Useful ‘Nobel’ Economics? You bet!
How a winner of the Economics ‘Nobel’ prize may help run organizations better
They say the way to hell is paved with good intentions. Arguably, anyone intent on paving the way to hell might find more paving material if they went for the unintended consequences of well-intentioned measures. There are several classic examples from history of incentives that went spectacularly wrong; one of these has even been used to provide a label for this problem: the Cobra effect.
Of cobras and contracts
When India was still a British colony, the governor of Delhi was worried about the high number of cobras in the city. Rather than have his staff address the apparent problem, he decided to rely on the services of the population and pay a bounty for every cobra skin that was brought in. This incentive initially seemed to work, but the number of skins brought in kept going up and up — because some enterprising people had begun to breed cobras. They sold these to the locals who then killed them and went to collect the bounty for the skins. Before long, the bounty scheme was stopped, leaving the cobra breeders with a sizeable population of cobras they could no longer sell. Unhindered by scruples, these surplus serpents were released, leaving Delhi with a bigger cobra problem than ever before.
Interestingly, the term cobra-effect was made popular by an economist, Horst Siebert, in his book Der Kobra-Effekt: Wie man Irrwege der Wirtschaftspolitik vermeidet (The Cobra effect: how to avoid mistakes in economic policy). And even more interestingly, it is also two economists who recently won the Economics “Nobel” prize, Bengt Holmström and Oliver Hart, whom we have to thank for a theory that can help avoid some of such problems.
Their discipline, contract theory, sounds perhaps more academic than it really is. It studies the widespread, everyday situations where market exchanges between two parties take place over a longer period of time, as opposed to one-off situations. Both parties have expectations about future mutual performance, which effectively form a contractual relationship. Some of this may be formalized in an actual explicit contract, some may be more implicit. A typical application area is that of employment — which is what Bengt Holmström has been mostly focusing on.
Few people would deny that it makes sense to pay employees according to their performance, i.e. how well they do their job. But is the visible output or outcome of what they are hired to do the best way to determine this? It is tempting to think so, and often that is exactly what happens. Quantifiable targets and hard objectives figure widely in how people’s performance is assessed at work. Yet this ignores the fact that in the real world the result of a staff member’s efforts is only partially attributable to their own actions. Much of that result is due to external circumstances not in the employee’s control, or to the contributions of others.
The less stable and predictable someone’s job environment, the more likely this kind of approach is to reward the employee for good luck and penalize them for bad luck. A clever employee would understand the risk of delivering poor performance as a result of bad luck, so they would demand a lot of money, either as a base pay or as a bonus when they’re lucky, in order to compensate for this risk. And of course, the more their performance depends on luck, the less hard they would work, as it would not make any difference anyway.
Problems behind the bar
To illustrate the kind of problems Holmström has studied, imagine working behind the bar in an establishment at a tourist hotspot. On sunny days, there is plenty of custom because there are many thirsty visitors; when the weather is bad, there are only few. If you are paid according to the drinks you serve, you can probably see how you would indeed not be particularly motivated to do a good job (your income would depend mostly on the weather, not on how good you are at barkeeping), and how you would expect to be paid a lot on busy days to counteract the lousy income on slow days. This is not a good state of affairs.
And there are further problems. Imagine the drinks you can sell are procured by a colleague who is quite haphazard in replenishing the stock, which means you will often run out of the popular beverages on busy days. So customers will go next door, and as you are being paid according to your turnover, you are penalized for the poor work of your colleague. Or imagine that on busy days, you get extra help from a waiter who will serve customers at their tables to allow you to concentrate on pouring the drinks. On these days, you and the waiter share the bonus 50/50 — but the waiter, being a lazy dude, spends his time chatting up the customers, while you need to jump in to serve customers and do much of his work. With this kind of performance incentive for the team, the temptation to free-ride is a real risk. Not a good state of affairs.
Many employees also have multiple tasks that require them to make trade-offs. Even if your boss has seen the light and pays you a decent fixed salary with a modest share in the profits on busy days, you still need to divide your time between serving customers and say, rinsing the glasses or wiping the counter. But if your bonus rewards the amount of drinks you sell, you may be tempted to concentrate entirely on pouring drinks and serve them in dirty glasses on a sticky counter. This will work in the short term, but it risks eventually putting off customers and tarnishing the bar’s reputation. This is also not a good state of affairs, certainly not for the bar owner.
This simple example illustrates how precarious it can be simply to use what is easy to measure as an indicator of actual performance, and reward people accordingly. In Moral hazard and observability, a classic paper from 1979, Holmström provides a comprehensive analysis of the kind of problem that is inherent in situations where only the payoff of a transaction is visible. Contracts that look no further are “second-best”, he says, and he shows that additional information, even if it is imperfect can significantly improve contracts (whether explicit or implicit).
Economic science vs Management science
Economics does show here that it is a social science rather than an exact one, though: unfortunately contract theory does not provide simple and unambiguous rules that tell us how to construct an ideal contract. But it does show the way to a better understanding of our quirky (but intelligent) human behaviour, and therefore to avoid falling into the trap of relying on all too simplistic interventions to influence it.
And it also shows that sometimes economists can successfully challenge respected management thinkers. You know, “What gets measured gets done” — even the great Tom Peters swears by it. But maybe just getting done what gets measured because it’s easy to measure it, is not quite what we want. We may end up with lots of cobra skins, but be no closer to solving the viper problem.
We had better acknowledge that much of what people do is not easy to measure, yet it can be at least as important as what is measurable. Holmström tells to look beyond the outcomes and seek to understand the inherent value of what someone does. That is arguably better advice than Tom Peters’.
And in this respect, Holmström’s contributions are supported by his behavioural economist colleagues. In Misbehaving, Richard Thaler says,
“Good leaders must create environments in which employees feel that making evidence-based decisions will always be rewarded, no matter what outcome occurs.”
This is a clear plea for looking at how people make decisions, rather than just to what the result of those decisions is. Dan Ariely thinks along the same lines. In an answer to a reader’s question whether it is wise to fire people who fail, he writes:
“Organizations reward and punish people based on the outcome of decisions, not on the quality of decisions. In general, you’d hope that good decisions would lead to good outcomes, but that causal link rests on probabilities, not certainties — so reward and punishment are often misapplied.”
We people may be, ultimately, simple — but we’re not that simple. We’d better bear that in mind.
Note: the scare quotes around the term ‘Nobel’ throughout this text indicate the fact that, as some pedants like to point out, this is not a proper Nobel prize, but the Prize in Economic Science in Memory of Alfred Nobel — here is more info on its origins.
Originally published at koenfucius.wordpress.com on October 16, 2016.
(Thanks for reading this far. If you really liked it, do consider clicking or tapping the heart below to recommend it to others!)