WHEN should policymakers stop stimulating an economy? America’s unemployment rate is 4.1%. At such a low level, the Federal Reserve, the central bank, would normally expect inflation to rise. In 2017 the economy grew by 2.5%, spurred by falls in joblessness that cannot go on forever. And inflation—though still below the Fed’s target—has overshot forecasts in recent months. All that might suggest that stimulus has become unnecessary. Yet America is cutting taxes and raising spending. As a result, in 2018 and 2019 it is poised to run an experiment. By stimulating economic activity when times are already good, it will find out what happens when the economy runs hot.

The target at which central bankers usually aim when setting monetary policy is the so-called “natural” rate of unemployment. It is elusive. After the financial crisis, when unemployment rose to 10% but inflation failed to subside, some economists speculated that the natural rate was as high as 6.5%. In hindsight those forecasts were far too pessimistic. Today, after years of non-inflationary job growth, there is a widespread consensus that the natural rate is around 4-5%. But this could be a reflection of groupthink. The natural rate is notoriously hard to estimate. Running the economy hot should reveal with certainty where the limits of the labour market really are.

It is also possible that pushing down against the natural rate of unemployment causes it to fall. Economists have long speculated that downturns might damage the supply side of the economy. They might, for example, cause workers to give up job-seeking and leave the labour force entirely. With plenty of cheap workers around after a recession, firms might be discouraged from investing in technology to boost productivity and reduce labour costs. Cheerleaders for a hot economy hope these forces will go into reverse as unemployment falls. Americans of working age are still participating in the labour force at a lower rate than before the recession, and productivity growth has been notoriously weak since the financial crisis. If, in a hot economy, rising wages entice more people to seek work, and spur firms to invest in, say, labour-saving artificial intelligence, then unemployment should be able to fall further without sparking inflation. The biggest beneficiaries of the hot economy are likely to be lower-skilled workers, who tend to reap the biggest wage gains when unemployment is very low.

The risk of a hot economy is that these arguments are wrong, and so inflation takes off. The relationship between unemployment and price rises has been so weak in recent years that hawks fear that if inflation overshoots the Fed’s target, it could take a deep recession to bring it back down. Thankfully, a sudden spike in price growth looks unlikely. Inflation expectations seem very firmly anchored, and wage growth remains very predictable. What is more, say doves, a small inflation overshoot, and expectations of slightly higher inflation, would be no bad thing. Because inflation erodes the value of a fixed return, expectations of higher inflation would depress the “real” interest rate. That would allow the Fed to set higher policy rates without harming growth. As a result, when the next recession comes, the Fed would have more room to cut rates before they run up against their lower bound, near zero. Running the economy hot now might make it easier to keep it warm later on.

See also: “What will result from America’s strangely timed fiscal stimulus?” and our Economics Brief on “The natural rate of unemployment