Welcome to your study space
Understand economics at your own pace: step-by-step guided courses, and exercises to practice.
Aggregate demand
The same formula as GDP — but it does not say the same thing, and that is the whole point.
AD(P) = C(P) + I(P) + G + (X − M)(P)Problem / motivation
This course displays almost the same formula as the course “GDP”: C + I + G + (X − M). Same letters, same order. Yet it does not say the same thing — and until you see the difference, everything that follows stays out of reach.
Here is that difference, in one sentence. GDP is an ACCOUNTING IDENTITY: after the fact, you add up spending that really took place. It is true by construction, whatever happens — and that is precisely why it can predict nothing. Aggregate demand is a FUNCTION: it says how much agents WOULD LIKE to buy at a given general price level. It is an intention, measured before the exchange happens, and it changes when that price level changes. The two coincide only at equilibrium — when what some want to buy meets what others agree to produce, which is the subject of the next course, “Shocks and AS-AD equilibrium”.
The whole course follows from that. If AD depends on the price level, then we must say WHERE prices are in the formula (they do not appear in it), through which mechanisms they act, what the resulting curve looks like, and what shifts it — keeping apart two movements that get confused all the time: sliding ALONG the curve, and shifting THE CURVE itself.
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
Take the GDP formula again: GDP = C + I + G + (X − M). It is recorded AFTER the fact, on spending that actually happened; it is an identity, it cannot be false. We keep the same letters, but we change the question: how much would agents want to spend IF the general price level were at this or that level? The answer is no longer a number, it is a list of numbers — one per price level. In other words a function, written AD(P). And while we are at it, let us settle a question the GDP formula leaves hanging: why subtract M? Because an imported product bought by a household is already counted in C; taking it out avoids counting as demand addressed to our producers what goes to foreign producers.
The formula displays no P at all — it acts inside three components. (1) On C, the REAL BALANCE effect, or Pigou effect: at an unchanged quantity of money, lower prices make each euro held more powerful; households feel richer and consume more. (2) On I, the INTEREST RATE effect, or Keynes effect: lower prices reduce the money needed for day-to-day transactions; the surplus gets lent, the interest rate falls, and credit-financed investment becomes more attractive. Careful, this channel assumes the quantity of money unchanged (assumption 3) — in the course “The policy rate and transmission”, by contrast, it is the central bank that steers the rate. (3) On (X − M), the NET EXPORTS effect: lower domestic prices make our products relatively cheaper than foreign ones, so X rises and M falls (course “What is an exchange rate?”). Three channels, three components: G is the only one that does not move with prices.
Now we can finally draw. On the HORIZONTAL axis, the total quantity demanded, in billions of constant euros. On the VERTICAL axis, the general price level, expressed as an index (base 100 in the reference year). Every point on the curve answers: “at this price level, how much do we want to buy?” Since the three channels all push the same way, the curve SLOPES DOWN from left to right: lower prices, larger quantity demanded. One order of magnitude to fix ideas: if a fall of 4 index points (from 100 to 96) raises demand by 2%, that is about €60bn, so the slope is worth −€15bn per index point. That number is ILLUSTRATIVE — no institution publishes “the” slope of aggregate demand — but it turns a vague sentence (“the slope is negative”) into a magnitude you can handle.
Now that the axes exist, the distinction becomes visible. If the PRICE level changes, we SLIDE ALONG the curve: it has not moved, we simply read another of its points. If something else changes — household confidence, a fiscal stimulus, foreign demand, the policy rate — then it is THE WHOLE CURVE that shifts: at every price level, we now want to buy more (or less). And by how much does it shift? Not by the amount injected: by that amount MULTIPLIED, because every euro spent becomes an income that is partly spent again. That is exactly the k of the course “The fiscal multiplier”. Click each term:
Solving / calculation
Let us take the real figures of the French economy — the 2025 national accounts published by INSEE, France's national statistics institute, in May 2026 — and do three things: check what the formula really adds up, turn it into a curve, then apply a shock to it while keeping the impulse apart from the shift.
