A Benefit of Membership

The Irony of Punishment

New research measures the impact of reward and punishment.

Meet Fred. Heís a new owner of an existing KFC franchise restaurant. Itís his first day on the job, and as he introduces himself to his new employees, he imagines that each one of them is as hard working as he is. One by one, his people smile and tell him theyíre glad to have him. ďThis will be a snap,Ē Fred muses to himself, and that evening, he gives his wife a positive report. Heís glad to have such a good staff, and he tells her that he doesnít expect to use punishment in managing this group of people. A month later, he may be singing a different tune. The honeymoon for new supervisors ends when they recognize free-riding for the first time.

Most jobs are organized to allow flexibility. An absence of one person should not close the business merely because the absent person is the only one who can carry out an essential task. But flexibility encourages free-riding, and often, itís the supervisor himself who has to step up and carry out a task that someone else should be doing. Thatís when punishment becomes more attractive.

Punishment takes many forms: a scowl, a critical remark, a good chewing out, a written warning, a deduction from pay, a day off without pay, and ultimately, termination. None of these are nice, but eventually, supervisors will use them all. Fred will worry about this and agonize about adding punishment to his management repertoire, but now, there is research to guide him. Ozgur Gurerk, from the Laboratory of Experimental Economics at the University of Erfurt in Germany conducted the study.

Gurerk set up an elaborate laboratory experiment using 20 groups with 6 people in each group. He randomly assigned one to be supervisor and gave each supervisor the freedom to adopt either a reward or a punishment incentive system: small additions to or subtractions from their monetary compensation. Next, he turned the groups loose on a 10-task assignment. Supervisors used incentives after each task to encourage desirable behavior. He repeated this experiment two more times keeping the groups together and giving the supervisors the opportunity to change the incentive system at the beginning of each new set of tasks.

Just as Fred began positively with high hopes for cooperative, productive efforts from his employees, 19 of the 20 supervisors in Gurerkís experiment also adopted positive incentive systems in the first set of tasks. Near the end of this first set, free-riding appeared, and where it was most pronounced, supervisors switched to punishment incentive systems in the second set of tasks and began handing out punishments to free-riders. It didnít take long for the free-riders to change their ways. For many, it only took the threat of punishment with the new incentive system to correct the free-riding.

Gurerk devised a way to measure each groupís response to their incentive system. For groups whose contribution had fallen and whose leaders switched to a punishment incentive system, the immediate increase amounted to 60%, and that surprised even Gurerk.

With the third repetition of the experiment, five more supervisors switched to a punishment system, but some who had switched to punishment in the second set switched back to positive. Gurerk was measuring all the reactions.

The five groups who found themselves in a punishment incentive system in the final set of tasks also responded with a strong improvement of their performance. They increased 28%. Those that switched to punishment for the second set and remained negative in the third set posted the highest performance of all. The poorest contributions came from groups who switched from punishment in the second set to reward in the third. Their performance actually declined.

In looking over his work, Gurerk made some observations that are helpful.

Reward systems reveal a curious paradox. Supervisors must remember to deliver the rewards, and even if they never fail to provide proper rewards for productive behavior, rewards quickly lose their attractiveness. They become expected, and receiving them is taken for granted. But when an expected reward doesnít show up, i.e. when a supervisor neglects to provide the expected reward, then this absence of reward is experienced as a punishment. Ironically, reward incentive systems are mostly experienced as punishment for proper behavior.

Punishment incentive systems, on-the-other-hand, require much less action from the supervisor. He only needs to act when free-riding demands it. The rest of the time, when employee conduct is constructive, the supervisor doesnít need to do any punishing, and thatís experienced as a reward. Curiously, punishment incentive systems, where punishment is rarely used, are mostly experienced as reward for proper behavior. The reward is merely the lack of punishment.

Not all of Gurerkís groups developed free-riders. A few groups displayed consistently positive performance and good cooperation, and they responded well to a reward incentive system. So the incentive system a supervisor adopts should respond to the needs of the employee group. Groups with free-riders need, want, and appreciate a punishment incentive system to correct free-riding. Groups without free-riders do best with reward. - J.E. Larsen, Ph.D.

Reference: Gurerk, Ozgur, Bernd Irlenbusch, and Bettina Rokenbach (2009) Motivating Teammates: The Leaderís Choice Between Positive and Negative Incentives. Journal of Economic Psychology, 30(20009), 591-607.