If bonus or “incentive pay” schemes work so well for senior executives and bankers, why does everyone not get them? After all, many jobs involve making important decisions or taking risks. Is there anything about corporate decisions and financial risks that makes these categories of work special in terms of how they need to be incentivised and rewarded?

The conventional answer is that a bonus scheme or incentive plan will indeed encourage the recipients to make more money for the shareholders or clients on whose behalf they act. And while there were some unfortunate abuses in the past, the schemes have been altered to focus on longer-term shareholder returns and to enable clawbacks. However, the conventional answer does not address my question: there are many workers acting as agents on behalf of others (taxpayers, shareholders, clients), so why are bonus and incentive schemes not more common?

A classic paper on the “principal-agent problem” (how you encourage somebody else to act in your best interests rather than theirs when you hire them) by Bengt Holmstrom and Paul Milgrom pointed out that the conventional answer makes the mistake of assuming that jobs are simple and consist only of one task. In reality, agents — such as executives acting on behalf of shareholders — have multiple tasks with many dimensions. Some of these will be easier to measure than others.

As the two economists note, if you use incentive pay to get teachers, say, to raise pupils’ test scores, the test scores will go up, but other desirable, unmeasurable attributes such as intellectual curiosity or exploration of the arts will suffer. Better to pay a fixed salary to achieve effort across the whole range of desired outcomes. Despite their warning, we have, of course, ended up with incentive pay in education in the UK, albeit indirectly via school inspection criteria and national targets; pupils are being taught to pass tests to their detriment and the economy’s.

The point is that pay structures not only incentivise effort and direct risk-taking, they also determine the worker’s allocation of effort between different tasks. The design of pay schemes would be easier if jobs could be designed to separate the different tasks. But if not, the general conclusion is that incentive pay schemes are less desirable the harder it is to measure and monitor outcomes.

This clearly helps explain why bonus schemes and incentive plans tend to be tied to easily observable indicators such as the share price or a year’s trading profits. It also helps explain why such schemes so badly distort behaviour, incentivising short-term over long-term gains and quarterly share price increases over customer satisfaction.

The more complex the job, the more dimensions involved — as in being a corporate chief executive, say — the less justification there is for an incentive reward scheme. This is reinforced in the specific context of shareholder principals and chief executives, when the latter are responsible for the value of the assets they are managing on behalf of the company’s owners. Incentives linked to whatever can easily be measured lead agents to turn their efforts away from maintaining and enhancing the value of the asset over time.

Professor Holmstrom and Prof Milgrom thought they had done a good job of explaining many pay structures that were puzzling from the perspective of the simple view that people will work harder in exchange for pay that rewards the effort. For example, sales commission is found in industries where the output is easy to measure, but not so often in those where there is a quality dimension. But they went badly wrong when they diverted sales peoples’ efforts away from quality, as in financial services mis-selling. As the economy becomes increasingly complex and intangible, monitoring and measuring seem ever harder. The simplistic case for bonus and incentive pay schemes grows ever weaker.

Indeed, the best arrangement would seem to be the opposite of the pattern we observe now. Corporate executives and senior bankers doing complex jobs involving many impossible-to-monitor activities are the last people who ought to be paid via an incentive scheme; while bonuses for fast-food workers or shop-floor employees make more sense.

Shareholders are right to be rebelling against ludicrous executive pay packages. But the rebellion needs to go much further and demand fixed salaries for the most senior people. If the chief executive of a company seriously tells me, as a shareholder, that he will not put in as much effort as he otherwise would unless I link his pay to a handful of metrics, I have every reason to be doubtful about hiring him to do the job.