Discussion

Journal of Economic Perspectives—Volume 32, Number 3—Summer 2018—Pages 195–214

L abor is supplied because most of us must work to live. Indeed, it is called “work” in part because without compensation, the overwhelming majority of workers would not otherwise perform the tasks. Of course, work can be rewarding and empowering, which adds to work’s compensation. However, absent compensation, work would be much rarer and confined to those activities that are enjoyable, but not necessarily most needed by society.

The evidence that compensation affects worker behavior is overwhelming. At the most basic level, almost all of the labor component of GDP that is derived from work is paid rather than voluntary. Beyond that, the literature is full of exam- ples where manipulating the pay structure alters worker behavior, by affecting either hours of work or output associated with it. Incentives are a necessary part of inducing the work that makes an economy go, even when those incentives must be self-imposed. Economists have understood the importance of incentives for decades; for example, they discussed it in the context of Soviet-style work environments.1 The theory developed further during the 1970s and 1980s with modern agency theory, with early examples being Ross (1973), Lazear (1979, 1986), and Hölmstrom (1979).

1 See Bergson (1944, 1978) and Weitzman (1984). Bergson discusses the importance of incentives within command economies towards the close of The Structure of Soviet Wages: A Study in Socialist Economics (1944, p. 204):

Compensation and Incentives in the Workplace

■ Edward P. Lazear is the Davies Family Professor of Economics at the Graduate School of Business, and the Morris A. and Nona Jean Cox Senior Fellow at the Hoover Institution, Stanford University, Stanford, California. † For supplementary materials such as appendices, datasets, and author disclosure statements, see the article page at https://doi.org/10.1257/jep.32.3.195 doi=10.1257/jep.32.3.195

Edward P. Lazear

196 Journal of Economic Perspectives

Providing incentives can be important, indeed game-changing. In the study of Safelite Auto Glass installers discussed later (Lazear 2000b), a switch from hourly wage pay to a piece rate structure had an almost immediate and enormous effect of increasing productivity by 44 percent. The study showed both that changing a compensation scheme could have large effects and that economic theory does well in predicting these outcomes.

Personnel economics in general and the theory of incentives in particular has made its way into business. The combination of the academic literature, the popular press, and cohorts of students who have been schooled in the new approaches and who now are managers has influenced the way in which business is practiced. Some Silicon Valley companies like Success Factors and Merced Systems use these methods explicitly in providing expertise to other companies. Others, like Safelite mentioned above, and many other companies that use bonuses and promotions as motivators, incorporate the findings and analyses of incentive theory into their compensation practices. This is true not only in the United States, but also in Europe, particularly in Switzerland and Germany.

Forms of Incentive Pay

It is common to associate incentive pay with payment that is directly related to output.2 This is too narrow. Virtually all pay methods provide incentives. A better taxonomy is to think of incentive compensation as being described by a two-by-three matrix, where columns relate to pay on input versus pay on output, and rows differ- entiate payment schemes as absolute payment that is discrete, absolute payment that is continuous, and payment that is primarily relative, as discussed in Lazear (2000c). Table 1 spells out the taxonomy.

For example, many workers face input-based pay with discrete incentives. The relevant input is that workers are paid per unit of time, either hours, weeks, months, or even years. Workers in this setting have no flexibility over the amount of time worked. For example, a retail clerk’s contract may specify that 40 hours of work per week are required. The worker is paid per hour, but inflexibility on the choice of hours is part of the job. In contrast, a number of part-time jobs use input-based pay, but incentives are continuous in that the worker has choice over the amount

“It is of great significance that Soviet equalitarianism was not of the utopian variety. That the worker requires a pecuniary incentive to acquire skill, to accept responsibility, to perform more arduous labor, and to increase his productivity, was an accepted principle of Soviet wage policy even in the period of War Communism. The equalitarian’s apprecia- tion of the magnitude of the incentive required may have been dim, particularly before 1920. But at least a sanguine appraisal of the conditions of supply set a lower limit to the reduction in differentials. If we may recur to the subject which was broached at the outset of this study, Soviet equalitarianism represented not an abandonment of capitalist wage principles, but at most a distorted application of them.”

