This article, updated for AME, needs no introduction.(n1) Even today, the original article is still widely reprinted. Now part of the lexicon, it truly qualifies as an Academy of Management Classic. For almost twenty years, its title has reminded executives and scholars alike--"it's the reward system, stupid!" We hope you enjoy the update!
Whether dealing with monkeys, rats, or human beings, it is hardly controversial to state that most organisms seek information concerning what activities are rewarded, and then seek to do (or at least pretend to do) those things, often to the virtual exclusion of activities not rewarded. The extent to which this occurs of course will depend on the perceived attractiveness of the rewards offered, but neither operant nor expectancy theorists would quarrel with the essence of this notion.
Nevertheless, numerous examples exist of reward systems that are fouled up in that the types of behavior rewarded are those which the rewarder is trying to discourage, while the behavior desired is not being rewarded at all.
Official goals are "purposely vague and general and do not indicate ... the host of decisions that must be made among alternative ways of achieving official goals and the priority of multiple goals ..."(n2) They usually may be relied on to offend absolutely no one, and in this sense can be considered high acceptance, low quality goals. An example might be "All Americans are entitled to health care." Operative goals are higher in quality but lower in acceptance, since they specify where the money will come from, and what alternative goals will be ignored.
The American citizenry supposedly wants its candidates for public office to set forth operative goals, making their proposed programs clear, and specifying sources and uses of funds. However, since operative goals are lower in acceptance, and since aspirants to public office need acceptance (from at least 50.1 percent of the people), most politicians prefer to speak only of official goals, at least until after the election. They of course would agree to speak at the operative level if "punished" for not doing so. The electorate could do this by refusing to support candidates who do not speak at the operative level. Instead, however, the American voter typically punishes (withholds support from) candidates who frankly discuss where the money will come from, rewards politicians who speak only of official goals, but hopes that candidates (despite the reward system) will discuss the issues operatively.
If some oversimplification may be permitted, let it be assumed that the primary goal of the organization (Pentagon, Luftwaffe, or whatever) is to win. Let it be assumed further that the primary goal of most individuals on the front lines is to get home alive. Then there appears to be an important conflict in goals--personally rational behavior by those at the bottom will endanger goal attainment by those at the top.
But not necessarily! It depends on how the reward system is set up. The Vietnam war was indeed a study of disobedience and rebellion, with terms such as "fragging" (killing one's own commanding officer) and "search and evade" becoming part of the military vocabulary. The difference in subordinates' acceptance of authority between World War II and Vietnam is reported to be considerable, and veterans of the Second World War were often quoted as being outraged at the mutinous actions of many American soldiers in Vietnam.
Consider, however, some critical differences in the reward system in use during the two conflicts. What did the GI in World War II want? To go home. And when did he get to go home? When the war was won! If he disobeyed the orders to clean out the trenches and take the hills, the war would not be won and he would not go home. Furthermore, what were his chances of attaining his goal (getting home alive) if he obeyed the orders compared to his chances if he did not? What is being suggested is that the rational soldier in World War II, whether patriotic or not, probably found it expedient to obey.
Consider the reward system in use in Vietnam. What did the soldier at the bottom want? To go home. And when did he get to go home? When his tour of duty was over! This was the case whether or not the war was won. Furthermore, concerning the relative chance of getting home alive by obeying orders compared to the chance if they were disobeyed, it is worth noting that a mutineer in Vietnam was far more likely to be assigned rest and rehabilitation (on the assumption that fatigue was the cause) than he was to suffer any negative consequence.
In his description of the "zone of indifference," Barnard stated that "a person can and will accept a communication as authoritative only when ... at the time of his decision, he believes it to be compatible with his personal interests as a whole.(n3) In light of the reward system used in Vietnam, wouldn't it have been personally irrational for some orders to have been obeyed? Was not the military implementing a system which rewarded disobedience, while hoping that soldiers (despite the reward system) would obey orders?
Theoretically, physicians can make either of two types of error, and intuitively one seems as bad as the other. Doctors can pronounce patients sick when they are actually well (a type 1 error), thus causing them needless anxiety and expense, curtailment of enjoyable foods and activities, and even physical danger by subjecting them to needless medication and surgery. Alternately, a doctor can label a sick person well (a type 2 error), and thus avoid treating what may be a serious, even fatal ailment. It might be natural to conclude that physicians seek to minimize both types of error.
Such a conclusion would be wrong. It has been estimated that numerous Americans have been afflicted with iatrogenic (physician caused) illnesses.(n4) This occurs when the doctor is approached by someone complaining of a few stray symptoms. The doctor classifies and organizes these symptoms, gives them a name, and obligingly tells the patient what further symptoms may be expected. This information often acts as a self-fulfilling prophecy, with the result that from that day on the patient for all practical purposes is sick.
