“The right remedy for the trade cycle is not to be found in abolishing booms and thus keeping us permanently in a semi-slump; but in abolishing slumps and thus keeping us permanently in a quasi-boom.”Thus wrote John Maynard Keynes in 1936. More than 80 years on, his insistence that governments should continue to stimulate demand even in a prolonged upswing is newly relevant.The long, slow recovery from the 2008 financial crisis shows how concerns that demand stimulus had exhausted its purpose and should be dialled back have been at best premature. So what have we learnt? The lessons are remarkably similar in all advanced economies.First, demand stimulus works. That simple observation faces surprising resistance. But those who reject stimulus because the economy is still in the doldrums and inflation below target have it the wrong way round. The fact is that a persistent expansionary policy — monetary stimulus everywhere, and fiscal stimulus in the US and a few other places — has kept growth on track well beyond the length of typical economic recoveries, and more stimulus has tended to go with more growth.This has pushed unemployment down and created more jobs than observers thought was safely possible. The “Phillips curve” that warns of inflation rising when labour markets become too tight has been quiescent.Second, the longer demand keeps expanding, the greater the benefits for the least fortunate. Those on the margins of the labour market are typically hurt first and worst in a downturn. Conversely, only when the economy is kept in Keynes’s “quasi-boom” does it bring them towards acceptable levels of unemployment and wage growth.In the US, according to researchers associated with the US Federal Reserve, “when the unemployment rate of whites increases by 1 percentage point, the unemployment rates of African Americans and Hispanics rise by well more than 1 percentage point, on average”. In a recent paper, they found that when labour markets are particularly tight, this extra advantage for marginalised groups becomes even stronger. In this sense, aggressive demand stimulus becomes more, not less, beneficial the longer it goes on.We can see this phenomenon in many of the numbers and stories that describe the current state of our economies. In the US and the UK, recent wage growth has been strongest for those paid the least. In much of Europe, more of the population has a job than ever before. Anecdotes abound about those previously given up as hopeless cases — former drug addicts and ex-prisoners, for example — now being not just hired, but trained to earn their keep by employers struggling to fill vacancies.We are entitled to hope for a third lesson: that the historical pattern of productivity increasing in an upswing will also come through this time. The reason for procyclical productivity is that robust demand growth creates incentives for businesses to do more with the same resources once it becomes difficult to expand by just hiring more.We have been lucky in that many policymakers, especially central bankers, have been more willing to believe their eyes than their theories. They have seen their policies working for longer than expected. They have noted the positive results — often highlighting the lack of inflationary pressures, the benefits of driving employment higher for those on the margins, and the possibility that supply capacity adjusts to satisfy demand pressures.Because of this sensitivity to economic reality, the end of demand stimulus has been postponed many times. Without it, more policymakers would have caused unnecessary slowdowns or downturns — as the eurozone did by tightening fiscal and monetary policy in 2011.But avoiding unforced mistakes is not enough for sound policymaking. At a minimum, the experiences of the past decade call for a much more tolerant attitude to stimulus, whether from finance ministries or central banks. Better still would be to incorporate this data formally in the formulation of policy targets — whether central banks’ mandates or rules for government budgets.In practice, this could mean at least three things. One would be to increase the burden of proof for scaling back stimulus on the grounds that the economy had reached full capacity. That would avoid premature tightening — and the subsequent need to loosen.Another would be to shift the focus from how the economy performs on average to how it performs for those on its margins. All-economy (un)employment and inflation rates could be complemented by explicitly taking account of labour market and wage performance for vulnerable groups, trading off higher tolerance for overall inflation risk against greater expected benefits for those often left by the wayside.Third, the effect of demand pressure on productivity growth could be explicitly included in forecasts.Such steps could introduce an inflationary bias — but only if there was no risk of leaving unused economic capacity on the table. But as the past few years have shown, that risk is real. In the 1980s, economic theory concluded that responsible macroeconomic policy required policymakers who were more hawkish than voters. Today, we need the opposite.