2 These schemes are sometimes referred to as “high powered,” as in Williamson (1975, 1985). This termi- nology is somewhat misleading because input-based, discrete schemes can also provide strong motivation to work, but load all incentives on achieving an exact target amount.

Edward P. Lazear 197

of input supplied. For example, in academia, part-time faculty may be paid on the basis of the number of courses taught, which is an input measure (the output is what students derive from the course), but the instructors may be given some choice over how many courses they would like to teach. Although input-based contracts induce workers to put forth a particular amount of measured input, like hours, they may fall short in providing incentives for effort. As a result, time-based contracts are almost always coupled with some implicit or explicit performance standard that must be met to avoid being dismissed.

Common examples of continuous output-based schemes include piece-rate workers like those in agriculture who are paid according to the amount of crop harvested, salespersons whose compensation depends directly on sales, taxi drivers who rent their taxis for a flat rate and keep all revenue generated, and Uber and Lyft drivers. One major advantage of this approach is that it accommodates a variety of worker preferences. The scheme motivates those who want to work at high levels of effort as well as those who choose to work at lower levels of effort. A disadvantage is that a pure piece-rate scheme makes the worker bear risk associated with variations in exogenous factors like business conditions. Workers, especially low-wage ones, are less well-suited to bearing risk than are capital owners who can diversify their holdings.

Discrete output-based incentive schemes induce all workers to focus on a partic- ular level of output. For example, a homeowner may hire a contractor to resurface a driveway at a given price. Payment for completion of the job motivates the contractor to perform, but all the incentives are concentrated on meeting the exact target—no more and no less. These all-or-nothing output-based contracts are less able to deal with heterogeneity, but they do create very strong incentives to get exactly the specified job done. Examples from the gig economy include Task Rabbit, where workers are paid a fixed amount to complete a specific task, or Upwork, which matches programmers with firms that need a specific piece of code to be written. Indeed, the rise of the gig economy may encourage a move away from time-based pay toward continuous versions of output-based pay, both because output is more easily measured in the gig economy and because hours worked are more difficult to measure.

Finally, both input- and output-based incentive schemes can be based on relative, rather than absolute performance. The classic form of relative scheme

Table 1 Taxonomy of Incentive Compensation

Payment on Input Payment on Output

Discrete Pay per hour with a specified hours requirement

Fixed payment for completion of construction project

Continuous Time-based pay that allows worker choice of labor units supplied

Piece rates

Relative Promotion tournaments based on subjective relative effort evaluation

Promotion tournaments based on some metric of relative output

198 Journal of Economic Perspectives

is a tournament in which the worker who does best receives a promotion. These schemes, discussed in more detail below, also provide incentives, even though they are both discrete and relative.

The theme of this essay is that incentives affect behavior and that economics as a science has made good progress in specifying how compensation and its form influences worker effort. This is a broad topic, and the purpose here is not a compre- hensive literature review on each of many topics. Instead, a sample of some of the most applicable papers are discussed with the goal of demonstrating that compensa- tion, incentives, and productivity are inseparably linked.

An underlying message of the discussion is that well-chosen compensation methods can affect positively both productivity and worker well-being. When firms provide an appropriate compensation structure, workers who join those firms benefit from being compensated at higher levels. Most of that additional compensa- tion is inframarginal, meaning that the additional compensation more than offsets the disutility from the additional effort provided.

Piece Rates and Continuous Output Incentives

When piece rates are paid, some measure of output is specified and workers are paid on the basis of the number of units they produce. Piece-rate pay is best suited for situations in which output is easily observed and quality is not much of an issue (and can be ensured by occasional inspection). A standard example is agricultural harvesting, although even in agriculture, piece rates may be used more when crops are not delicate than where quality is more of an issue, as pointed out by Moretti and Perloff (2002). The literature is virtually unequivocal in documenting that for the circumstance where piece-rate pay is well suited, it provides incentives for workers to produce as predicted by standard theory.