- The four letters, with the real 2025 figures (€bn)
1,545.9 (C) + 663.6 (I) + 719.6 (G) + (1,009.4 − 1,023.8)€2,914.7bn - Yet 2025 GDP is €2,991.1bn: the four-letter formula does NOT add up
2,991.056 − 2,914.735 (unrounded values: rounding to one decimal would give 76.4, which is the trap)€76.321bn is missing — the textbook formula is a simplification - What those missing €76.3bn are
69.478 (non-profit institutions serving households) + 5.499 (change in inventories) + 1.344 (valuables)€76.321bn — the identity then closes EXACTLY - The weight of each component in GDP
C: 1,545.9/2,991.1; I: 663.6/2,991.1; G: 719.6/2,991.1; X − M: −14.4/2,991.1C 51.7% · I 22.2% · G 24.1% · net exports −0.5% - From the number to the CURVE: we start again from FULL GDP (2,991.1 — the missing 76.3 are demand too) and vary the price level (illustrative toy, sensitivity 0.5% per index point)
at P = 100: 2,991.1; at P = 96: 2,991.1 × (1 + 0.5% × 4) ≈ 3,051. Five points to see the curve: P = 90 → 3,141; P = 95 → 3,066; P = 100 → 2,991; P = 105 → 2,916; P = 110 → 2,842the curve SLOPES DOWN to the right — and in every one of those cases we SLID along it, it has not moved a millimetre - A shock, now: the state spends €40bn more (the IMPULSE)
ΔG = +€40bn, i.e. 1.34% of GDP — a plausible order of magnitude: the French recovery plan announced in September 2020 came to €100bn spread over several yearsbut that is not what the curve shifts by - The SHIFT of the curve: the impulse multiplied (k ≈ 1.4, course “The fiscal multiplier”)
1.4 × 40+€56bn at EVERY price level, i.e. 1.87% of GDP
Three lessons to keep. First, the four-letter formula is a simplification: on the real accounts it falls €76bn short of GDP (non-profit institutions and inventories) — worth knowing before claiming that “it always adds up”. Second, one and the same number — €2,991bn — is an aggregate demand only if you say which price level it corresponds to: that is what makes it a curve rather than a total. Third, and this is the costliest mistake: the impulse (+40) and the shift of the curve (+56) are not the same number. A commentator announcing “€40bn of stimulus, hence €40bn more demand” forgets the multiplier; one promising “€40bn, hence €40bn more GDP” forgets on top of that that demand becomes output only by meeting a supply — the subject of the next course, “Shocks and AS-AD equilibrium”.
AD(P) = AD₀ × [1 − s × (P − 100)/100] + k × ΔGThis simulator separates the two movements everybody mixes up. First move the PRICE LEVEL alone: you slide along the curve, and the line “Shift OF the curve” does not budge. Then move the SHOCK: the curve shifts, and the effect is the same at every price level. Finally compare the impulse you inject with the shift you obtain — the gap is the multiplier.
Economic interpretation
Three warnings, two of which correct very widespread shortcuts.
One is tempted to say: the demand curve for a good slopes down because, if it gets dearer, people switch to another good (course “The law of supply and demand”). That reasoning does NOT work here: when the GENERAL price level rises, there is no other good to switch to, since everything rises together. That is why the slope of AD needs the three channels of the narrative — real balances, interest rate, net exports — and not the substitution argument. Moving from individual reasoning to whole-economy reasoning without changing the mechanism is called a fallacy of composition.
To the RIGHT (we want to buy more at unchanged prices): a rise in household or business confidence, a fiscal stimulus (more G or lower taxes), a cut in the policy rate (course “The policy rate and transmission”), stronger foreign demand, a depreciating currency, a rise in wealth. To the LEFT: the same in reverse, fiscal austerity in particular. One single factor NEVER shifts the curve: the general price level itself — it makes you slide along it. That is the test to apply every time: “did the price change, or something else?”
The most quoted mechanism is also the most fragile. It plays only on net NOMINAL assets — money and fixed-value savings — not on a house or on shares, whose price falls too. Above all, it has a dark mirror: if prices fall, the real value of DEBTS rises just as much, and indebted households have to tighten their belts. Irving Fisher named that phenomenon debt deflation (course “What is deflation?”), and it can outweigh the balance effect. A general fall in prices is therefore not the good news the slope of the curve suggests.
Limits / critiques
Exercises
C = 900, I = 300, G = 400, X = 200, M = 250 (€bn). What is aggregate demand at this price level (in €bn)?
True or false: a fall in the general price level shifts the aggregate demand curve to the right.
In 2025, investment (GFCF) was worth €663.6bn for a GDP of €2,991.1bn. What share of GDP does it represent (in %, one decimal)?
The price level falls. Which channel raises INVESTMENT in that movement?
This country launches a stimulus: G goes from 400 to €460bn. With a multiplier k = 1.4, by how much does the aggregate demand CURVE shift (in €bn)?
Among these four events, which one makes you SLIDE along the AD curve instead of shifting it?
Why are imports subtracted in the formula?
True or false: the aggregate demand curve slopes down for the same reason as the demand curve for a particular good — if the price rises, you buy something else.