Welcome to your study space
Understand economics at your own pace: step-by-step guided courses, and exercises to practice.
GDP
Measuring a nation's output in a single number — and knowing where the line runs around what it counts.
GDP = C + I + G + (X − M)Problem / motivation
How can you tell whether an economy is “doing well” this year compared with last year, or compared with its neighbour?
Before the 1930s, nobody could answer: activity was judged from scattered indicators — steel output, rail traffic. A single thermometer was missing. It was to put a figure on the collapse of the Great Depression that the economist Simon Kuznets delivered, in 1934 and at the US Senate's request, the first official national income accounts.
After the war, a standardised way of keeping a country's accounts — “national accounting” — took hold everywhere. In Europe it even has the force of a regulation: ESA 2010, which France's INSEE applies. That common manual is what makes a French GDP comparable with a German one — and it is why this course spends as much time on the CONVENTIONS of counting as on the formula.
Assumptions
In your view, which assumptions are needed for this model to hold? Jot down your ideas — no lead is wrong, this is your worksheet.
Your worksheet is still empty. Go for it: propose at least one idea.
0 idea(s) proposedFormalization
GDP is the VALUE — at market prices (assumption 3) — of everything a country PRODUCES in the sense of the production boundary (assumption 2), keeping only final goods and services (assumption 1), over a given PERIOD and for that period alone (assumption 4), INSIDE its territory (assumption 5), without deducting the wear on capital (assumption 6). Every part of that sentence is one of the six conventions you have just unfolded: you have already built the definition.
Asking millions of producers is out of reach; watching what people BUY is far easier. Is that legitimate? Follow a batch of bread worth €110. Three exits take it out of the circuit for good — a FINAL USE, economists say, because nobody transforms it again this year: a household eats it (C), a public body buys it to hand over as it is, like reimbursed medicines (G), a foreign customer takes it away (X). A fourth is no exit at all: the sandwich shop buys it to produce something else — the bread becomes INTERMEDIATE, its value re-emerging in the price of the sandwich, counted once and never twice. And the OVEN? Bought “in order to produce” too, but it outlives the year: its value will never re-emerge in the price of any loaf. So it is a final use — the fourth letter, I. Adding up FINAL uses does come to the same thing as adding up production.
The reasoning above has a hole a quarter of GDP wide: a state school is sold to no pupil, an owner-occupier pays rent to nobody. Those productions are nonetheless inside (assumptions 2 and 3), and no purchase would ever pick them up. The accountant's fix is disarming: since there is no purchase, he INVENTS one at an estimated price and files it under the appropriate letter — imputed rent into C, the public service valued at its cost into G. Entered twice, once as production and once as a fictitious purchase, each imputation makes the identity true by construction. The price of the trick: over the non-market part, GDP does not measure what a service is WORTH, but what it COSTS.
An intruder has slipped into the count. When a household buys a Korean television for €500, those €500 go into C: the accountant observes a purchase, not a factory. Yet nothing was produced here — the “− M” takes them back out. It is therefore NOT a penalty inflicted on imports, it is the subtraction of what had been counted in excess: they already sit inside C, I, G and even X (an exported car contains imported steel). Check it in the simulator — raise C and M by the same amount and GDP does not move; raise M alone and it falls, because the country consumes just as much while producing less. (The same total could be reached by adding up value added, or distributed income: three routes, one number — except that the sum of values added gives GDP NET of taxes on products, some €300bn that INSEE adds back. The course “The three approaches to GDP” runs them.) Click each term:
Solving / calculation
The figures below are a textbook example, scaled to France without being its actual accounts — French GDP came to 2,991 €bn in 2025. Replace each letter with its value, put the formula to the test, then move the sliders: every value below can be reached on them. Amounts in billions of euros, written “€bn”.
- Label them, then sum (C = 1,480, I = 680, G = 700, X = 740, M = 800)
GDP = 1,480 + 680 + 700 + (740 − 800)GDP = 2,800 €bn - Put to the test: households consume 100 more, all of it imported
GDP = 1,580 + 680 + 700 + (740 − 900)GDP = 2,800 €bn — unchanged: the “− M” cancelled the 100 counted in excess inside C - Back to step 1. The following year, with a little more activity
GDP = 1,530 + 690 + 710 + (760 − 834)GDP = 2,856 €bn - Growth g: the gap (56 bn) divided by the starting point (2,800), ×100
g = (2,856 − 2,800) / 2,800 × 100g = 2%
GDP rose by 2% in one year — and that number does not yet say much: part of the rise may come from PRICES alone. Disentangling the two is what the course “Nominal vs real GDP” is about, and growth itself what “GDP growth” is about.
