Sam Bowman is Executive Director of the Adam Smith Institute.

Free marketeers are sometimes accused of crying wolf over the minimum wage. Despite what we believe about price floors causing an oversupply (which, when it comes to workers, we call unemployment), there are examples of this not happening. The UK’s National Minimum Wage, for example, seems to have been quite benign.

That was one reason George Osborne got more or less a free pass from his own side when he introduced the National Living Wage in 2015 – a substantial hike to the Minimum Wage by another name.

Sure, some of his backbenchers warned that it would risk destroying jobs, but hadn’t those same things been said by the same people back in 1998 and been proved wrong?

The minimum wage looks like it’s been a free lunch. Well, not so fast.

We know that economies are complicated things, and looking at only a handful of examples might not give us the full story because other factors (strong economic growth driven by globalisation, for example) might obscure the effect of that which we’re trying to study.

One way to get around that is to look at a large number of examples of where it has happened, or to try to compare like with like – states that are similar in many respects and growing at the same clip, but where only one has seen a rise in the minimum wage, for example.

When we look at minimum wage rises across the United States, where each state has its own rate (a treasure trove for data-hungry economists), and other developed countries, the evidence and the economists’ model begin to fit each other much more nicely.

There are exceptions, but overall we see a pretty clear and consistent pattern of minimum wage rises leading to job losses – not massive ones, but significant and not to be ignored. In a new Adam Smith Institute paper out today, I review the research into these effects and discuss why the National Living Wage might be different.

The graph below shows a scatter plot of over one thousand different data points – the further left of the centre line the dot is, the worse the job losses associated with a minimum wage rise were in that case.

The red line is the estimate of David Neumark, an expert in minimum wages, as to the average unemployment they create – every ten per cent rise in the minimum wage should cause a two per cent fall in employment for people affected by it. Of over one hundred studies Prof Neumark reviewed, two-thirds showed a statistically significant unemployment effect.

ASI plot

Recent evidence suggests that we might be missing the real story. What if firms don’t react immediately to changes in labour costs, instead adjusting their long-term plans?

Most firms are loath to fire workers unless they absolutely have to, and might even hold on to a worker who is no longer making them money to spare their other employees the morale hit of losing a colleague.

Evidence from long-term, time-adjusted data in the US suggests that many firms won’t fire workers when the minimum wage rises, even if simple accounting might imply they should, but they will forgo hiring other workers in future that they would have taken on.

Employment growth, rather than the number of actual jobs right now, is what gets hit. That’s no better for the person who would have been able to take a job that now doesn’t exist, but it’s much more difficult to spot in a straightforward eyeballing of the data.

Then there’s a problem of ‘non-linearity’ – a technical way of saying that a one pound rise from £7 to £8 might be less affordable for firms than a one pound rise from £5 to £6. Again, we’re better off looking at the broad evidence, instead of just looking at recent experience.

Ultimately, minimum wages are a form of redistribution. That doesn’t make them bad, but remember that the money comes from somewhere. Politicians like them because they’re done off-balance sheet – it’s not as clear to the people paying that they’re losing out as when taxes go up.

If the result is higher prices, then we might just be taking money from poor consumers to give a pay rise to, say, second earners in a better-off household who need the money less.

There are other reforms that are much less risky, from the perspective of the poor: planning changes that make it easier to build dense, beautiful housing in cities and green suburbs outside our cities that would lower the cost of housing. And we could make changes to childcare regulations that bring us into line with much cheaper countries like Denmark in Europe.

If we want to redistribute to the working poor, simplifying and strengthening tax credits would avoid some of the unintended consequences we find with minimum wages.

The National Minimum Wage was set by a panel of economists, industry representatives, and trade unionists, with a mandate not to endanger jobs. They were restrained, and avoided the large rises that many people feared.

The National Living Wage is being set without any care for that evidence. George Osborne’s main goal was to undermine Labour after the 2015 election. Well, fine, but they’ve done a good enough job of that themselves.

Politics shouldn’t trump the well-being of people who need to work, and whose prospects might be seriously harmed by further rises to the Living Wage. Let’s stick with what works.