DURING the Battle of Britain, German fighter pilots were paid a competitive bonus for each enemy Spitfire or Hurricane they shot down.
Sounds sensible? Luckily for us, it was a disaster. The aces behaved like prima donnas, the non-aces were demoralised and resentful, and they were all diverted from their strategic job of protecting bombers. Still worse (or better), they systematically overestimated the number of British kills, giving German high command a falsely optimistic picture of the overall situation, perhaps even changing the course of the battle.
Almost everyone instinctively accepts that pay should reflect contribution rather than simple presence. With many private-sector companies won over, spreading performance pay to the public sector has become an article of faith for public service reform. Next week's strike by job centre and benefits staff, the bitterest civil service dispute in a decade, is partly about performance pay, as is the stand-off between employers and university teachers.
Yet despite its near ubiquity, there is strikingly little evidence that pay for performance has the desired effects in practice and plenty that it is often counterproductive. As the Second World War example illustrates, incentives are all too often divisive and highly vulnerable to the law of unintended consequences.
Making them dependent on appraisal - a favourite but truly toxic combination - ensures that results are arbitrary and subject to favouritism, as does the assumption that performance can be judged in isolation from the system as a whole. Ed Lawler, an acknowledged guru of high performance, has one word to say of merit-based pay assessment: 'Don't'.
So why do managers so obstinately believe in it? The most obvious reason is that rewards do alter behaviour - so it's easy for managers to believe that they have fixed the problem, or they will have after another tweak to the incentives. What is less obvious is the damage they do in the long term by, in effect, managing people by bribes.
Performance-related rewards derive ultimately from behaviourist experiments with rats in labs. In this context, rewards are simply the opposite side of the coin of punishment - as Alfie Kohn pointed out in a still-to-be-bettered Harvard Business Review article on the subject a decade ago: 'Do this and you'll get pounds 100' is no different in kind from 'Do this and you'll get a good kicking'. Both are manipulative and controlling, classic manifestations of command-and-control management. And just like punitive management, management-by-reward produces movement rather than motivation, compliance rather than commitment. Kohn says: 'Do rewards motivate people? Absolutely. They motivate people to get rewards'.
So what should a good pay system look like? Pay being incapable of motivating on its own, it follows from the above that one of its most important jobs is actually negative: not to demotivate. That has some significant implications. For the individual, fairness, openness and simplicity become the most important characteristics.
At the same time, there's no denying that pay is a powerful shaper of organisational culture. The snag is that the traditional hierarchical job-based pay systems that cemented bureaucratic organisations can do little to foster learning new skills or encourage teamworking - the main prerequisites of today's high-performing organisations.
From the organisation's point of view, therefore, the need is also negative: to eliminate the pinch-points where current pay policies work against desired results such as customer focus and improving the system.
Some companies are basing part of their pay policy on competence rather than the job, so people are paid more as they acquire extra skills.
Often both cases are served by dispensing with individual incentives, whose 'gains' can easily be wiped out by their disadvantages for the wider organisation: reinforcing barriers among functional areas and business units, demotivating others, promoting competition rather than teamwork. The latter is a powerful reason why performance pay is so strongly resisted in professions like teaching.
Crudely, incentives often cause people to do things that benefit themselves but harm the company. Alternatively, they spend a disproportionate amount of effort on one aspect of the work to the detriment of others - a particular problem where jobs are complex - making performance measures hard to set up.
Incentives are particularly damaging where they take the form of payments triggered by reaching fixed targets. Incentives of this kind trigger intense political lobbying to negotiate achievable figures and then cheating of one sort or another to hit them. Success becomes an internal matter of satisfying the boss rather than an external one of satisfying the customer.
To change this inward-looking focus, some large companies are throwing out pay systems that rely on fixed performance contracts. Instead they judge performance relative to competitors and pay appropriate bonuses retrospectively, allowing for special circumstances to be taken into account.Typically, the bonus is determined by the performance of the business unit rather than the individual subsidiary, so that people have an incentive to work for the greater good.
The disadvantage is that the larger the unit, the less opportunity individuals have to make a difference.
The inconvenient truth is that pay is both extremely important and a blunt instrument. It can make things go horribly wrong but by itself can't make them go beautifully right. Individual incentives are the last thing an employer should want for multi-dimensional jobs needing lots of teamwork. Don't throw the straightforward salary out just yet.
The Observer, 11 April 2004