In Lazear (2000b), mentioned above, I focus on Safelite Auto Glass installers, a company that switched from hourly wage pay to a piece-rate structure. Workers earned on average $11 per hour before the change to the piece rate and installed an average of 2.7 windshields per eight-hour day. The piece rate was set at approximately $20 per windshield, coupled with a minimum earnings guarantee. Productivity increased almost immediately by 44 percent. About half the increase is traceable to the workers who were present at the time of the switch, while the rest is attributable to the higher productivity of workers hired after the piece rate was put into place.

In some other prominent studies of piece rates, Shearer (2004) studies tree planters in British Columbia. He conducts a field experiment where nine randomly selected workers are observed for 120 days. During half the period, the workers were paid a fixed wage. During the other half, they were paid a piece rate. Again, both mean and variance are higher under piece rates than under fixed wages. Average output is about 21 percent higher under piece rates than under a fixed wage system. Bandiera, Barankay, and Rasul (2007) perform a field experiment where front-line supervisors, namely field managers on a fruit farm in the United Kingdom, were

Compensation and Incentives in the Workplace 199

given a performance bonus halfway through the season. The field managers could increase fruit-picking productivity by their subordinates by working harder them- selves, which involved clearing the filled crates faster; by assigning workers more strategically to rows of fruit; and by hiring better workers. The introduction of the performance bonus to field managers increased the overall productivity of their subordinates by about 25 percent.3

These examples of piece rates, although telling, of course do not mean that piece rates always result in higher output, compensation, and profit. In settings where output and/or quality are not very observable, piece rates may not work well. Even in settings well suited to piece rates, a piece rate that is set too high could raise cost per unit of output, or even reduce the level of output, if the income effect of the wage increase were large enough. But as a practical matter, firms are unlikely to put in place and maintain a piece-rate scheme that reduces profits.

Because worker output may be multidimensional and difficult to measure, a traditional piece-rate system is rarely used. This is the subject of some early work by Fernie and Metcalf (1998), which studies four firms in the United Kingdom, three of which are call centers and one of which is a bookmaker (a licensed betting office). The main thrust of this work is to compare the predictions of personnel economics to those of the older institutional literature with respect to pay schemes chosen. It is the choice of compensation scheme, rather than the effect of that choice on productivity, that is the subject of the analysis.

Performance-based pay, like piece-rate pay, seems to be associated with higher levels of output and pay for the average workers, but also with a higher dispersion of pay. As one example, Booth and Frank (1999) analyze performance- related pay using data from the British Household Panel Survey. Although they do not have information on output, they have detailed data on earnings and find that performance-related pay is associated with about 9 percent higher earnings for men and 6 percent higher earnings for women. In another study, Jirjahn and Kraft (2007) analyze the Hanover Panel of German Manufacturing Firms (1997). They have a variety of measures of compensation, including wage dispersion and the type of pay (piece rate or fixed wage) offered. They find that higher produc- tivity at a firm is associated with higher wage dispersion. The elasticity is about .2, meaning that a 10 percent increase in wage dispersion is associated with 1.8 percent higher productivity. Additionally, they find that the productivity-dispersion elasticity is considerably higher in firms that operate an explicit piece-rate system, varying between .7 and .9, depending on the type of system used. The evidence on perfor- mance pay and piece rates suggests that when pay is overly compressed, it will tend to reduce incentives and output.

3 In a study not focused on alternative methods of compensation (Lazear, Shaw, and Stanton 2015), we also find evidence of the importance of front-line supervisors on performance in the context of a company-based dataset on a technology-based services job, where the output of workers at a firm can be monitored by computer.

200 Journal of Economic Perspectives

Team-Based Incentives

Consider a high-skill worker who also has the ability to help others develop their own skills. If that worker is paid a straight piece rate based on individual output, the worker would have no incentive to help others. However, if that worker is assigned to a relatively small team and paid partly on the basis of team output, the worker would face a tradeoff between spending time on personal work and spending time helping others. If compensation is based solely on the output of the team, the worker might even find it useful to spend most or all of the time helping others.