Why does this happen? Why are physicians so reluctant to sustain a type 2 error (pronouncing a sick person well) that they will tolerate many type 1 errors? Again, a look at the reward system is needed. The punishments for a type 2 error are real: guilt, embarrassment, and the threat of a malpractice suit. On the other hand, a type 1 error (labeling a well person sick) is a much safer and conservative approach to medicine in today's litigious society. Type 1 errors also are likely to generate increased income and a stream of steady customers who, being well in a limited physiological sense, will not embarrass the doctor by dying abruptly. Fellow physicians and the general public therefore are really rewarding type 1 errors while hoping fervently that doctors will try not to make them.
A current example of rewarding type 1 errors is provided by Broward County, Florida, where an elderly or disabled person facing a competency hearing is evaluated by three court-appointed experts who get paid much more for the same examination if the person is ruled to be incompetent. For example, psychiatrists are paid $325 if they judge someone to be incapacitated, but earn only $125 if the person is judged competent. Court-appointed attorneys in Broward also earn more--$325 as opposed to $175--if their clients lose than if they win. Are you surprised to learn that, of 598 incapacity proceedings initiated and completed in the county in 1993, 570 ended with a verdict of incapacitation?(n5)
Society hopes that professors will not neglect their teaching responsibilities but rewards them almost entirely for research and publications. This is most true at the large and prestigious universities. Cliches such as "good research and good teaching go together" notwithstanding, professors often find that they must choose between teaching and research-oriented activities when allocating their time. Rewards for good teaching are usually limited to outstanding teacher awards, which are given to only a small percentage of good teachers and usually bestow little money and fleeting prestige. Punishments for poor teaching are also rare.
Rewards for research and publications, on the other hand, and punishments for failure to accomplish these, are common. Furthermore, publication-oriented resumes usually will be well-received at other universities, whereas teaching credentials, harder to document and quantify, are much less transferable. Consequently it is rational for university professors to concentrate on research, even to the detriment of teaching and at the expense of their students.
By the same token, it is rational for students to act based upon the goal Displacement(n6) which has occurred within universities concerning what they are rewarded for. If it is assumed that a primary goal of a university is to transfer knowledge from teacher to student, then grades become identifiable as a means toward that goal, serving as motivational, control, and feedback devices to expedite the knowledge transfer. Instead, however, the grades themselves have become much more important for entrance to graduate school, successful employment, tuition refunds, and parental respect, than the knowledge or lack of knowledge they are supposed to signify.
It therefore should come as no surprise that we find fraternity files for examinations, term paper writing services, and plagiarism. Such activities constitute a personally rational response to a reward system which pays off for grades rather than knowledge. These days, reward systems--specifically, the growing threat of lawsuits--encourage teachers to award students high grades, even if they aren't earned. For example:
When Andy Hansen brought home a report card with a disappointing C in math, his parents ... sued his teacher.... After a year and six different appeals within the school district, another year's worth of court proceedings, $4000 in legal fees paid by the Hansens, and another $8500 by the district ... the C stands. Now the student's father, auto dealer Mike Hansen, says he plans to take the case to the State Court of Appeals.... "We went in and tried to make a deal: They wanted a C, we wanted an A, so why not compromise on a B?" Mike Hansen said. "But they dug in their heels, and here we are."(n7)
It is axiomatic that those who care about a firm's well-being should insist that the organization get fair value for its expenditures. Yet it is commonly known that firms seldom bother to evaluate a new TQM, employee empowerment program, or whatever, to see if the company is getting its money's worth. Why? Certainly it is not because people have not pointed out that this situation exists; numerous practitioner-oriented articles are written each year on just this point.
One major reason is that the individuals (in human resources, or organization development) who would normally be responsible for conducting such evaluations are the same ones often charged with introducing the change effort in the first place. Having convinced top management to spend money, say, on outside consultants, they usually are quite animated afterwards in collecting rigorous vignettes and anecdotes about how successful the program was. The last thing many desire is a formal, revealing evaluation. Although members of top management may actually hope for such systematic evaluation, their reward systems continue to reward ignorance in this area. And if the HR department abdicates its responsibility, who is to step into the breach? The consultants themselves? Hardly! They are likely to be too busy collecting anecdotal "evidence" of their own, for use on their next client.
Most coaches disdain to discuss individual accomplishments, preferring to speak of teamwork, proper attitude, and one-for-all spirit. Usually, however, rewards are distributed according to individual performance. The college basketball player who passes the ball to teammates instead of shooting will not compile impressive scoring statistics and is less likely to be drafted by the pros. The ballplayer who hits to right field to advance the runners will win neither the batting nor home run titles, and will be offered smaller raises. It therefore is rational for players to think of themselves first, and the team second.