GDP = C + I + G + (X − M)Two experiments are worth the detour. (1) Raise C and M by the SAME amount: GDP does not move, the imported purchase cancels out — that is the demonstration of the 4th beat of the narrative. (2) Raise M alone: GDP falls even though the country consumes just as much, because it produces less of what it consumes. Then look at the weights, set against France. ⚠️ The France line is brought back to the SAME FOUR items as this toy: the real accounts carry two more lines (non-profit institutions and the change in inventories, €76.4bn out of 2,991.1), so the four letters cover only 2,914.7. French consumption comes to 1,545.9 €bn — that is 51.7% of whole GDP, but 53.0% of those 2,914.7. One fact, two denominators. (Weights rounded to a tenth: their displayed sum may come to 100.1%, as in INSEE's own tables.)
Economic interpretation
A GDP on its own says nothing: 2,991 €bn — is that a lot or a little? The number only takes on meaning once COMPARED — and every comparison has its precaution.
France's 2,991 €bn is what was produced IN ONE YEAR, not what the country owns (assumption 4). Hence the right reading of “public debt = 115.6% of GDP” at the end of 2025: not “France owes more than it owns”, but “the debt is worth 1.156 years of production”. A debt-to-GDP ratio is a DURATION, since it divides a stock by a flow — with this unsettling consequence: it can fall without a single euro being repaid, if the denominator grows (course “The dynamics of public debt”).
A GDP going from 2,800 to 2,856 €bn may have risen because more was produced, or simply because everything costs more. The principle that separates them fits in one line: recompute this year's output AT LAST YEAR'S PRICES. If you land back on 2,800, quantities did not move and the whole rise was prices. The index that sums up the gap is the GDP deflator (course “Nominal vs real GDP”). Divided among France's 68.6 million inhabitants, that GDP also comes to about €43,600 per person — an AVERAGE, and the entire subject of the course “GDP per capita”.
This was the promise of the opening question. Converting at the exchange rate is not enough, because a euro does not buy the same thing everywhere. In 2024 China produced 18.7 trillion dollars against 29.2 for the United States at market rates: second place. Corrected for price differences — at purchasing power parity — it produced 38.2 against the very SAME 29.2, and takes first place. The American figure does not move because it is the reference: everything is converted into “international dollars”, defined to be worth exactly one dollar in the United States. Same year, same economies, ranking reversed by the choice of converter alone (course “The real exchange rate and PPP”).
GDP measures what is produced ON the territory, whoever owns it; gross national income (GNI) measures what RESIDENTS earn. In most countries, France included, the gap runs to fractions of a point. In Ireland it is a chasm: foreign multinationals base patents and brands there whose output inflates GDP before returning to their owners. The Irish statistical office therefore publishes a modified GNI, worth only 57% of GDP in 2024 — a spectacular GDP per capita that does not end up in Irish pockets.
Kuznets warned as early as his 1934 report that the welfare of a nation can scarcely be inferred from a measurement of national income alone. He went further: he opposed — and lost — the decision to count government spending as final use, that is, the very “G” you have just learned; it was the wartime accountants, in Keynes's wake, who prevailed. GDP is therefore no law of nature: it is a built tool, made for a purpose, by people who disagreed with one another — and the next station says what that costs.
Limits / critiques
Exercises
In a country, households consume 600 €bn. Firms invest 150, government invests 50, and government spends a further 250 running its public services. Firms have also bought 70 €bn of land. Exports are 120, imports 170. Compute GDP by the expenditure approach (in €bn).
A French grower sells €30 of cotton to a French weaver. The weaver sells €70 of cloth in all: €50 to a French shirt maker, and €20 exported to a foreign customer. The shirt maker sells the shirts for €120 to French households. By how much do these operations raise French GDP?
A small country hosts a foreign-owned plant whose value added comes to €10bn a year; €4bn of profits go back each year to its foreign parent. The rest of its economy generates €90bn of value added. What is that country's GDP (in €bn)?
Of these four situations, which does NOT enter French GDP?
A firm spends, this year: €400,000 on a new machine, €250,000 to buy a warehouse built ten years ago, €90,000 on raw materials entirely transformed within the year, and €60,000 on bespoke software it will use for several years. What amount enters INVESTMENT (I) in French GDP, in €?
C = 900, I = 300, G = 400, X = 250, M = 280 (€bn). That same year, households resold €60bn of existing homes and the state paid out €120bn in retirement pensions. Compute GDP by the expenditure approach.
True or false: cooking a meal at home and paying for exactly the same meal in a restaurant have the same effect on GDP.
A country produces only two things: a market sector whose value added comes to 800 €bn, and a free state school whose cost — wages only — comes to 200 €bn. The state raises teachers' pay by 10%, with teaching hours strictly unchanged. What is measured GDP in value after that rise (in €bn)?