May not be redistributed commercially without the author's permission.

Index

SHOULD CORPORATIONS ENCOURAGE MORE RISK-TAKING?

Arnold Kling, "Arguing in My Spare Time", No. 16

June 22, 1998

Recently, I attended a conference organized by FastCompany, an interesting 2-year-old business magazine. Many of the attendees were middle managers of large corporations. Although I took no formal survey, I got the impression that most of them would agree with the following statements:

(a) Corporations should give middle managers more freedom to take risk.

(b) Corporations should be more willing to make mistakes and accept failure.

(c) Corporations should offer more rewards and incentives for innovation.

Indeed, it seemed to me that many middle managers came to the conference looking for ways to change the culture where they work in order to foster (a) - (c). They believe that this would lead to better corporate decision-making and higher profits. However, some basic economic analysis suggests otherwise.

Suppose we re-phrase (a) - (c) in terms of economic risk and reward. We might express them as:

(a') Corporations should enable middle managers to make larger bets with corporate resources than is the case currently.

(b') The downside risk of these bets should be borne more by the corporation and less by the middle manager than is the case currently.

(c') More of the upside of these bets should accrue to middle managers than is the case currently.

Put in these terms, it is easy to see why (a) - (c) sounds like the holy grail to middle managers. It also makes it easier to understand why shareholders may see it differently.

The general issue here is the alignment of incentives between the middle managers and shareholders relative to risk-taking. If the incentives are aligned properly, then the correct decisions will be made. Economists have coined the term "incentive compatibility" to describe this sort of issue.

The current incentive structure for middle managers can be described as follows:

(a1) Large bets are not permitted without multiple layers of approval.

(b1) Middle managers have limited personal downside risk if an initiative fails.

(c1) Middle managers have limited upside benefit from an innovation that succeeds.

This incentive structure encourages middle managers to promote innovations that have low risk and modest expected return. An example would be an initiative that could gain or lose $200,000 for the corporation, with an 80 percent chance of success.

An alternative incentive structure is the incentive structure for an entrepeneur:

(a2) Large bets require financing from a venture capitalist.

(b2) The entrepeneur has considerable downside risk, but some of the risk is borne by a venture capitalist.

(c2) The entrepeneur has considerable upside benefit, but some of the upside benefit is shared by the venture capitalist.

This incentive structure encourages entrepeneurs to promote innovations that have high risk and high expected return. An example would be an initiative that could gain or lose $10 million, with a 60 percent chance of success.

The incentive structure that middle managers want, as embodied in a' - c' above, also would promote innovations that have high risk and high expected return. However, in addition, it would encourage innovations that have high risk and low expected return. An example would be an innovation that could gain or lose $10 million, with a 20 percent chance of success. While such a bet obviously would be bad for the corporation, it would be good for the risk-incented middle manager, who gains something if the innovation works but loses nothing if it fails.

For corporations, encouraging middle managers to take good risks is not as easy as it sounds. Middle managers understandably do not want the same degree of personal downside risk as entrepeneurs. However, in the absence of personal downside risk, the middle manager's incentives would be skewed toward taking unjustifiable risks. Bureaucratic controls and limits on upside incentives may be an appropriate adaptation for correcting this potential bias.