Big bonuses for bosses can have adverse effects upon productivity. A new paper by Jorg Oechssler, Anwar Shah and Nikos Nikiforakis concludes:

Managerial bonuses have the potential of generating conflict between managers and their subordinates. Managerial bonuses can be a disincentive for subordinates...firms should exercise caution when using high-powered incentives for managers. The benefits from managers' increased motivation need to be weighed carefully against the adverse effects for other employees.

This is based upon laboratory experiments in which a manager and subordinate cooperate on a project: they found that when the manager is offered a big bonus, subordinates' effort levels drop.

This might well have external validity, given that in the real world managerial oversight of subordinates is less tight than it was in the lab. As HSBC's CEO Stuart Gulliver said, "Can I know what every one of 257,000 people is doing? Clearly I can't."

But of course, this is but one of several pieces of evidence that big individual bonuses can have adverse effects. Experiments (pdf) by Uri Gneezy and colleagues have shown that they can reduce performance because people can be over-motivated and so crack under pressure. Other experiments have found that bonuses for traders can encourage asset price bubbles, or that collective bonuses can reduce effort by encouraging free-riding. It's also possible that bonuses can crowd out intrinsic (pdf) motivations (pdf) and encourage short-termism and earnings manipulation. And Nobel laureate jean Tirole has argued that bonuses can cause "signifi…cant efficiency losses" by over-incentivizing some roles and under-incentivizing others: traders tend to get big bonuses, risk managers not so much - so guess what happens?

I suspect that the main justification for bonuses is not so much that they elicit effort but rather that they are a form (pdf) of efficiency wage: bosses and CEOs who cannot be effectively monitored must be bribed handsomely not to steal the firm's assets.

Subject to this caveat, all this implies that bonuses can be inefficient. Insofar as they contribute to inequality, this in turn is more evidence that inequality might reduce productivity. How lucky it is for the rich, therefore, that our politicians don't care about productivity.

Another thing: if bonuses do reduce productivity, it's not obvious that the solution is to tax them more heavily; it could be that it's pre-tax bonuses that matter, to the extent that these are associated with individuals' sense of self-worth and intra-firm perceptions of fairness.