The speed of the whole economy.

Every decade or so the statisticians rebuild GDP on a fresh base year: new weights, new source data, new prices to hold constant. The rebuilt series is a better map of today's economy and a different map of yesterday's: the same quarter gets a new growth rate, sometimes a point away from the old one. Neither number is a lie. They answer the same question with different rulers.

This atlas never splices rulers. The decade plate below stays on the 2011–12 base it was drawn on; this panel is the same economy on the 2022–23 base, where the quarterly releases now continue. When enough new-base history accumulates, the new base takes over the whole guide: side by side until then.

The same eight engines as the 2011–12 plate above, on the new ruler. The new base re-weights them: each panel's share of GVA is computed afresh from 2022–23 prices, so a sector's weight here is not comparable with its weight on the old plate. Read each panel's shape, and read the shares as today's economy.

Each line compares one quarter with the same quarter a year earlier. The deep V is the pandemic: a real collapse in mid-2020, then a rebound a year later that is arithmetic, not boom: growth measured against a destroyed base. Read the 2020–22 stretch as a wound and its scar, not as information about trend.

GDP and GVA move together because they are nearly the same thing: GDP = GVA + net indirect taxes. Consumption (PFCE) is the demand engine; when it outgrows GDP, households are confident, and vice versa.

What households spend on goods and services: the largest and steadiest block of demand. It tracks wages, rural incomes and consumer confidence, and it is the single best summary of how ordinary life is going. Sustained strength here is what turns a statistical recovery into a felt one.

Spending on machines, buildings and infrastructure: the economy's bet on its own future. Note the y-axis: investment collapsed more than 40% in the pandemic quarter and rebounded over 60%, the widest swing of any major component. Judge its level over a year, never a quarter.

GVA measures producers' output net of input costs, before product taxes and subsidies. Economists often prefer it as the cleaner read on supply conditions; newspapers usually quote GDP. When the two diverge, the cause is almost always net taxes: check before reading meaning into it.

The three engines the headline plates skip. Government consumption is the state's own spending on salaries and services: about a tenth of GDP, and the only engine a budget can start at will. Exports add demand from abroad; imports subtract it, because GDP counts what is made here, not what is bought here. A quarter where imports outrun exports is a quarter where net trade dragged on the headline.

All three swing harder than consumption. Government spending lurches with the fiscal calendar; trade answers to the world cycle and the oil bill. Read them for composition, not for trend.

This is the supply side of the same economy: who made the output, not who bought it. The eight groups sum to total GVA, so the headline is always their weighted average: services are over half the weight, and they set the level. Watch the scales: each panel has its own, and a flat-looking farm panel hides less drama than a flat-looking finance panel.

Agriculture answers to the monsoon, not the cycle: it was the one engine still running through the lockdown quarter. Construction and trade are the swing engines; they fell close to half and rebuilt from there. Public administration moves with the budget, which makes it steady and tells you little about private demand.

The gap between the two GDP lines is the deflator: the price of everything the economy produces, not just what households consume. It is the broadest inflation gauge India has: and the least familiar, because no press release leads with it. Tax revenue, debt ratios and corporate toplines all live on the nominal line; the real line is what the growth debate is about.

The deflator leans heavily on wholesale prices, a known quirk of Indian national accounting, so it tracks the WPI more than the CPI. That makes it honest about producer prices and unreliable as a cost-of-living measure. It can even turn negative while consumer inflation stays positive, as it did in FY16. When the two price worlds diverge, the deflator sides with the factory gate.

The quarterly print is one number; the annual accounts are a ledger, and the ledger teaches the identities. GDP is GVA plus taxes on products minus subsidies: that wedge is the whole difference between the two headlines, and it moves when policy moves, not when factories do. Net domestic product strips out capital consumption: the wearing-out of machines, roads and software. Gross counts everything made; net counts what is left after replacing what wore out.

Gross national income asks a different question: not what was made here, but what residents earned anywhere. The two travel together for India; the gap is remittances and profits crossing the border.

Households, firms and government earn income and either spend or keep it. The kept share is gross saving: the pool that funds domestic investment. When investment runs above saving, the difference arrives from abroad and the current account turns to deficit. The line is the economy's own savings habit, distinct from how fast it grows.

GDP measured by spending: what households buy (consumption), what is spent on buildings, machines and roads (fixed investment), what the state spends on its own services (government consumption), goods piled up in warehouses (change in stocks), gold and jewellery bought as a store of value (valuables), and what the world buys from us less what we buy from it (net exports). Each panel is that line's share of GDP in rupees of the day.

The long view is a story of a country growing richer. In the early 1950s almost everything earned was eaten, worn or lived in; there was little left over to build with. As incomes rose, the consumption share fell and the investment share rose. That shift is what development looks like in the accounts.

