Abstract
This journal attempts to bring together the practical and the theoretical, the business world and the academic world. Although I have a doctorate degree, I have never had true research interests. I learned early that I enjoy solving problems. I occasionally dig into the research literature but I have been selective. For that reason I took for granted that I missed important studies on pay and performance.
Now as I reconnect with people in academia I have learned that I missed very little. I am not at all sure why but there has been little useful research for years. It is badly needed.
There are two excellent articles that highlight the need for research. The most comprehensive, “Performance Evaluation and Pay for Performance,” by Sara Rynes, Barry Gerhart and Laura Parks, was published in Annual Review of Psychology, 2005. The second, a Harvard Business School Working Paper, is more recent and more theoretical, “The Psychological Costs of Pay-for-Performance: Implications for the Strategic Compensation of Employees,” 2011. The authors are Ian Larkin, Lamar Pierce and Francesca Gino. Both have lengthy bibliographies of articles summarizing research on pay and performance.
For virtually every employer, payroll is the largest cost, after excluding the cost of goods sold. In an earlier era, when employees were seen strictly as a cost, the goal was to keep salaries low. When jobs involved routine tasks, there was little need to assess individual capabilities or their performance. Obeying the rules and following orders was all that was required. Employees were paid to do their job; there was no real interest in understanding what caused the differences in performance and therefore little research.
More recently, with knowledge jobs and increased worker autonomy to tackle problems, there is a recognition that under the right circumstances people can perform at much high levels. That has triggered research on how high-performance environments emerge and what factors come together to create what is now referred to a high-performance organization. Significantly, several popular books on high performance ignore or downplay pay and performance as factors that create a high-performance environment.
However, pay is universally seen as a source of motivation. There are other books, not surprisingly by consultants who include compensation in their practice, that cite the importance of pay and performance in achieving high performance.
Having read a number of books and studies on high performance, I am reminded of the tale from India of the blind men and the elephant. As with the elephant, compensation is a complex “beast” that is interwoven with every occupation and operation across an organization. Its impact can be very different from work group to work group. It would be enormously useful to sort all this out and to know how to plan and manage pay to gain competitive advantage and avoid the potholes.
Actually, one of the first points that needs clarification is that the phrase pay for performance is used to refer to two different practices—the linkage of wage or salary increases to performance and the use of bonus or incentive plans. For practitioners, studies that fail to focus explicitly on one of the two practices make the conclusions difficult to interpret and use.
It has also become clear, at least to me, that much of the research in the past is suspect. In my opinion, studies that focus on the behavior of students or volunteers performing artificial tasks in psychology labs are not useful. The experiments are far from a true work setting where financial rewards are granted once a year and nonfinancial rewards, positive as well as negative, are granted haphazardly throughout the year.
From the same perspective, the studies from years ago in industry reflect work settings that are very different than today. The most prominent are the classic studies by Herzberg, reported in the 1959 book, The Motivation to Work, that found pay to be a cause of dissatisfaction. He failed to recognize that employees react to two separate issues—their pay level and their pay increase. His research design is not surprising, however, since “merit” pay was not nearly as prevalent in the 1950s as it is today.
It is not clear how salary level can trigger “satisfaction” (although I have no evidence to support this). An employee could be satisfied or dissatisfied with his or her pay level but that is not the same as satisfaction. In my consulting, I have met highly paid executives who are dissatisfied with their salary. It would be very useful to understand what factors contribute to how employees view their pay.
More research is needed. Bibliographies of publications on pay are dominated by studies done a decade or more ago. The most prominent articles were published more than 20 years ago. For employers, the basic questions are largely unanswered. A chasm remains between the theory and the need for practical answers for employers that need to manage compensation systems to benefit their organization.
Salary systems often appear to be clones but a close look makes it clear that the variations in the policies are almost endless, and change from year to year. Simply stating that pay increases are linked to performance does not disclose how a policy works in practice. Moreover, the work settings and cultures in which pay decisions are made and communicated introduce endless variations. Finally, each employee arrives at his or her job with his or her individual background, including demographics and prior work experience, which influences his or her reactions. These factors are comingled in the work place, and I am not aware of any research that isolates the impact of pay increases on performance.
Base salaries are presumably a key in recruiting and retaining employees (although again the impact of base pay is difficult to separate from other job issues). There is of course the standard argument from economics that higher salaries, relative to market levels, will attract better qualified job candidates. I often discuss that as a pay strategy with clients. It is an important policy consideration because employers have to balance the added cost with the expectation of better performance from better qualified employees. The currently popular author Daniel Pink (Drive: The Surprising Truth About What Motivates Us) has speculated that it will enable employers to save money! There has been virtually no research on this question.
Pink is also one of the most prominent current critics of pay for performance. He is one of several individuals who have argued extrinsic rewards are less effective as motivators than intrinsic rewards. Over the years, I have found it curious that the critics, starting with Dr. W. Edwards Deming, the quality management guru, have virtually no experience managing employees. Pink was a speech writer for Vice President Al Gore and has a law degree from Yale. For the most part, the critics in the academic world are all paid for their performance (although the typical faculty salary system is hardly a best practice model)—and no doubt might protest loudly if they were paid on a uniform step increase policy.
