Welcome to your study space
Understand economics at your own pace: step-by-step guided courses, and exercises to practice.
Okun's law
Why it takes growth to bring unemployment down.
Δu ≈ −c × (g − ḡ)Problem / motivation
Everyone calls for “more growth for jobs”. But how much growth, exactly, does it take for unemployment to fall?
The American economist Arthur Okun — an adviser to President Kennedy — measured a regular relation in 1962: growing at the economy's “usual” pace (its POTENTIAL) is only enough to hold unemployment steady; to bring it DOWN, you have to grow faster than that potential, and below it, unemployment rises. Why? Because every year, growth is already needed just to absorb two silent forces — like a moving walkway sliding backwards under your feet; the formalisation unfolds it. And every economy has ITS OWN potential: about 2% for Okun's United States, of the order of 1% for France today.
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
Two silent forces work against employment every year. PRODUCTIVITY first: with machines and organisation helping, the same employees produce a little more (say +1% a year) — at unchanged output, FEWER hands would therefore be needed. NEW ENTRANTS next: young people finish their studies, inactive people come back looking for a post — the activity rate (the share of 15-64s working or searching) is at a record in France, course “Measuring unemployment” — so EXTRA jobs are needed just for them. Add them up: it takes roughly productivity + new entrants of growth just to keep the unemployment rate still. That is potential ḡ — the moving walkway: ~1 + 1 = 2% for Okun's United States.
Once the walkway is paid for, what counts is the gap g − ḡ: above it, the economy hires beyond replacement and the unemployment rate falls; below it, the rate rises. The change in the rate is written Δu — “delta u”, Δ being the Greek letter for CHANGE: Δu = +0.5 means “the unemployment rate gains 0.5 point in a year” (a POINT: from 7.5% to 8%, not “+0.5% of something”).
That leaves the strength of the link: c. Why not one for one? Because firms do not adjust employment on a sixpence: faced with a dip, they first keep their employees (“labour hoarding”), cut hours, freeze hiring. And one last shock absorber, counter-intuitive: discouraged people drop out of the count of the unemployed altogether (course “Measuring unemployment”), which cushions the rate without creating a single job. Result: one point of growth lost costs only about half a point of unemployment. c ≈ 0.5 here — estimated by regression, variable across countries.
Put it together: Δu ≈ −c × (g − ḡ). It reads like a sentence — unemployment moves in the OPPOSITE direction (the − sign) to the gap from potential, with strength c. And it turns over — the route takes two moves: aiming for Δu = −1 means setting −c × (g − ḡ) = −1, hence g − ḡ = 1/c. It takes potential PLUS 1/c points of growth, that is ~2 points with c = 0.5: the numerical answer to the opening question. Click each term:
Solving / calculation
A toy country with American features: starting unemployment u = 7.5%, potential ḡ = 2%, coefficient c = 0.5. This year, growth stalls at g = 0.5%. Beware the trap of the two 0.5s — growth g and coefficient c happen to have the same value by coincidence, so keep track of which is which. Let's unroll it, then take the controls.
- Growth gap: the economy is running 1.5 points below its walkway
g − ḡ = 0.5 − 2.0= −1.5 point - The coefficient turns it into unemployment — and minus times minus makes plus: the NEGATIVE gap pushes u UP
Δu ≈ −0.5 × (−1.5)= +0.75 point - From Δ to level: the starting rate absorbs the change
u: 7.5% + 0.75 pt → 8.25%+0.75 point of unemployment - And the other way round — aiming for −1 point would require running FASTER than the walkway by 1/c = 2 points
g = ḡ + 1/c = 2 + 2g = 4%
Growing below potential (0.5% instead of 2%) pushes unemployment UP by ~0.75 point; growing at 3.5% would bring it down by as much — and aiming for −1 point would take 4% growth. There is the numerical answer to the opening line: “growth” means “clearly above potential”, not merely “positive”.
Δu ≈ −c × (g − ḡ)Set the starting unemployment rate, growth, potential and the coefficient: the simulator deduces the change, the rate ONE YEAR LATER, and the growth it would take to gain 1 point.
Economic interpretation
Okun's law sheds light on why fighting unemployment demands SUSTAINED growth — and on what can, or cannot, be made of it.
Below potential, the unemployment RATE rises; but jobs can perfectly well be created below the threshold — simply more slowly than new entrants arrive. France in 2023 showed this: job creation still positive with growth of only around 1%. The rate relates the unemployed to a labour force that is GROWING (course “Measuring unemployment”): that is the race Okun arbitrates.
“How much growth does it take to bring unemployment down?” Okun's answer: potential, plus 1/c per point of fall aimed at in one year. With c ≈ 0.5: ḡ + 2 points — that is ~4% for Okun's United States, of the order of 3% for a France whose potential is close to 1%. Orders of magnitude, not prophecy (limits 1 and 2).
Our formula speaks in SPEEDS (growth rates). Economists also use a LEVELS version: one excess point of unemployment corresponds to about 2% of GDP not produced — the “output gap”, taken as given here. Same spirit, different snapshot: it is the cost in wealth of cyclical unemployment — the unemployment created by the cycle, the ups and downs of activity, and which growth can absorb.
Okun speaks only of unemployment linked to the business CYCLE. The STRUCTURAL part — skills mismatch, labour market institutions — is not absorbed by growth: it is equilibrium unemployment, built in the course “Equilibrium unemployment (NAIRU / WS-PS)”. A live illustration: in early 2026, French unemployment gained +0.7 point over a year, sluggish growth obliging — pure Okun; but the floor around which it has oscillated for years is structural.
Limits / critiques
Exercises
Growth = 1%, potential = 2%, coefficient = 0.5. By how much does the unemployment rate change (in points)?
True or false: with a coefficient of 0.5, a 1 point rise in unemployment reflects about 2 points of growth MISSING relative to potential.
Potential = 2%, coefficient = 0.5. What growth g does it take to bring unemployment down by 1 point in a year (in %)?
g = 0.5%, ḡ = 1.5%, c = 0.3. By how much does unemployment change (in points)?
To bring unemployment DOWN, the economy has to grow…
Unemployment is at 8.1%. The economy grows at 3.5% for a potential of 2% and c = 0.5. What unemployment rate one year later (in %)?
In early 2026, the French unemployment rate rose by 0.7 point over a year. Without knowing the exact coefficient, what can Okun tell you?
True or false: if growth is below potential, the economy necessarily creates NO jobs at all.