The three rulers never join. The 2022–23 base counts consumption differently (new surveys, new weights), so in the years both bases cover it puts the consumption share several points lower. Neither ruler is wrong; they are different measurements of the same spending.

GDP counts what is produced inside India. Two flows across the border change what residents actually have. Primary income is earnings on work and capital that cross the border: interest on foreign loans, profits of foreign-owned companies sent home, wages of Indians working abroad for short spells. India pays out more than it receives, so the line sits below zero. Adding it to GDP gives gross national income.

Current transfers are mostly remittances: money a nurse in Riyadh or an engineer in Texas sends to parents in Kerala or Punjab. Adding them gives gross national disposable income, the widest measure of what the country can spend or save. In the 1950s the line was close to nothing; the Gulf boom of the 1970s lifted it, and the 1990s and 2000s lifted it again.

India's GDP is built from production: what each industry adds. Spending is estimated separately, from surveys, tax and trade data. The two should match, and they never quite do. The statisticians publish the gap as a line called "discrepancies" and leave it in. Above zero, measured output ran ahead of measured spending; below zero, the reverse.

The API's filter offers an approach cut (production, expenditure, income), but the data it returns is the same for every approach, so this plate computes the gap from the published levels. The annual figures are in the data table below.

The accounts split the economy into kinds of owner: households (with the small charities and trusts that serve them), private companies, public-sector companies, financial firms and government. Each saves part of its income and invests part of it in new buildings, machines and roads. Where a sector saves more than it invests, the surplus flows to the others through bank deposits, insurance, provident funds, mutual funds and government bonds.

Households are India's lender. Their saving pays for their own homes and tractors and still leaves a surplus, lent to companies building factories and to governments running deficits. A government below zero on saving is spending more on salaries, pensions, interest and subsidies than it collects: it borrows before it builds a single road.

Household saving here includes physical saving (houses, gold, farm equipment) as well as financial saving. The split between the two is in MoSPI's yearly publication, not in the data this guide reads; for household borrowing see the credit and debt guides.

Two sectors move the national saving rate most: households, the largest saver, and government, the largest dissaver. The faint lines are the 2011–12 base; the strong ones the 2022–23 base, which measures household income and spending afresh and puts household saving at a different level for the same years.

Below the eight industry groups the accounts publish eighteen sub-industries: inside farming (crops, livestock, forestry, fishing), inside manufacturing (food, textiles, metals, machinery, the rest), inside trade (shops and repair, hotels and restaurants) and inside transport (rail, road, water, air, storage, communication). Each bar is the compound real growth of that sub-industry across the whole series; the figure beside the name is its share of the economy at the end.

A long bar with a tiny share, such as storage or air transport, is a small business growing quickly. A short bar with a big share, such as crops, moves the economy more than any of them.

A year's growth is printed six times or more. The first advance estimate comes in early January, before the year is over, from three quarters of data and a projection. The second advance follows at the end of February, and the provisional estimate at the end of May, once the year has ended. Then come the first, second and third revised estimates, one each February, as company accounts, the industry survey, farm data and tax returns arrive. Each line starts at zero, the first print, and shows how far later prints moved from it.

Revisions are not mistakes. The early estimates lean on fast indicators (listed companies, government accounts, the industrial index); the informal economy is extrapolated from them until slow surveys arrive. When the slow data disagree with the fast, the number moves. A line above zero is a year that turned out better than first thought.

Stage growth is computed from the published levels, against the previous year's figure in the same release. MoSPI's own one-decimal rate is kept beside it in the data table; it is sometimes cut rather than rounded (it gives 9.1 for 2023-24's first revised estimate, where the press note and the levels give 9.2).

The blue line is the NSO's back series: the national accounts re-estimated at 2011–12 prices, all the way back to the first years of the Republic. The statisticians built it so that one ruler covers six decades. The two grey lines are the series the NSO publishes today, on the 2011–12 and the 2022–23 base. Each is drawn as published; none is joined to another.

Read the decades, not the years. A single year can be a drought or a war; a decade is a regime. The averages below are of the blue line alone.

Saving is the share of income the country keeps; fixed investment is the share it spends on buildings, machines and roads. When investment runs above saving, the difference is borrowed from abroad. The gap between the two lines is, roughly, the current account.

The make-up of what India produces, industry by industry, over seventy-five years. Each panel is one industry's share of value added in rupees of the day. A falling line is not a shrinking industry: farming produces far more food today than in 1950. It means the rest of the economy grew faster.

The share of farming fell, but the share of people who farm fell far more slowly. That gap, a large share of workers producing a small share of output, is why farm incomes lag and why so many households want a job off the land (see the jobs guide).