From the gap to the jobless: Okun's bridge
By now you have met two big pictures that never quite touched. From the cycle rung you know the output gap — the distance between what an economy actually produces and the potential output it could sustain. From this rung you know cyclical unemployment — the joblessness that swells in a slump and drains away in a boom. It would be strange if these two were unrelated: when factories run cold and shops stand empty, surely people are out of work too. The question is *how much* — and that is exactly what an American economist named Arthur Okun measured in 1962.
Okun simply plotted the two against each other across years of US data and found a stubborn, almost embarrassingly tidy relationship. Okun's law says that for every extra slice of output an economy loses below potential, unemployment rises by a smaller, fairly steady fraction. In its most-quoted rough form: each *percentage point* the unemployment rate climbs above the natural rate tends to come with the economy producing roughly two percentage points below its potential. The exact number varies by country and era — somewhere near 2 in the postwar US, less elsewhere — but the shape holds: lost output and lost jobs march together, in a known ratio.
Okun's law (rough US form):
Output gap ~= -2 x (unemployment - natural rate)
Say the natural rate is 4%, and unemployment jumps to 7%:
unemployment - natural rate = 7% - 4% = 3 points
estimated output gap = -2 x 3 = -6%
=> the economy is producing about 6% BELOW potential
(~3 jobless points cost ~6 points of lost GDP)
Why more than 1-for-1? In a slump firms also cut HOURS,
freeze hiring of new workers, and idle machines -- so output
falls faster than the headcount of the unemployed alone.What Okun is, and what it is not
It is worth being precise about what kind of thing Okun's law is, because the word "law" oversells it. It is not a law of nature like gravity; it is an empirical regularity — a pattern that has held remarkably well across decades of data, but only *because* the underlying machinery (cutting hours, freezing hiring, idling plant) keeps behaving the same way. Change that machinery and the number drifts. The ratio is looser outside the US, it wobbles from cycle to cycle, and economists openly argue about whether recent recessions have bent it.
Notice too that Okun's law speaks only about the *cyclical* gap — the part of unemployment caused by weak demand, measured against the natural rate. It says nothing about structural unemployment, the joblessness that comes from skills not matching openings. If a region's coal mines close for good, no amount of closing the output gap brings those exact jobs back; that is a structural problem wearing a cyclical disguise. Okun is a thermometer for the demand side of the cycle, not a cure for every kind of joblessness — a distinction the next two sections turn on.
When a slump leaves a scar: hysteresis
Everything so far assumes a comforting symmetry: demand falls, cyclical unemployment rises, and when demand returns, the jobless go back to work and the natural rate is right where it was. But what if a deep enough recession could *change* the natural rate itself — could make the floor permanently higher? That unsettling idea is hysteresis, a word physicists use for systems that stay bent even after the force that bent them is gone.
The human channels are painfully concrete. A worker out of a job for two years sees their human capital erode — skills rust, professional networks fade, and employers, scanning a long gap on a résumé, quietly move the application to the bottom of the pile. The long-term unemployed start to exert little downward pressure on wages, because firms barely consider them, so the economy can run "hot" without their ever being rehired. Through channels like these, what began as ordinary cyclical unemployment can harden into structural unemployment — and the natural rate the recovery returns to is higher than the one the slump began from.
If hysteresis is real, the stakes of recession management change completely. A downturn is no longer a passing storm you can wait out; every extra month of high unemployment risks baking permanent damage into the economy's foundations. That logic — most associated with the economists Olivier Blanchard and Lawrence Summers, who studied Europe's stubbornly high unemployment in the 1980s — is a powerful argument for acting *fast and hard* against slumps. It is honestly debated: the *size* of hysteresis is hard to pin down, and some recoveries have clawed back nearly all the lost ground. But the asymmetric risk it implies — that waiting can cost you forever — now shapes how serious policymakers think.
Two toolkits: active vs passive policy
So what can a government actually *do* for jobs? It helps to split the toolkit in two. Passive labour-market policy cushions the fall: unemployment benefits, income support, early-retirement schemes. It does not try to get you back to work; it keeps you afloat while you are out of it. Active labour-market policy does the opposite — it spends to push you *back into* a job: training and re-skilling programmes, public employment services that match workers to openings, wage subsidies, and direct public-job creation.
The two are not rivals so much as a matched pair, and they fit the different *types* of unemployment you met earlier. Cyclical joblessness — a demand shortfall — is best met by reviving demand itself, the work of the fiscal and monetary tools from the macro rungs; passive support buys time until that demand returns. Structural and frictional joblessness need the active tools: retraining for the laid-off miner, better job-matching for the worker who simply cannot find the opening that fits. Matching the remedy to the *kind* of unemployment is the single most important move in this whole field — using a demand cure on a structural problem just stokes inflation without curing the joblessness.
The evidence on active policy is genuinely mixed, and worth being honest about. Some programmes work well — careful job-search assistance and well-targeted wage subsidies often pay for themselves. Others disappoint: hastily designed training that teaches yesterday's skills, or subsidies that pay firms to hire people they would have hired anyway. The economists' verdict is not "these tools are useless" nor "they are magic," but the duller, truer "design and targeting decide everything." A policy that closes the output gap but ignores the long-term unemployed can leave hysteresis untouched.
The honest limits — and the rung in one breath
Time to be plain about what macro policy *cannot* do. It cannot drive unemployment to zero, because some is frictional and some structural — floors that demand management never reaches. It cannot push below the natural rate for long without paying in accelerating inflation, the lesson of the long-run Phillips curve. It cannot instantly retrain a welder into a software tester, however generous the budget. And it works through long, uncertain lags, so even a perfectly aimed stimulus may land after the moment that needed it has passed. Policy can shrink the cyclical part of joblessness and blunt the worst of hysteresis; it cannot abolish unemployment, and any politician who promises to is selling something.
Step back and the whole rung snaps into one shape. You learned that the headline rate measures one narrow thing and hides discouraged and underemployed people just outside its frame. You learned that its single number is stitched from frictional, structural, seasonal, and cyclical stories. You met the natural rate and the cruel short-run trade-off of the Phillips curve. And now Okun's law has tied joblessness back to lost output, hysteresis has warned that slumps can scar, and the two policy toolkits have shown what — within honest limits — can be done. Behind every tick of that famous number is a person waiting for work; this rung was about seeing them clearly enough to help.