PsyFi Search

Tuesday 13 December 2011

When Incentives Go Bad

Strings Attached: Untangling the Ethics of Incentives by Ruth W. Grant

Way back in ’76, when heels were high, flares were wide and hair was long (for the men, at least) a couple of dudes called Michael Jensen and William Meckling proposed a solution to a problem that had vexed even the founders of economics. They attempted to resolve a puzzle that started with Adam Smith and which has ended in the crash to end all crashes; and as ever we’re looking at a set of consequences which was unforeseeable in advance but seems inevitable in retrospect.

The problem was how do you align the interests of the owners of corporations – the shareholders – with those of the managers of the companies? The solution was ingenious, but the chain of events it set off has exposed the nature of the concept of incentives. Because, it turns out, incentives are not neutral economic things, but strike at the very heart of what it means to be human.

Agency Revisited

Jensen and Meckling’s paper explored an economic concept that we’ve seen before, in On Incentives, Agency and Aqueducts, the issue of agency, where different parties with a similar economic interest have different goals. So, for instance, shareholders of a firm want to increase the value of their shareholding while the firm’s managers want to see their remuneration do the same. These are not necessarily aligned goals, because managers, who have day-to-day executive control, will tend to operate in a way which maximises their personal returns; and this may not equate to maximising the value of the company from a shareholder perspective.

Jensen and Meckling’s insight was that these competing interests could be aligned by ensuring that managers were also significant shareholders. This led to the growth of the culture of bonuses based on shares, share option schemes and cash rewarded for an increasing share price – the essence of maximising shareholder value. Unfortunately this makes a strong assumption about the nature of incentives, one that’s strongly entwined with modern theories of incentives and one that’s quite probably wrong. It assumes that there are no ethical issues associated with incentives.

Trades and Incentives

Technically economics assumes that an incentive is equivalent to a trade between two consenting parties, and trades are believed to be voluntary, so neither party needs to agree unless they want to. Which is all fine and good but it assumes that if I consent to an incentive it’s the right thing for me to do. At the heart of this is a rather odd point – that if this is true, why would I need an incentive at all? Indeed, as we’ve seen in Money Can't Buy Happiness, sometimes introducing monetary incentives decreases compliance, rather than improving it.

A fascinating book, Strings Attached: Untangling the Ethics of Incentivesby Ruth Grant, attempts to untangle the moral maze that underpins incentives. As she points out, there are times when treating incentives as a trade between two parties simply doesn’t work. A criminal could bribe a judge, which is an incentive but is also immoral as well as illegal. In particular she takes aim at the U.S. system of plea bargaining which she shows can never deliver what the justice system is intended for – justice. Either criminals get away with a lesser sentence than they deserve or the innocent get a harsher one: arguing that this makes economic sense misses the point, and is a perversion of the whole purpose of the system, in the name of cost-savings.

Moreover, not all incentives are open to being freely rejected. In some cases they are coercive – so a prosecutor may offer a plea bargain where the alternative to accepting it may be so harsh that the defendant really has no choice. Even worse, as Grant points out, some incentives should never be offered – again, if a prosecutor offers a plea bargain when they know their case is exceptionally weak this is an abuse of power.

External v Internal

At the heart of the argument is the point that incentives are always external – extrinsic motivators, as they’re called – and are not the same as intrinsic motivation, where we do things because we want to, not because we're induced to:
“Intrinsic motivation is a form of autonomy: self-directed behavior, not depending on external forces. When extrinsic and intrinsic motivation conflict, or when incentives conflict with autonomous action and judgement, either the intrinsic motivation and autonomy are “crowded out” or the incentive is rejected and backfires”.
Worse still, in heavily incentivised environments people:
“Can become desensitized, and the assumption of incentives – that people are primarily moved by calculations of extrinsic benefits and costs – becomes self-fulfilling. This is what it means to say that incentives can corrode institutional culture”.
This seems to rather accurately describe the system of bonuses that has spread through the financial system over the last thirty years, as executives have been “aligned” with the shareholders, to the extent that they’ve profited more by maximising shareholder value than the shareholders have. By accepting incentives, many of which were virtually impossible to turn down, financiers made a Faustian pact, accepting the structure of the world in terms of the models peddled by economists in order to meet the targets and reap the rewards of the incentives.