Consider the cost-plus contract or its next of kin, the allocation of next year's budget as a direct function of this year's expenditures--a clear-cut example of a fouled up reward system. It probably is conceivable that those who award such budgets and contracts really hope for economy and prudence in spending. It is obvious, however, that adopting the proverb "to those who spend shall more be given," rewards not economy, but spending itself.
The past reward practices of a group health claims division of a large eastern insurance company provides another rich illustration. Attempting to measure and reward accuracy in paying surgical claims, the firm systematically kept track of the number of returned checks and letters of complaint received from policyholders. However, underpayments were likely to provoke cries of outrage from the insured, while overpayments often were accepted in courteous silence. Since it was often impossible to tell from the physician's statement which of two surgical procedures, with different allowable benefits, was performed, and since writing for clarifications would have interfered with other standards used by the firm concerning percentage of claims paid within two days of receipt, the new hire in more than one claims section was soon acquainted with the informal norm: "When in doubt, pay it out!"
This situation was made even worse by the firm's reward system. The reward system called for annual merit increases to be given to all employees, in one of the following three amounts:
- If the worker was "outstanding" (a select category, into which no more than two employees per section could be placed): 5 percent
- If the worker was "above average" (normally all workers not "outstanding" were so rated): 4 percent
- If the worker committed gross acts of negligence and irresponsibility for which he or she might be discharged in many other companies: 3 percent.
Now, since the difference between the five percent theoretically attainable through hard work and the four percent attainable merely by living until the review date is small, many employees were rather indifferent to the possibility of obtaining the extra one percent reward. In addition, since the penalty for error was a loss of only one percent, employees tended to ignore the norm concerning indiscriminant payments.
However, most employees were not indifferent to a rule which stated that, should absences or latenesses total three or more in any six-month period, the entire four or five percent due at the next merit review must be forfeited. In this sense, the firm was hoping for performance, while rewarding attendance. What it got, of course, was attendance. (If the absence/lateness rule appears to the reader to be stringent, it really wasn't. The company counted "times" rather than "days" absent, and a ten-day absence therefore counted the same as one lasting two days. A worker in danger of accumulating a third absence within six months merely had to remain ill--away from work--during a second absence until the first absence was more than six months old. The limiting factor was that at some point salary ceases, and sickness benefits take over. This was usually sufficient to get the younger workers to return, but for those with 20 or more years' service, the company provided sickness benefits of 90 percent of normal salary, tax-free! Therefore ....).
Thanks to the U.S. government, even the reporting of wrongdoing has been corrupted by an incredibly incompetent reward system that calls for whistleblowing employees to collect up to thirty percent of the amount of a fraud without a stated limit. Thus prospective whistleblowers are encouraged to delay reporting a fraud, even to actively participate in its continuance, in order to run up the total and, thus, their percentage of the take.
I'm quite sure that by now the reader has thought of numerous examples in his or her own experience which qualify as "folly." However, just in case, Table 1 presents some additional examples well worth pondering.
Extremely diverse instances of systems which reward behavior A although the rewarder apparently hopes for behavior B have been given. These are useful to illustrate the breadth and magnitude of the phenomenon, but the diversity increases the difficulty of determining commonalities and establishing causes. However, the following four general factors may be pertinent to an explanation of why fouled-up reward systems seem to be so prevalent.
1. Fascination with an "Objective" Criterion
Many managers seek to establish simple, quantifiable standards against which to measure and reward performance. Such efforts may be successful in highly predictable areas within an organization, but are likely to cause goal displacement when applied anywhere else.
2. Overemphasis on Highly Visible Behaviors
Difficulties often stem from the fact that some parts of the task are highly visible while other parts are not. For example, publications are easier to demonstrate than teaching, and scoring baskets and hitting home runs are more readily observable than feeding teammates and advancing base runners. Similarly, the adverse consequences of pronouncing a sick person well are more visible than those sustained by labeling a well person sick. Team-building and creativity are other examples of behaviors which may not be rewarded simply because they are hard to observe.
In some of the instances described the rewarder may have been getting the desired behavior, notwithstanding claims that the behavior was not desired. For example, in many jurisdictions within the U.S., judges' campaigns are funded largely by defense attorneys, while prosecutors are legally barred from making contributions. This doesn't do a whole lot to help judges to be "tough on crime" though, ironically, that's what their campaigns inevitably promise.
