Where the number comes from

Chapters 1 to 3 treated the number as given. This is where it is made, and the method is more provisional than its presentation suggests.

Quarterly GDP is extrapolated, not counted

The single most important sentence in the methodology section:

The compilation of Quarterly Estimates of GDP is based on Benchmark-Indicator methodology in which the estimates computed for the previous financial year are extrapolated using the relevant indicators reflecting the performance of various economic and institutional sectors.

Read that carefully. Nobody counts the economy every three months. What happens is:

The benchmark. A detailed annual estimate for the previous financial year, built from surveys, company accounts, administrative records and censuses.

The indicators. A set of fast-moving, frequently published measures that are believed to track each sector.

The extrapolation. Move the benchmark by what the indicators did.

So a quarterly growth rate is a statement about indicators, converted into a statement about the economy by an assumption that the relationship between them held. It is a well-founded estimate made under time pressure, following the International Monetary Fund's Quarterly National Accounts Manual of 2017, and it is not a measurement.

That is not a scandal — every country does it this way, because the alternative is waiting two years. It is a reason to treat a 0.2-point difference between quarters as noise, and chapter 5 is about what happens when the benchmark later arrives.

What the indicators actually are

The press note lists them, and the list is the most grounding thing in this subject. On the production side:

• Advance estimates of crop production; inland and marine fish production; season-wise livestock products
• Granular Index of Industrial Production
• Consumption of natural gas; consumption of steel
• Financial results of listed companies on BSE and NSE
• GST data
• Domestic sales of commercial vehicles and three-wheelers
• Domestic and international air traffic statistics
• Railway net tonne-kilometres, passenger-kilometres, goods and passenger earnings
• Cargo handled at major and minor ports
• Wireless data usage, minutes of usage, internet telephony, telephone subscribers
• Bank-wise loan, deposit and interest information; the same for non-banking financial companies
• Revenue, profit-loss and operating expenses of life and non-life insurers
• Public Financial Management System data of the Central Government
• Tax and subsidy information of state governments from the CAG, and of the Centre from the CGA
• Granular CPI and CPI (Industrial Workers); granular output Producer Price Index

And on the expenditure side, the RBI's balance of payments table, vehicle registrations from e-VAHAN, commodity-wise imports and exports, food grain offtake, and a commodity-wise Unit Value Index.

India's quarterly growth rate is built, in part, from how much freight the railways moved and how much mobile data people used. That is worth knowing for two reasons. It makes "how do they know?" concrete and answerable. And it tells you exactly where the estimate is weakest: wherever an indicator stops tracking the thing it stands for.

Consider the listed-company results in that list. They cover the organised, listed part of corporate India. A quarter in which listed manufacturers do well and unlisted small manufacturers do badly will be read through an indicator that sees only the first group. The method assumes the ratio between them is stable, and in a period of structural change it is not. That assumption is the main reason quarterly estimates get revised.

Double deflation, and an unusually honest warning

The new 2022-23 series changed how manufacturing's real value added is computed, and the press note explains it.

Previously, single deflation: output and the inputs consumed were both deflated using the same price index. Now, double deflation: output and intermediate consumption are deflated separately, with granular producer price indices, and real value added is the difference between the two real figures.

Double deflation is the internationally recommended approach and it is more correct. It is also more volatile, and the Ministry says so in its own release:

under the Double Deflation approach, the implicit price deflator for Manufacturing GVA may decline, or even show a negative growth rate despite an increase in both output and input prices.

It then explains when: if input prices rise faster than output prices, the price component of value added can fall even as every individual price rises. Margins are being squeezed, and value added is a margin.

Two things to take from that. The mechanism is real and will produce headlines about "negative manufacturing inflation" that are not errors. And a statistical office that writes the counter-intuitive consequence of its own new method into the release, before anyone has complained, is behaving the way you would want it to.

The honesty at the end

The methodology section closes with this:

Improved data coverage and revision in input data made by source agencies would have a bearing on subsequent revisions of the estimates. Estimates are, therefore, likely to undergo revisions for the aforesaid causes in due course, as per the release calendar. Users should take these into consideration while interpreting the figures.

The release tells you, in advance, that the number will change and that you should allow for it. Almost nobody quoting a quarterly GDP figure mentions this. Chapter 5 is about how much it changes.

The calendar

Quarterly estimates arrive roughly two months after the quarter ends. The Q1 figures for April to June 2026 were released on 31st August 2026, and the same note scheduled Q2, covering July to September, for 30th November 2026.

Two months is fast for a national accounting exercise and slow for a decision. It is also why the Monetary Policy Committee in chapter 5 of the inflation subject was setting rates in October on a picture of the economy that ended in June.

The point

Quarterly GDP is not measured but extrapolated: last year's detailed benchmark moved by about twenty fast indicators, from railway tonne-kilometres and GST receipts to wireless data usage and listed-company results, under the IMF's 2017 manual. The new series deflates manufacturing output and inputs separately, which is more correct and more volatile — and the Ministry warns in the release that it can show falling value-added prices while every underlying price rises. The same release tells you the estimates will be revised and that you should interpret them accordingly.