Undercover Economist

The pillars of tax wisdom

‘The Budget bears not even a passing resemblance to how we economists were taught that taxes should work’

I sometimes feel that seeing the world through the eyes of an economist is like seeing the world through the ears of a bat. We notice a lot that others miss, and we miss a lot that others notice.

The annual rituals of the chancellor’s Budget, and next week’s Autumn Statement — focal points of the British political calendar — are good examples. Tax thresholds are nudged up and down, allowances introduced and withdrawn, and much attention is given to the tax on a pint of beer. None of this performance bears even a passing resemblance to how we economists were taught that taxes should work.

So how would the Budget look if designed by economists? Economists’ ideas about taxation are based on three pillars. The first, developed by Cambridge economist Arthur Pigou in 1920, is that we should tax things that have unpleasant spillover effects on bystanders — “externalities”. The classic example is to tax activities that produce pollution. A few Pigouvian taxes exist in the UK but they are patchy: the tax on petrol is implausibly high, for example, while that on domestic fuel is strangely low.

The second pillar was erected in 1927 by another Cambridge man, Frank Ramsey, shortly before his death at the age of 26. Ramsey showed that taxes should be focused on products that aren’t very responsive to price. This is because a tax on a price-sensitive good will simply destroy demand. The consumer won’t buy the good, the retailer can’t sell the good, and the taxman doesn’t collect any tax. Everyone loses out.

Ramsey’s ideas, too, are patchily implemented. Basic foodstuffs such as rice and bread might look like excellent candidates for high taxes in the pages of a learned journal but less so on the front page of a newspaper.

The third pillar was unveiled in 1971 by James Mirrlees — now a Nobel laureate — who tried to figure out what could be said about optimal income taxation. One of his conclusions, surprising to him as much as anyone else, was that an optimal income tax might impose flat or even falling marginal tax rates.

This counter-intuitive idea requires us to see the difference between the marginal rate of tax — the headline rate, paid on each extra pound earned — and the average rate of tax that an individual pays. The two can be very different: if everyone pays a marginal tax rate at 50 per cent, with a £10,000 allowance, then someone with an income of £10,000 pays no tax; an income of £20,000 will attract a 25 per cent average rate; an income of £1m will attract a 49.5 per cent average rate. Yet while the tax burden is progressive, everyone must give half of any extra earnings to the taxman.

Why did Mirrlees argue that the best marginal rate might be falling or flat, as in the example above? The answer is that high marginal rates on top incomes are almost a pure discouragement for the rich to earn money. But high marginal rates on lower incomes will raise money from lots of people, without discouraging work. If you earn £40,000 and the chancellor raises income tax in the £20K to £30K band, that should encourage you to work harder.

This isn’t conclusive proof that marginal tax rates should fall rather than rise — there are lots of other factors at play — but it was a surprising and powerful argument, and one of the few that politicians did seem to absorb.

These three pillars have been standing for a while. So what’s new in the economics of taxation?

. . .

One answer: better data. The next generation of economists, people such as Raj Chetty and Amy Finkelstein, are drawing on data that the likes of Pigou and Ramsey could hardly have imagined. As a result they are able to blend ideas from mainstream economic theory with the psychological insights beloved of behavioural economists. It’s a pragmatic approach, depending on the problem at hand and what the data tell us.

A few years ago, Finkelstein looked at what happens when tollbooths offer electronic toll collection, allowing drivers to breeze through without fiddling for change. She found persuasive evidence that the electronic toll weighed less heavily in people’s minds — they forgot exactly what the price was and began to ignore it. Toll collectors, quite rationally, respond by raising the toll.

More recently, Chetty and co-authors tried to estimate whether the earned income tax credit (EITC) in the United States, a work-related subsidy paid to parents, encouraged people to work more. With a truly mind-boggling dataset boasting 78 million taxpayers and 1.1 billion income statements, they found that the EITC can work very well — if people know about it. In areas with lots of claimants, new parents tended to be well-informed and to respond to the EITC. In other areas, the EITC was not widely understood, and it was less effective.

Perhaps this new data-driven, psychologically realistic approach to tax will win political support. After all, Finkelstein discovered a tax that works best when concealed, while Chetty found a benefit payment that works best when widely trumpeted. Boasting about the good news and hiding the bad? That is the kind of economic theory that any politician can love.

Written for and first published at ft.com.