Sometimes consideration of other factors prevents the establishment of a system which rewards behavior desired by the rewarder. The felt obligation of many Americans to vote for one candidate or another, for example, may impair their ability to withhold support from politicians who refuse to discuss the issues. Similarly, the concern for spreading the risks and costs of wartime military service may outweigh the advantage to be obtained by committing personnel to combat until the war is over. The 1994 Clinton health plan, the Americans with Disabilities Act, and many other instances of proposed or recent governmental intervention provide outstanding examples of systems that reward inefficiency, presumably in support of some higher objective.
Managers who complain about lack of motivation in their workers might do well to consider the possibility that the reward systems they have installed are paying off for behavior other than what they are seeking. This, in part, is what happened in Vietnam, and this is what regularly frustrates societal efforts to bring about honest politicians and civic-minded managers.
A first step for such managers might be to explore what types of behavior are currently being rewarded. Chances are excellent that these managers will be surprised by what they find--that their firms are not rewarding what they assume they are. In fact, such undesirable behavior by organizational members as they have observed may be explained largely by the reward systems in use.
This is not to say that all organizational behavior is determined by formal rewards and punishments. Certainly it is true that in the absence of formal reinforcement some soldiers will be patriotic, some players will be team oriented, and some employees will care about doing their job well. The point, however, is that in such cases the rewarder is not causing the behavior desired but is only a fortunate bystander. For an organization to act upon its members, the formal reward system should positively reinforce desired behavior, not constitute an obstacle to be overcome.
An irony about this article's being designated a management classic is that numerous people claim to have read and enjoyed it, but I wonder whether there was much in it that they didn't know. I believe that most readers already knew, and act on in their non-work lives, the principles that underlie this article. For example, when we tell our daughter (who is about to cut her birthday cake) that her brother will select the first piece, or inform our friends before a meal that separate checks will be brought at the end, or tell the neighbor's boy that he will be paid five dollars for cutting the lawn after we inspect the lawn, we are making use of prospective rewards and punishments to cause other people to care about our own objectives. Organizational life may seem to be more complex, but the principles are the same.
Another irony attached to this "classic" is that it almost didn't see the light of day. It was rejected for presentation at the Eastern Academy of Management and was only published in The Academy of Management Journal because Jack Miner, its editor at the time, broke a tie between two reviewers. Nobody denied the relevance of the content, but reviewers were quite disturbed by the tone of the manuscript, and therefore its appropriateness for an academic audience. A compromise was reached whereby I added a bit of the great academic cure-all, data (Table 1 in the original article, condensed and summarized in this update), and a copy editor strangled some of the life from my writing style. In this respect, I would like to acknowledge the extremely competent editorial work performed on this update by John Veiga and his editorial staff. I am grateful to have had the opportunity to revisit the article, and hope the reader has enjoyed it also.
(n1) Originally published in 1975, Academy of Management Journal, 18, 769-783.
(n2) Charles Perrow, "The Analysis of Goals in Complex Organizations," in A. Etzioni (ed.), Readings on Modern Organizations (Englewood Cliffs, NJ: Prentice-Hall, 1969), 66.
(n3) Chester I. Barnard, The Functions of the Executive (Cambridge, MA: Harvard University Press, 1964), 165.
(n4) L.H. Garland, "Studies of the Accuracy of Diagnostic Procedures," American Journal Roentgenological, Radium Therapy Nuclear Medicine, Vol. 82, 1959, 25-38; and Thomas J. Scheff, "Decision Rules, Types of Error, and Their Consequences in Medical Diagnosis," in F. Massarik and P. Ratoosh (eds.), Mathematical Explorations in Behavioral Science (Homewood, IL: Irwin, 1965).
(n5) Miami Herald, May 8, 1994, 1a, 10a.
(n6) Goal displacement results when means become ends in themselves and displace the original goals. See Peter M. Blau and W. Richard Scott, Formal Organizations (San Francisco, CA: Chandler, 1962).
(n7) San Francisco Examiner, reported in Fortune, February 7, 1994, 161.
Legend for Chart:
A - We hope for...
B - But we often reward...
long-term growth; environmental
setting challenging "stretch" objectives
achieving goals; "making the numbers"
downsizing; rightsizing; delayering;
adding staff; adding budget; adding Hay
commitment to total quality
shipping on schedule, even with defects
candor; surfacing bad news early
reporting good news, whether it's true or
not; agreeing with the boss, whether or
not (s)he's right
By Steven Kerr
Steven Kerr (Ph.D., City University of New York) is a visiting professor of management in the University of Michigan business school, and has recently assumed the position of Vice President, Corporate Management Development, for General Electric Co. (managing Crotonville). He has been on the faculties of Ohio State University and the University of Southern California (where he was Dean of the Faculty from 1985-89), and was President of the Academy of Management from 1989-90.