It may of course be correct that intrinsic rewards are more effective although that is an extremely broad conclusion since it ignores the amounts and circumstances governing monetary rewards. I happen to believe that is true but there is also, within our society, the expectation that job success will lead to a higher income. It is a mistake to discount the importance of pay. Again, the conclusions were largely based on artificial experiments. Moreover, in many of those occupations—health care specialties, for example, employees still react with anger if they feel they are not adequately compensated.
The belief that opportunities to earn money motivate behavior is widely accepted and supported by psychological theory and research. Financial incentives are the foundation of a capitalist system. Anyone with experience in executive compensation knows that cash incentives and stock ownership plans are universal in every industry. If anything, the critics argue that the incentives work too well. Sales compensation systems are based largely on the use of financial incentives and have few critics. Experience with profit-sharing plans as well as gain-sharing plans—group incentives—supports the impact of incentives—as long as people feel their performance affects results. Incentives are also prevalent in professional sports and entertainment. Parents rely on the “theory” in raising children. The idea is used in marketing as well.
There is research that highlights the distinction between bonus plans—where payout decisions are made largely subjectively and after the fact—and incentives where payouts are linked explicitly to predefined results. Evidence shows that the latter are far more powerful.
There is also evidence that group and team incentives, with explicit linkages to results, are also powerful motivators. With gain-sharing plans, studies show that the money saved (the goal of gain sharing is reducing costs) can be four or five times the payouts. Although it is rarely recognized, management incentive plans are a form of group incentive plan since at least a portion of the payouts depends on company performance.
But the research tells us very little about the design of incentives. How does the prospective size of the payout affect performance? How does the number of performance measures affect performance on individual measures? How does the relative weighting of the measures affect performance?
The research is also silent on how the work circumstances come together with the prospect of cash payouts to influence behavior. Autistic individuals may not be affected by their surroundings but that is obviously not the case in a typical work setting. The notion of a performance culture, for example, is known to influence behavior but as with many constructs, we do not have a solid definition or agreement on the factors that create the culture. There are a number of management practices that highlight the importance of performance and contribute to the culture, and very possibly influence behavior.
So there is broad acceptance of the argument that financial rewards motivate behavior. That was the conclusion of a National Academy of Sciences committee created in 1990 at the request of the U.S. Office of Personnel Management. They reviewed the research evidence to that point and recommended that the federal government should shift to a pay for performance policy. The report and recommendation was quickly forgotten but it remains the most comprehensive review of the literature to date.
But with all of that, new critics surface every few years arguing against policies that link salary increases to individual employee performance. Daniel Pink is the most recent. One of their themes is that there is little evidence showing that policies contribute to higher performance. The National Academy of Sciences committee also acknowledged the evidence is scarce but then that was 20 years ago.
The fact is that the evidence would be difficult to develop. First, the practice is close to universal which makes comparative studies almost impossible. Second, it is extremely difficult with many jobs to measure performance and even more difficult to confirm performance has improved. Third, all the variations discussed previously make attempts to generalize suspect.
In the private sector, the critics have had virtually no impact, but pay for performance continues to be a front burner issue in government and still is not solidly established in health care and higher education. Developing a better understanding of what factors affect its acceptance would have significant value to those employers. Is it the culture? The managers? Unions? Weaknesses in the performance management process? The absence of confidentiality? It would also be advantageous to understand what planning and implementation strategies work best. There are not only solid success stories but also failures.
A simple but unanswered question relates to how the differences in salary increases affect motivation and performance. To illustrate, with the typical “merit” increase policy, if the budget for salary increases is 3%, a star performer might expect 6%. For an employee paid a base salary of $60,000, a star performer would be granted an extra $1,800 or roughly $35 a week before taxes. That is hardly enough to make a difference in lifestyle. It is difficult to believe the prospect of that amount viewed over a year would be a strong motivator.
But that is a very narrow view of the policy’s impact. It focuses only on the money. As a comparative study, it would be valuable to contrast performance in an organization under both a merit increase policy and an “automatic,” step increase policy. That is of course common in many public agencies.
We really do not understand how switching from a step increase policy to merit pay impacts performance or how to avoid the problems so often encountered. In addition to elevating the attention to performance, the change is often associated with a new performance system and revised individual ratings. On the positive side, the change places more emphasis on recognition. New performance systems typically place greater emphasis on individual performance goals and success (or failure) in achieving the goals would likely affect employee satisfaction. So the change in policy is not simple, and that is the issue.
My experience suggests the success or failure of pay for performance is in the hands of managers and their acceptance of this new responsibility. It redefines their relationship with their people. Their skill in managing performance is a central issue but more important is their commitment to this new role. If that’s a valid observation, it makes the preparation and acceptance by managers for their role a key issue. Researchers should develop a better understanding of how to get managers on board with the change.
The questions for researchers are endless. The fact is that practical research conclusions would have a high value to practitioners. The size of the compensation consulting “industry” suggests how willing employers are to spend money for help with pay issues. Consulting firms conduct their own “research,” but their purpose is usually to generate information that they can use in marketing their services. They rarely develop “new” knowledge. Researchers who develop knowledge that helps employers improve their pay systems would find a ready audience willing to pay for the answers.