Joint production poses problems for incentive theory because workers may not have the right incentives to motivate their coworkers. Peer effects can interact with incentives and affect productivity, whether through monitoring, coaching, or motivation. In Kandel and Lazear (1992), my coauthor and I lay out the theory of the specific ways in which peer effects may operate, and Hölmstrom (1982) looks at free-riding in teams. An accumulating body of empirical evidence has validated a number of theoretical predictions from these models; for example, Mas and Moretti (2009) document positive, albeit small, peer spillover effects. (Substituting an above-average peer for a below-average peer increases a given worker’s produc- tivity by about 1 percent.)

Although team compensation suffers from free-rider effects, in small number settings, like medical practices where there are only a few physicians, incentive dilution may not be pronounced. Gaynor, Rebitzer, and Taylor (2004) find that incentives to conserve on costs, which increase physician take, are more effective in smaller physician groups—that is, groups in which the actions of individual physi- cians will have a greater effect on total savings.

But team incentives can prove useful in larger settings as well, perhaps by encouraging the adoption of more efficient methods of production. Employee stock option or ownership plans can be viewed as a broad-based team incentive. An obvious problem is that if a worker owns just a tiny fraction of the company, the return to the effort of that individual worker may be small relative to the cost. However, the ability to affect other workers magnifies the effect of effort on pay.

The effects of team incentives, although small, can be observed in real world data. Kruse (2016) reports that a review of about 100 studies suggests that there are positive effects on productivity, pay, job stability, and firm survival of employee ownership, although causation and interpretation questions remain. These results are consistent with those presented by Hamilton, Nickerson, and Owan (2003), who analyze a garment production plant that moved to team compensation in two stages, the first being voluntary and the second being compulsory. The authors find that average output rose by almost 20 percent after the switch to team production from individual piece-rate pay. In the context of this production process, if effort is defined more broadly to include effort in helping other workers on the job, a stan- dard piece rate will fail to accomplish that effectively.

Bartel, Cardiff-Hicks, and Shaw (2017) study an international firm of lawyers in which compensation switched from pay on the basis of individual billings to one based on team revenues. Under the first system, the senior lawyers were less willing

Edward P. Lazear 201

to share the work with their subordinates because their compensation depended strictly on what they billed themselves. After the switch, the same senior attorneys allocated more of the work to their subordinates.

Bandiera, Barankay, and Rasul (2005) explore the interaction between incen- tives and social networks. As in their other work, the data here come from studying agricultural workers on a UK farm. They compare the behavior of workers under piece rates, where compensation is independent of the performance of others, to that under a relative compensation scheme, where high output of a given worker imposes negative effects on the compensation of others because it raises the stan- dard against which others are compared. Output is lower under the relative pay scheme. However, altruism does not seem the correct explanation. The disincentive of potentially high producers to cause negative externalities for others are internal- ized only when their actions can be monitored by the affected parties—specifically, when those who lose as a result of their efforts work alongside them and can observe the effort taken.4 Their evidence demonstrates that compensation has significant effects on the way that workers behave toward one another.

Finally, some have documented how changing the pay structure relative to expectations might affect performance. Mas (2006) finds that performance changes when pay is exogenously altered from some reference point. Krueger and Mas (2004) find that at a Firestone tire factory, output quality declined (as evidenced by increased complaints about defective tires) following labor strife.

Relative Pay and Tournament Theory

“Tournament theory” is the name used to describe the literature that focuses on providing incentives to workers on the basis of their relative performance within a firm. The title is not much of an abstraction from the way many workers— especially those climbing the corporate ladder—think about their situations.

To grasp the intuitition behind this approach, consider two players competing in a tennis match—say, the Wimbledon finals. The two players are (presumably) fairly close in skill in an absolute sense, but even when the match is very close, the prize for the winner is considerably larger than that for the loser. The winner’s prize is based on relative, not absolute, performance. There is an optimal spread between the prize for the winner and the runner-up, and optimality is defined so as to elicit the efficient level of effort, not the maximum effort possible. For example, it might be possible to set the prize so high that players would be willing to risk death or try to kill the other player to win, as was the case in gladiatorial tournaments, but that would not be efficient because the added output would not cover the cost.

