22 January 2018

The free-market case for basic income


Basic income is often dismissed as far-Left utopian thinking. Yet long before Shadow Chancellor John McDonnell and Green Party MP Caroline Lucas were backing the idea, it was being proposed by free-market royalty. Nobel Laureates Friedrich Hayek, Milton Friedman and George Stigler have all argued that unconditional cash payments would be an improvement on today’s labyrinthine welfare state.

Left-wing councils may be piloting it in Scotland, but Donald Rumsfeld and Dick Cheney carried out the first experiments in the late 60s. It is important that basic income isn’t the sole preserve of the anti-market Left, because it could be free marketeers’ best way of responding to the challenges posed by automation.

In one sense, there has never been a worse time to fret about automation. Employment is at record levels, while productivity is stagnant. Outlandish forecasts of technological unemployment have been proven wrong again and again. Internet pioneer J C R Licklider’s prediction that machines would complement (and not replace) humans has (so far) been correct. Economists have been worrying about automation since the industrial revolution, yet they typically underestimate the ability of entrepreneurs to find new uses for labour and discount how technology can make existing workers more productive. After all, spreadsheets led to more, not fewer, jobs for accountants.

Only a fool would say “this time is different”. But, this time might just be different. In the past, machines have been very good at tasks that humans are bad at. At the same time, workers have benefitted from the fact that we know more than we can tell. For example, it’s very difficult to program a machine to distinguish between street signs and shopfronts. Our tacit knowledge of how to break an egg or ride a bike has given us a competitive advantage over machines.

Machine learning is changing that. It combines statistics with trial-and-error reasoning to provide best-guess answers in the absence of clear rules. AI isn’t just beating the world’s best chess players, it’s even beating the world’s best poker players. Until recently, supercomputers that could name the sixty-trillionth digit of Pi couldn’t drive. Now, 100,000s of hauliers and cabbies may be redundant in the next decade. But it’s not just lorry drivers. Lawyers are also under threat. AI search systems are already beating junior lawyers and paralegals at finding relevant documents to settle disputes. By automating expensive tasks, AI will raise living standards massively.

But there is a real risk that the speed of technological change outpaces the ability of entrepreneurs to create new jobs. And even if entrepreneurs keep up, the new jobs they create could be less secure and pay less than the tasks they automated. As my former boss once said, “automation is a lot like going to Australia, it’s great when you get there – but the journey is a pain.”

Our welfare system is well-equipped for a world where most jobs offer a tolerable standard of living, but the system isn’t adapted to in-work poverty. We cannot guarantee that all the new jobs created will pay as well as the older jobs they replace. If well-organised groups see automation as lowering their living standards, they will fight for Luddite regulation to slow the pace of technological change, leaving the majority worse off.

We’ve already seen a backlash against trade in the US, where workers hit by the China Shock backed protectionist Trump. In London, Black Cab drivers have lobbied hard to restrict choice to prevent Uber drivers from driving down their wages. Trade with China and the ability of Londoners to use Uber have both been massively beneficial. But they also threaten the privileged position of some workers who inevitably fight hard to maintain their status.

Basic income could be the answer. Instead of only paying benefits to the jobless, cash would also go to those in work on low incomes. The steelworker now working in a call centre may be less likely to call for protection if their drop in income is less steep.

There are other benefits too; reduced means-testing means fewer bureaucrats in Whitehall, allowing welfare spending to stretch further, while benefit withdrawal rates (that can be as high 76 per cent) will no longer discourage work.

If free marketeers lack a response to in-work poverty, the ground will be ripe for protectionists and socialists to protect jobs at the expense of greater prosperity for the majority.

But, there are big questions for basic income advocates that still need answering. Should it replace welfare altogether or merely some benefits? How generous should the payments be? Should we pay out a flat payment and tax it back through general taxation (UBI) or claw it back directly at a fixed rate (the negative income tax advocated by Friedman)? Or is it simply enough to apply the insights of basic income to in-work tax credits replicating a beefed-up version of the US’s Earned Income Tax Credit?

All these questions are vital. But they can’t be answered by theory alone. We need hard evidence. In the Adam Smith Institute’s new report, Basic Income Around the World, Otto Lehto highlight’s six studies currently taking place that will help us better answer those questions. But we should do more. The government should work with philanthropists to fund pilots of basic income. Perhaps basic income’s critics will be proven right, but it’s time for hard data to settle the debate.

Sam Dumitriu is the Head of Research at the Adam Smith Institute