Incentives Multiplied

Worse still, this approach to incentives and choice is beginning to spill over into other areas of life. The use of incentives to encourage the behavior desired, rather than actually engaging in constructive persuasion, is becoming commonplace for governments as well as corporations. Unfortunately if incentives crowd-out autonomy they cause resentment, to the point where non-compliance sets in. As Bruno Frey has described:
“In many countries tax noncompliance is a rampant problem … Economists have a relatively straightforward suggestion to make: As the incentive to cheat negatively depends on expected punishment, the government should raise the probability of detection and/or the penalty for tax fraud”.
His research suggests that this doesn’t actually work: there’s a “crowding-out effect” where institutional incentives push out intrinsic motivation such as civic virtue or tax morale. Which is what happens when you define the world in terms of self-interest. Why would anyone expect anything else?

Power and Autonomy

Of course, this mental attitude cascaded down through chains of management in a morally corrosive manner. Here's Ruth Grant again:
“Incentives are also deceptive because, when governors and managers employ incentives, they are attempting to steer people’s choices in certain directions. But because incentives do leave people with a choice, managers remain in a position to deny the extent of their own control and to evade responsibility. This is another aspect of incentives that makes them an attractive tool from a manager’s point of view. They are an exercise in power than conceals power”.
For Grant incentives are not ethically neutral exchanges between people, they are exercises in power, where that power can range from simple bargaining to utter coercion, dependent on circumstances. Because of this voluntariness is not the only criteria necessary to determine whether incentives are ethical and all incentives must respect our individual rights to self-determination – to autonomy. Incentives designed to remove that autonomy are just plain wrong.

Rethinking Agency 

It’s a good book, and one that deserves to be widely read, even if you don’t agree with the conclusions. In short, though, these suggest that Jensen and Meckling were, at least partially, wrong. While incentivising managers to align them with shareholders’ interests addresses the agency problem at the heart of the modern corporation it also leads to a loss of any kind of moral compass. By removing individuals’ freedom of choice by introducing ever more powerful incentives we’ve created a self-fulfilling prophesy in which self-interest becomes the defining approach of the age and incentives are no longer optional.

Here’s the rub – if an incentive is so powerful it can’t be rejected it no longer offers a choice. And if an incentive comes without choice it’s no longer an incentive, its an exercise in power. If this is the beating heart of modern capitalism then it’s basically based on a fundamental deceit – and if we’re all being forced to live a lie in order to get those all important incentives why should anyone expect anyone in the corporate world to behave ethically?  Incentives are not the solution to agency problems, because their unintended, perverse, consequences create a different set of problems and a less just world.

Time for a rethink, methinks.


  1. This comment has been removed by a blog administrator.

  2. For those who have read Daniel Dennet's work on freewill, the result of compelling or irresistible incentives will be unsurprising.

    In his book "The Illusion of Conscious Will" Dennet opines that we have far less freewill than we imagine. However, he argues that the belief in our freewill is ESSENTIAL to our model of hour our thoughts lead to action.

    Even if we are misapprehended in our belief that conscious thoughts precede our actions, without a belief in freewill we cease to interpret our actions through the prism of our own personal agency. That is to say that our consciousness is more often engaged in rationalization or explanatory functions AFTER observing ourselves committing some act.

    When people perceive themselves to subject to determinism they cease to ask themselves the question 'why did I just make this choice'? This is problematic because the inquiry into the authorship of our actions is necessarily the origin of our moral compass.

    This has been well-tested empirically. An obvious example is to ask inmates where they caused or were responsible for their crimes. Almost invariably the answer is "no" because they externalise the agency. I did it because... Or I couldn't help it because.. It is not difficult to see why this approach is problematic. Acceptance of culpability for our actions is the cornerstone of morality.