Within a firm, receiving a promotion is akin to winning a tennis tournament. The contestants in this case are typically managers at a lower level in the firm’s

4 An older literature documents how workers may react to “rate-busters,” defined as workers who perform at high levels in an industrial setting and thus have the effect of lowering compensation per unit of output for all others. For an example, see Roy’s (1952, 1954) studies of a machine shop.

202 Journal of Economic Perspectives

hierarchy. Promotion decisions almost always require relative rankings. The raise associated with a promotion motivates young associates at consulting firms to exert high levels of effort in order to win a promotion to partner, just as the hope of tenure motivates assistant professors in academia. However, if the value of winning a promotion is too high, the output from the additional incentives provided will fall short of the additional wages that must be paid to induce workers to accept the job at the outset. The difference between the spread in wages for the promoted and the unpromoted should be just large enough to induce the efficient level of effort, but not so high as to exceed it.

These intuitive ideas have been derived formally in early game-theoretic anal- yses (Lazear and Rosen 1981; Green and Stokey 1983; Nalebuff and Stiglitz 1983), and demonstrated in laboratory experiments (for instance, Bull, Schotter, and Weigelt 1987; Falk, Fehr and Huffman 2008) and in sports environments (for an example from golf, see Ehrenberg and Bognanno 1990).5

In the more germane context of business, Eriksson (1999) finds that many implications of tournament theory hold when looking at data on about 2,600 executives in 210 Danish firms. First, the larger the number of candidates for a promotion, and thus the smaller the probability of being promoted, the higher the pay jump associated with that promotion, which is necessary to provide incen- tives. Second, the jump in pay for promotions at high levels is greater than that at low levels. This is implied by the theory because part of the reward for low-level promotions is the option value of obtaining higher-level additional promotions (Rosen 1976).

The second finding is also corroborated by Belzil and Bognanno (2008), who examine 600 US firms having 25,000 executives. Like Eriksson, they find that pay jumps are larger for promotions near the top than near the bottom. Classifying levels within firms into nine categories, a promotion from the bottom to the next level generates a 15 percent raise, whereas a promotion from the second-to-the-top level to the top generates a 94 percent raise.

Effort is difficult to observe in data, but absenteeism can, in some situations, be used as a proxy for the effort that workers are willing to commit to a job. Drago and Garvey (1998) examine data from 23 Australian firms. They find that the larger the spread in wages between workers and supervisors, the less absenteeism. They interpret this as workers being willing to commit higher effort to the job when the incentives from promotion to a higher rank are larger.

There are indirect tests of incentives to put forth effort that use observable outcomes as a proxy for effort. The Danish data used by Eriksson (1999) provide a test of outcomes in tournament settings. Eriksson finds that those firms that award larger raises to promotion have better outcomes as measured by profits—and also pay higher average wages. Using US data, Kale, Reis, and Venkateswaran (2009)

5 Gneezy, Niederle, and Rustichini (2003) explore the way men and women respond to piece-rate and tournament structures. They find no statistically significant difference by gender in response to piece rates, but men outperform women in tournament incentive settings. Noteworthy is that women’s perfor- mance is higher in single-sex tournaments than in a piece-rate treatment.

Compensation and Incentives in the Workplace 203

find that the difference between the pay of chief executive officers and vice presi- dents is strongly and positively related to return on assets, return on equity, and market-to-book value. Using data from the Swedish Registry, which surveys about 10,000 managers, Heyman (2005) finds a positive and significant effect of intra-firm wage dispersion on profits and average pay. He also finds that dispersion tends to be larger in firms that experience higher market demand volatility, which is consis- tent with tournament predictions. When workers compete in a noisy environment, larger incentives and wage dispersion are required to compensate for the diluting effects on incentives of higher risk. These studies document that performance varies in a way consistent with tournament theory, and they provide indirect evidence that effort is affected by altering compensation structures in the ways predicted by tour- nament theory.

