sectoral estimates, quantitative information is needed for all aggregate economic variables for that year, such as output and prices.

As GDP/GVA is the sum of all goods produced in an economy, we need a summary measure of prices – called the price index. The index is prepared for a particular year, and changes in the value of the index are taken as the measure of aggregate price changes over the years. So GDP/GVA is prepared with respect to the base year. And changes in the GDP/GVA over the year, net of price changes, represent the measure of “real” goods and services produced during the year. Thus, GDP/GVA is prepared with respect to a base year. And it is used for a certain number of years. Hence, universally, whenever real GDP is mentioned, it is always referred to a particular “base-year”. For instance, in India, the ongoing GDP series is with the base year 2011-12. To illustrate, in the year 2023-24, India’s GDP was Rs. 3,01,22,956 crore in current prices, and Rs. 1,76,50,591crore at constant prices. Thismeans that the difference between the two numbers (Rs. 1,24,72,365 crore) represents the effect of prices.

 

A delay in base-year revision may lead to under-reporting/estimation of output of newer or emerging activities, and over-representation of older or declining activities. Aldo, ss relative prices change, the underlying price indices may become distorted. Hence, it may wrongly represent output growth.

Questions raised last time with the base-year  revision

The major decision in the last (2011-12) base-year revision was to change over to UNSNA 2008, replacing the templates of the earlier UNSNA 1994.

Following UNSNA 2008, the Central Statistics Office (CSO) – now renamed as the National Statistics Office (NSO) – decided to prepare national income estimates also by institutional categories such as the private corporate sector, household sector, and so on. To implement the decision, the CSO decided to use the recently available Ministry of Corporate Affairs (MCA) database to prepare the estimate for the private corporate sector (PCS), replacing the Reserve Bank of India’s  (modest) sample of companies with large paid-up capital. The MCA database also replaced the Annual Survey of Industries (ASI) for estimating manufacturing sector output and capital formation.

Likewise, for output estimation of the informal/unorganised sector, the CSO replaced the average labour productivity measure (obtained as a product of the number of workers and average productivity per worker) with the effective labour input (ELI) method for three categories of workers, by using a production function approach, and summing them up.

Preparation of the Back Series

Preparing the back series of the GDP with a new base year is mandatory with every base-year revision to ensure the availability of consistent long-term economic statistics. In India, it has usually gone back to 1950-51. The last time, (base year 2011-12, after a prolonged delay, the CSO expressed its inability to prepare the back series due to the substantial changes made to the estimation procedures and lack of availability of the older data, especially of the corporate sector database.

However, maybe due to the pressure from  end-users, the CSO did finally prepare a  back series using some ad hoc measures. The back series went back to the 1990s , and not to 1950-51 as earlier. The back series estimates surprised users as they systematically reduced the growth rates of the previous decade (2000s), which made the GDP growth rates in the 2010s higher than those of the previous decade. This change in the decadal growth estimates baffled the user community. Moreover, the official back series were quite at variance with the back series prepared by the National Statistical Commission, compounding the doubts and confusion among data users. Thus, the statistical exercise of back series preparation acquired a political taint in the case of the preparation with the base year of 2011-12.

 How the states measure their GDP:

The CSO/NSO decides the methodologies for estimating state domestic product (SDP) in line with the national accounts’ methodologies, reportedly after due consultation. State statistical bureaus seem to have little leeway in modifying the estimation procedures handed to them, as the national level agency is mainly focused on maintaining consistency at all levels of government.

However, after the 2011-12 base year estimates were released, many serious issues cropped up in SDP numbers that sprang from the use of the MCA database for the manufacturing sector and production function for informal sector output estimation. The crux of the problem was that the national-level parameters were used to obtain the state-level data. Such a methodology overlooked interstate variations in the production structures, thus distorting the absolute levels of estimates and their relative rank across states. This resulted in a reduction of the use of primary data in SDP estimation, from 50 per cent in the earlier series to 30 per cent in the 2011-12 series – a clear regression in the methods of measurement. The CSO set up a committee headed by Professor Ravindra Dholakia to iron out the wrinkles in the SDP estimation whose report is available here (

IMF’s C grade for India’s National accounts:

In the last two years, the International Monetary Fund (IMF) has initiated a process of grading the quality of statistics of its member-countries. According to its mandate, every year under Article IV, in its consultations with its member-countries, the IMF prepares a country report. In November 2025, the IMF released its report for India, which reported its assessment of the country’s statistics. IMF has given C grade to India’s National Accounts, along with China. However, all other Indian statistical series have a grade of B. The reasons mentioned in the report (available here) for the poor grade mentioned in the report are  “(i) an outdated base year (2011/12), (ii) use of wholesale price indices as data sources for deflators due to the lack of producer prices indices, and excessive use of single deflation, which may introduce cyclical biases, (iii) at times sizable discrepancies between production and expenditure approaches, that may indicate the need to enhance the coverage of the expenditure approach data and the informal sector, and (iv) lack of seasonally adjusted data and room for improvement of other statistical techniques used in the quarterly national accounts compilation. On granularity, further breakdown of Gross Fixed Capital Formation by institutional sector (published with a significant lag) and further disaggregation of the quarterly production and expenditure approach estimates would allow for a more detailed analysis of economic trends.”

(For an insightful  commentary, see “Are There Gaps in India’s Economic Data?” by Biswajit Dhar, Business Line, December 16) ().

IMF’s poor grade for the quality of India’s official statistics has put it in an adverse light internationally. However, many of the issues that the IMF report has flagged are widely known and and have been debated in India over the last decade after the revised estimated with the base year 2011-12 was released in 2015 (See R Nagaraj and TN Srinivasan, 2016, India Policy Forum, 2017; Amey Sapre and Vaishali Bhardwaj, Status and Compilation Issues in National Accounts Statistics: A Short Summary, NIPFP Working Paper No. 397 Jul, 2023).

Looking forward, the government has announced the much-delayed revision of the base year of the National Accounts to 2022-23. The NSO is also coming up with a new series of Consumer Price Index (CPI), Wholesale Price Index (WPI) and the Index of Industrial Production (IIP), also with 2022-23 as the base year that would feed into the National Accounts estimation. Recently, the NSO released a paper outlining the proposed changes in the methodologies in the National Accounts estimation ()

Conclusions

The GDP or SDP is the most widely used aggregate economic indicator of economic performance. Their accurate and up-to-date measurement are vital for economic policy. The base year of the measure needs to be revised periodically to capture changing output composition and to net out the effects of price changes. The base year revision is a statistically challenging task requiring credible data from factories, farms and offices. Sound statistical measures are needed to use the mass of data to produce meaningful economic statistics.

 

Schedule of Release of Quarterly/Annual GDP Data in India

The following is the official release calendar as for 202-26. The same holds for all years, though at times there could be a change

India’s Calendar of Release of Quarterly GDP for each financial year

1. Quarter 1 (April-June): August 29
2. Quarter 2 (July-September): November 28
3. Quarter 3 (October-December): February 27
4. Quarter 4 (January-March): May 30 (of the next financial year)

 

India’s Calendar of Release of GDP for each financial year

1. First Advance Estimates of GDP o(on January 7 of the same financial year)
2. Second Advance Estimates of GDP (on February 27 of the same financial year)
3. Provisional Estimates of GDP (on May 30 of the next financial year, T2)
4. First Revised Estimates of GDP (on February 27 of the next financial year, T2)
5. Final Estimates of GDP (on February 27 of financial year T3)