Tuesday, 12 October 2010

Should we pay for performance?

Tom Powdrill has a question: why do companies think that higher pay leads to better performance?

He's right to point out that there are lots of studies now showing the opposite: higher pay can lead to worse performance. But these are mostly lab studies which don't replicate all the conditions of real life.

To explain why companies might think this, notice that the implicit basis of pay-for-performance is that there is a cost to the worker of doing good work. Otherwise, why would we need to pay them more to do it? So, is it true that workers bear a cost if they do well?

In traditional work, where the quantity of labour largely controls the value of the output, this is certainly true. Digging more coal takes effort and time, hence we might want to compensate the worker for doing so. One can make a similar argument for writing more lines of Java code, though that's more controversial - mainly because more is not necessarily better.

But Tom is talking about management decision making, where it's not the quantity of work done but the quality of decisions made that matters. So is there a cost to managers of making better decisions? Yes, actually.

If they need to do more research, try more alternatives and learn more things in order to make a better decision, that's a cost. All of those things take time, and a lazy manager could just do the first thing that comes to mind and go home.

If they need to simply spend time thinking things through - especially if they have lots and lots of decisions to make - that's also a cost. Think of the classic naive junior health minister, sitting up till 2am going through the red box and taking shortcuts. Then think of the more cynical, less junior health minister who just signs what the civil servants put in front of her and goes to the pub.

In a few cases, the cost of making a decision that's good for the company is the option to make an alternative choice that's good for you. Especially in finance, where there are lots of opportunities to trade on your own account or siphon off money - high, performance-related pay is meant to incentivise you to act in your employer's interest and not your own.

And there's a more subtle cost to making good decisions on behalf of one company: the opportunity cost of making good decisions at the other company you could have worked for. If company B is willing to pay you well to make good decisions - simply because you're naturally talented at making them - then company A had better pay you good money too.

This last point explains why pay might be high, not why it might be performance related. But that's quite easily explained by a related argument. Companies offer performance-related pay because that lets the good employees select themselves for the job. If you're not willing to be paid per (goal scored/extra penny on the share price/lawsuit won) then maybe you're not so confident in your abilities after all. And who'd know your own abilities better than you?

These are all good rational explanations for performance related pay. And most of them are not reflected in the lab experiments which argue against it. So what of the cognitive and behavioural aspects?

Well, managers may not always know whether their decisions are good. If this is the case, then all of the above arguments are weakened.

And, as those experiments show, high stakes sometimes lead to counterproductive behaviour, probably due to the distraction of the potential rewards.

Tom mentions Deci and Ryan: undoubtedly intrinsic motivation is important to most people, and money can displace some of that, potentially diminishing the desire (and ability?) to do a good job.

So while the rationale for performance-related pay is clear, the counterarguments are evident too. Deci has attempted to measure the comparative effects (for example in "Experiment II" on this page) but the main result shows that when extrinsic motivation is added then removed, intrinsic motivation can fall. I don't know of any good attempt to measure or estimate the strength of the countervailing effects during the time that the monetary reward is provided.

Unfortunately this doesn't give us a clear conclusion (though I should probably obey blogging protocol by manufacturing one). There are good arguments for and against performance pay; some situations will call for it and others won't.

2 comments:

Min said...

"why do companies think that higher pay leads to better performance?"

Is that why they "reward" poor performance with high compensation? ;) To encourage better performance next time?

Seriously, the existence of the "glass ceiling" seemed to me to indicate the threshold above which compensation exceeds the value of performance. You do not hear about the glass ceiling much anymore, but that does not mean that the threshold has ceased to exist. It separates those who are "members of the club" from those who are not. No?

Min said...

As to why I thought that the glass ceiling indicated a threshold above which compensation was greater than the value of the person to the company, it was because that explains why the company could afford to keep employees below the glass ceiling. If they let them above the ceiling, they would be paying them too much.

And why a ceiling or threshold instead of gradually increasing overpayment? That has to do with the structure of groups, in-groups and out-groups. In short, in any sufficiently large but still cohesive group, there is at least one internal boundary. Club membership also seems to extend across corporations. :)