Various other findings also support the tournament view of incentives. Mobbs and Raheja (2011) find that firms that have multiple candidates competing for promotion among top executives do better than those with a groomed successor. Actually, the prediction is that the relationship between number of candidates and effort exerted should be an inverted U. If there is a groomed successor, the poten- tial competitors are not highly motivated to compete for the job, just as the authors find. But with an extremely large number of potential competitors, incentives to put forth effort would be reduced as well, because the change in probability of winning the tournament/promotion from exerting additional effort becomes very small.

Compensation may affect managers’ risk-taking behavior. The classic concern is that risk-averse managers adopt overly safe strategies to protect their jobs. Because shareholders can diversify their portfolios, managers who implement the desires of the principals should behave as if they are risk neutral. Tournaments may encourage more risk taking among risk-averse managers because winning a multicontestant tournament is a tail-event. A sports parallel is that, in Olympic downhill skiing, a cautious approach will assure a loss of the gold medal, even if it minimizes expected time to complete the course. Winning requires great talent, but also a high draw of positive luck. Choosing low-variance strategies precludes getting a large positive draw on luck.

Properly structured stock options can also be used to address managerial risk aversion. Just as was the case in the downhill skiing analogy, call options create incentives for adopting riskier strategies (as noted in Jackson and Lazear 1991). If the exercise price of the option is high, then only a manager who adopts a high-risk strategy will end up with a stock that is “in the money,” that is, has a value higher than the exercise price.6 Kini and Williams (2012) find that increasing tournament-like

6 An intriguing question sometimes posed in the compensation literature is whether firms might wish to grant put options, rather than traditional call options. With a traditional stock call option, managers have incentive to put forth effort that translates into higher stock prices, because then they can sell the stock at a higher price differential over the exercise price of the option, which is given and predetermined. But consider an alternative. Managers could be given higher base salaries and required to short put options on stock. If the stock price falls below a certain level, managers would have to purchase company stock at a price higher than the market price, forcing them to overpay for shares of their company’s own stock. Both kinds of options provide incentives. The difference is that even risk-neutral managers who are short

204 Journal of Economic Perspectives

incentives increases risk taking. To the extent that the losing prize—the wage of the unpromoted worker—is relatively constant, an increase in spread is akin to increasing variance of an option. Kini and Williams find that a bigger pay gap at a firm implies more cash flow volatility and more volatility in returns. Additionally, firms with bigger pay gaps undertake riskier investments (like emphasizing research and development over tangible assets) and have higher leverage ratios.

Not all incentives associated with more risk taking are positive. Hass, Muller, and Vergauwe (2015) find that a larger variance in the pay structure creates more dysfunctional responses like fraud, as reflected in class-action lawsuits against the firm. More generally, when a worker’s promotion depends on relative perfor- mance, workers have an incentive to be uncooperative with fellow workers and in the extreme, even to sabotage their efforts. Drago and Garvey (1998) find that workers are less willing to be helpful to co-workers when the promotion rewards are greater. Falk, Fehr, and Huffman (2008) document the same behavior in a lab setting, and even document a rise in sabotage that increases with the prize spread. Additionally, they find that the pay setter compresses pay in response to this. Thus, a tradeoff arises here, as predicted in Lazear (1989). Reducing the spread between the promoted and the unpromoted has adverse consequence for good effort, but mitigates the uncooperative behavior that is inherent in any relative compensation scheme. The dangers of sabotage or uncooperative behavior are another reason for limiting the size of the promotion incentives. In real-world settings, compensation- induced incentives can end up being beneficial or adverse, when poorly designed. But either answer confirms an underlying theme of this essay, which is that compen- sation schemes affect worker incentives and behavior, both in theory and in practice.

Career Incentives and the Experience–Earnings Profile

For many workers, promotion is not a realistic possibility. For example, consider a middle-aged manager who has been in the same position for many years with little hope of moving up in the hierarchy. Other than paying the worker directly on output, which may not be easily observed, it may be difficult to motivate workers in these posi- tions. For such workers, the experience–earnings relationship can serve as a way to motivate workers who are neither paid piece rates nor are candidates for promotion.

