Undercover Economist

Does the altruism theory help anyone at all?

People respond to incentives, so if you want something done, reach for your wallet.That’s what you’d expect an economist to say, but it is a belief that infuriates many commentators.

I will concede that offering cash is not always productive. In the days when I was young, free and single, I was never tempted to try to seduce cute girls at parties by slipping them a couple of crisp twenties. (Perhaps I should have done it. It is not as if my hit rate on an unpaid basis was particularly good either.)

Yet many policy wonks believe not just that there are some things that money can’t buy, but that cash incentives are counterproductive and even morally corrosive. The touchstone of this school of thought is Richard Titmuss’s book The Gift Relationship, published in 1970.Titmuss’s most memorable and influential claim was that the British system of voluntary blood donation led to better outcomes – healthier blood, supplied in a more timely fashion – than the American system of paying blood donors.

Titmuss’s argument was influential. In an example made famous by Freakonomics, when parents at an Israeli kindergarten were fined a small amount for showing up late to collect their children, their punctuality actually declined. The behavioural scientists Dan Ariely and James Heyman asked experimental subjects to perform a boring task; those paid a few cents did less work than those paid nothing at all.

Another experiment showed that when kindergarten students were given little awards when they drew with crayons, they tended to refuse thereafter to draw for nothing. Who colours in for free, anyway?

And there is even the evergreen idea that by underpaying nurses, we might attract the right kind of people – those with a vocation.

I can’t help feeling that we believe in the altruism thesis so strongly because it feels like the way the world should work. But the empirical support for the belief is thinner than we like to think. In the Ariely-Heyman experiment, a payment as low as five dollars was enough to remotivate their subjects. It was only the laughably small payments that caused problems. Bear in mind, too, that Ariely and Heyman were not trying to test a hypothesis about payment for performance, which is why they did not make the reward conditional on results.

Attachment to the Titmuss hypothesis has the power to cause harm. The psychologist Barry Schwartz used the kindergarten experiment to excoriate an experimental New York schools programme which paid older children to show up and work hard. It was depressing to see a write-up of a small experiment being used to argue that New York should not run a large and realistic one.

The “cheap nurses are good nurses” thesis is demonstrably false; in the British health service, the economists Emma Hall, Carol Propper and John van Reenen have shown that lower real wages for nursing staff mean more temporary staff, more overpromoted staff – and more patient deaths. As for blood donation, Titmuss’s thesis is far less pressing now that better blood-screening techniques have been developed. It is not clear how solid the idea was, since he himself complained about the lack of good data. But perhaps he was right that paying for blood was counterproductive.

Still, it is interesting to see a new study by the economists Nicola Lacetera, Mario Macis and Robert Slonim concluding that paying for blood increases the quantity donated without lowering the quality. Distasteful it may be, but sometimes the way to get results is to pay for them.

Also published at ft.com.