What GDP does not tell you

Five chapters of what GDP is and how it is built. This one is about the claims it cannot support, which matters because it is routinely asked to support them.

It was never a welfare measure

The point is as old as the measure. Simon Kuznets, who built the first national accounts for the United States in the 1930s, told Congress in 1934 that the welfare of a nation can scarcely be inferred from a measurement of national income.

The man who invented it said it does not measure what people want it to measure, in the document introducing it, and it has been asked to ever since.

This chapter is not an argument that GDP is bad. It is an argument that it is a production measure, and that production is one input into how well a country is doing rather than the thing itself.

Four things it counts that you might not want counted

Repairing damage is output. A flood destroys ten thousand homes; rebuilding them adds to GDP. The destruction subtracts nothing, because the houses were a stock and GDP measures flows. A country can raise measured GDP through disasters it then repairs.

Depletion is free. Chapter 1 noted that GDP is gross of the capital used up. It is also gross of natural capital. A felled forest is output; the missing forest is in no account. A depleted aquifer is output now and nothing later. This is the sharpest version of the "gross" problem and the one with the longest consequences.

Harm sold at a price counts. Tobacco sales, the healthcare that treats the consequences, the legal costs of an accident, the fuel burned sitting in traffic. All transactions, all output.

Defensive spending counts as progress. Money spent on security, pollution control and remediation enters GDP identically to money spent on things people actively want.

None of this is a measurement error. It all follows correctly from "market value of final goods and services produced". The error is in reading the result as a score.

Three things it leaves out

Who got it. GDP has no distributional content whatever. A quarter in which all the growth accrued to a thousand people and a quarter in which it was spread evenly are the same quarter to this measure. Nothing in the accounts can distinguish them, and nothing in them is meant to.

Unpaid work. Chapter 1's production boundary. Cooking, cleaning, childcare and elder care are excluded unless paid for. The consequence is systematic rather than random — in India, where a large share of this work is done unpaid and overwhelmingly by women, GDP excludes a quantity of real production that would be substantial if valued.

Everything without a price. Clean air, safety, free time, the quality of public space. Not mismeasured — absent.

Per capita, which changes the story

The single most useful correction, and it requires no new data.

Aggregate GDP growth of 7.8% in a country whose population is growing means that output per person grew by less. The difference compounds: over a decade, a percentage point of population growth is a materially different trajectory for the person living in the country, while the headline looks the same.

So for "is the country bigger", use GDP. For "is the average person better off", use GDP per capita — and remember that "average" is doing a lot of work, because of the distribution point above. Both corrections run the same way, and both make the headline look better than the lived experience.

Reading a quarter's composition, and a trap in doing it

The expenditure side tells you where growth came from, and the exercise above asks you to do it. Here is the arithmetic for Q1 2026-27, at constant prices, with each component's change expressed against the previous year's GDP of ₹75,46,000 crore:

Component	Growth	Contribution
Private final consumption (PFCE)	7.1%	+3.95 pts
Gross fixed capital formation (GFCF)	11.9%	+3.95 pts
Net exports	—	+3.01 pts
Government final consumption (GFCE)	4.3%	+0.47 pts
Those four together		+11.39 pts
Published real GDP growth		7.8 pts

The four obvious components sum to more than eleven points against published growth of 7.8. Something subtracted about three and a half points, and it is the components nobody quotes: changes in stocks, valuables, and the statistical discrepancy.

This is the trap. A contribution analysis using only consumption, investment, government and trade will always be wrong, sometimes badly, and the direction is unpredictable. The residual is not noise — in this quarter it was larger than the entire contribution of government spending, several times over.

Two legitimate findings survive the caveat, and they are worth having.

Investment did as much as consumption. Fixed capital formation grew 11.9% against consumption's 7.1%, and the two contributed about equally. An economy where investment is growing faster than consumption is building capacity, which is the pattern that supports future output rather than current comfort.

Net exports contributed about three points, because exports grew 12.0% while imports fell 1.1%. Note what that means arithmetically: since GDP subtracts imports, falling imports raise measured GDP. That is correct accounting and a poor description of strength — imports fall when domestic demand weakens as readily as when a country substitutes production for them. A quarter flattered by shrinking imports is a quarter to look at twice.

What to do with all this

Not discard the number. Three habits.

Pair it. GDP with GDP per capita, growth with its composition, this quarter with the vintage question of chapter 5.

Ask who. The aggregate is silent on distribution and no amount of staring at it will help. That question needs consumption surveys and income data, which are separate exercises.

Remember what it is for. GDP was built to tell a government how much an economy could produce — originally, how much a country could produce for a war. It does that job well. It is a thermometer, and a thermometer is extremely useful and tells you nothing about whether the patient is happy.

The point

GDP measures production, not welfare, and the man who built it said so at the outset. It counts repairing damage and the depletion of natural capital as output, and leaves out unpaid work, distribution and everything without a price. Use per-capita figures to ask whether people are better off, and read the expenditure composition carefully — in Q1 of 2026-27 the four headline components summed to 11.39 points against published growth of 7.8, and the gap was stocks, valuables and the discrepancy. Falling imports raised that quarter's measured GDP, which is correct accounting and a weak description of strength.