The theory behind these schemes was exposited in Lazear (1979, 1981). The idea is rooted in the empirical observation that experience–earnings profiles tend to be nonnegatively sloped throughout a worker’s career. This pattern holds despite both casual and more serious empirical evidence suggesting that productivity declines with age. For example, age at some point brings a decline in the ability to conduct logical thinking and reasoning (Ruth and Birren 1985), a diminution of

a put option would want to choose safe strategies to avoid a big downside effect on stock price whereas managers who are long a call option would want to take more risk. The fact that firms make managers long on call options and rarely (if ever) short on put options suggests that shareholders and the boards are trying to increase managerial risk taking, rather than decrease it.

Edward P. Lazear 205

creativity (Florida 2002), and declining abilities to store and process information, solve problems, deal with complexity, and adjust to new situations (Kaufman and Horn 1996; Ryan, Sattler, and Lopez 2000).

Young and middle-aged workers often produce more than the more senior workers who receive higher compensation. How can this pattern survive in equilib- rium? The theory argues that efficiency is enhanced by underpaying the young and overpaying the old. Middle-age and senior workers have incentives to keep their levels of effort high because they earn rents on the job relative to what they would receive if they were to be forced to leave the job. The scheme requires some moni- toring, but this monitoring can occur occasionally and stochastically.

The theory has a number of testable implications. When first laid out, the goal was to provide an economic rationale for mandatory retirement. Because senior workers are overpaid relative to their productivity, they want to work longer than is efficient, which can be defined as retiring when the value of their leisure exceeds their productivity at work.

Because it is difficult in most settings to observe the experience–productivity relationship, direct tests on the slope of experience–earnings versus experience– productivity profiles are virtually nonexistent.7 In an early paper, Hutchens (1987) uses a clever indirect method that relies on the National Longitudinal Survey combined with the Dictionary of Occupational Titles in order to “test whether jobs that involve repetitive tasks tend to be characterized by an absence of pensions, mandatory retirement, long job tenures, and high wages for older workers.” The hypothesis is that jobs with technologies that are well-suited to supervision or over- sight can use more direct monitoring schemes and at the extreme, output-based pay, to motivate workers. For those where the job tasks are less easily measured, one would expect to observe more pensions, long tenures, mandatory retire- ment, and wages that increase more rapidly. He finds that workers who perform nonrepetitive tasks are 9 percent more likely to have a pension, have wages that are 28 percent higher (which, because of controls, reflects a steepening of the experience–earnings profile), and an 18 percent longer job tenure.

In another test, Goldin (1986) argues that women historically had shorter attachments to the labor force and as a result, should be more likely to have direct incentive pay and less likely to have experienced-based incentive pay. She docu- ments that women were more likely to be on piece rates than men, although over time, the difference between men and women should disappear as female labor force behavior resembled that of men. Indeed, other authors have found that to be the case. Booth and Frank (1999) using British data find that women were actually less likely to be on piece rates.

How might laws against age discrimination affect this experience–earnings relationship? Neumark and Stock (1999) use variation in state and federal age discrimination laws to show that stiffer laws against age discrimination actually lead

7 The Safelite data allow for an explicit test of the theory because productivity is measured precisely. The learning curve was quite steep for the installers, but wages continued to grow beyond the point at which experience effects on productivity flattened (Lazear 1999).

206 Journal of Economic Perspectives

to steeper experience–earnings profiles. Their interpretation is that age discrimina- tion laws make it harder to fire older workers, which means that a firm will find it more difficult to renege on its promise to pay older workers more. In this sense, age discrimination laws make the firm’s promise to pay experienced workers high sala- ries credible for younger cohorts, and experience-based incentives are more likely to be provided after age discrimination laws are passed.

More recent evidence is provided by Huck, Seltzer, and Wallace (2011). They use an experimental setting and find that when firms cannot commit successfully, there is “breakdown of worker-firm relations and a dramatic loss in efficiency … It is this comparison that really underlines the success of the Lazear model—the difference deferred compensation makes.”

