At last weekend’s European summit Tony Blair sent a message that has gone largely unnoticed – that armchair economists are wrong: orthodox incentives don’t always matter. On the eve of his departure the former Prime Minister had no conventional motive to cut a good deal in Brussels. A good settlement would not have increased his power, and a bad one would not have lessened his chances of staying in office.

Instead, his motives were patriotism, self-respect, a sense of how a prime minister should behave. And everybody seems to accept this. None of Mr Blair’s many critics claims that he would have got a good deal if only he’d had stronger incentives.

This shows a fact often lost on simplistic economists. People are not always driven by what the philosopher Alasdair MacIntyre called the goods of effectiveness – money, power, fame. Often, altruism or a sense of vocation (MacIntyre's goods of excellence) are enough.

Take the NHS. It is remarkable just how well it works, given that it is, as Patricia Hewitt, the former Health Secretary, said recently, even more centralised than the Cuban economy. Why does it do so well? Because doctors, nurses and (yes) managers are partly driven by what the LSE’s Julian Le Grand calls “knightly motives”.

Indeed, in some cases, conventional incentives can backfire. If you give a fund manager big incentives to hit particular targets, he might take absurd risks if he fears falling short of them.

A recent paper by Marco Daniele Paserman, of the Hebrew University of Jerusalem, shows that in women’s grand slam tennis tournaments, two fifths of the most important points end in unforced errors. Great golfers can miss important putts; footballers miss important penalties. People “choke” under pressure. Psychologists call it the Yerkes-Dodson law – performance worsens when incentives are high.

Economists at the Massachusetts Institute of Technology have confirmed this with some cunning experiments. They got Indian villagers to play some children’s games; Simon Says, throwing balls at targets and the like. They found that if they offered the villagers big money – a month’s wages – for doing well, their performance deteriorated. They became overmotivated. We should therefore be worried by Gordon Brown’s promise to try his utmost.

These are not the only ways in which incentives can go wrong. Steve Levitt, the author of Freakonomics, points to some kindergartens in Haifa, Israel. They had a problem with parents being late to pick up their toddlers. So they fined latecomers. And the numbers of them subsequently rose.

What happened? The same thing that researchers in New Zealand discovered when they found that two fifths of blood donors said they would stop giving blood if they were offered payments.

It’s something pointed out 50 years ago by the sociologist Richard Titmuss – financial incentives can be counterproductive if they crowd out altruistic motives. Parents picked up their children on time out of consideration to the kindergarten teachers, but when fines were introduced they saw them as a charge for looking after their children.

In other cases incentives simply conflict. A rise in wages encourages people to work longer. But it also means that they need to work less to pay the bills. That encourages them to work less.

These aren’t merely textbook examples. Researchers at the Institute for Fiscal Studies have estimated that, although working tax credits encouraged more people to work in general, they discouraged mothers in couples from working.

Though these examples show that incentives don’t always work, that does not mean that orthodox economics is wrong. Certainly, it is usually sensible to suppose that people are selfish. When Adam Smith said that it is not to the benevolence of the butcher, baker and brewer that we owe our dinner, but to self-interest, that was what he meant.

Instead, it shows that we should remember the caveat attached to orthodox economics. Yes, incentives matter at the margin. But the margin needn’t be particularly wide. And many people aren’t on it.

This matters. First, it means changes in prices can have small effects. A recent paper by economists at the University of Warwick suggests that the weak dollar will do little to reduce America’s huge trade deficit. This is because higher import prices and lower export prices don’t change trade volumes very much.

Secondly, supporters of free markets (of whom I'm one) forget that most people aren’t on the margin. The national minimum wage demonstrates this. Because it has destroyed fewer jobs than some predicted, its supporters have been able to claim that it works. In fact, if you look carefully, the minimum wage has led to small cuts in employment and hours. That confirms both parts of proper economics. Yes, incentives matter: the incentive to employ people has weakened. But they matter only at the margin: the effect on jobs and hours has been small.

Thirdly, it means there are dangers in public sector reform. Reforms that weaken the “knightly motives” of doctors and teachers may prove counterproductive.

Talk of incentives is often nakedly political. Chief executives need multimillion pound pay packets as an incentive to do a good job? Phooey. They get big money because they can – because they have the power to take money off shareholders and workers.

Married couples should have tax incentives to stay together? Sure, tax breaks will cause a few to get or stay married who otherwise wouldn’t. But mainly, such tax breaks would be a transfer from single people to married ones.

There’s more to motivation than many would have us believe.

Times Recommends

• In the real world, the public sector must pay
• Playing politics in a land of double Dutch
• Be like the plinth people: not set in stone