Labor Supply and Incentives for Inputs

Incentive theory is less frequently applied to or discussed in the context of payment for input, but it is just as relevant. At the most basic level, determining the elasticity of labor supply is fundamentally a question of incentives. This is not the place to attempt a survey of a vast literature on labor supply, but a few issues are directly relevant to the study of compensation incentives. First, raising wages induces individuals to work more. Although the elasticity of labor supply estimates vary from very small to substantial, depending on the group and margin considered, most of the literature supports the view that hours worked and labor force participa- tion respond positively to pay. Second, government programs, which exogenously alter the benefit from work, affect the amount of work in an economy.

Among the earliest papers to study the impact of exogenous wage changes on worker behavior is Rosen (1976), who examines whether tax changes that affected take-home wages led to a change in labor supply. Rosen’s main conclusion based on the Survey of Work Experience for Women 30–44 (1967) is that married women’s labor supply is highly responsive to tax changes that affect the take-home wage rate. Another analyses of the effects of tax changes on labor supply is Eissa and Hoynes (2004), which studies the effect of the Earned Income Tax Credit on behavior, with a focus on labor force participation. One finding is, “A $1 increase in the net wage raises the likelihood that wives work by 2.7 percentage points, or 4.2%.”

Yet another example of exogenous government programs that affect worker behavior is disability insurance. In an early study, Gruber (2000) compared work and disability behavior in different Canadian provincial regimes. Changes made in 1987 in the benefits offered by provinces other than Quebec allow Gruber to conduct a differences-in-differences analysis. Gruber reports that a 36 percent rise in disability benefits resulted in a rise in nonemployment of 11.5 percent from the baseline value. Recent analysis from 2006–2016 with US data reveals a correlation of .97 between the number of new applications for disability insurance (annual) and the unemployment rate (Lazear 2017). In this case, the interpretation is that when a recession occurs, incentives for work are reduced, and some who had qualifying disabilities will exercise their rights to receive disability insurance, forgoing work.

Compensation and Incentives in the Workplace 207

Other Views and Questions

A wide range of empirical evidence demonstrates that the form and level of workplace compensation has important effects on the incentives and behavior of workers and managers. Two additional questions are relevant. First, what is the role of nonmonetary incentives and of intrinsic motivation? Second, are there situations in which market forces may systematically provide incentives that are too large or too small for efficiency? The issue of overcompensation of chief executive officers and other high-level managers is a special case of providing too strong or perhaps the wrong kind of incentives.

Nonmonetary Incentives and Intrinsic Motivation The fact that workers are motivated not only by money is neither novel nor

controversial. The idea dates back to Adam Smith (1776, Bk. 1, Ch. X) and falls under the general rubric of “compensating differentials.”8 Rosen (1974) modern- izes the idea and derives the market equilibrium, understanding that the firm’s profit incentives interact with worker preferences to determine the price that workers must pay to have their tastes accommodated. Rosen’s insights imply that attributes that could be provided at no cost to the firm would not carry with them a compensating differential, even if workers had preferences for the attributes.

Consider intrinsic motivation. Some work occurs in the complete absence of pay. Charitable contributions equal about 2 percent of GDP (Zinsmeister 2016; Giving USA 2017), but that is not the same as work done without pay. Most chari- table contributions go to organizations like those in health care research that hire workers in a competitive labor market and pay wages. Volunteer work is a more direct measure of the amount of work done without monetary compensation and presumably based on intrinsic motivations, which might include altruism or group identification. However, its value is only in the range of 1 percent of GDP (as reported by Hrywna 2017).

Household work accounts for much more work that is unpaid, at about 25 percent of GDP (Bridgman, Dugan, Lal, Osborne, and Villones 2012). However, even though household work is unpaid, it is associated with an almost ideal incen- tive structure because it is akin to self-employment, where a high proportion of the rewards to effort are captured either by the provider of the services or by imme- diate family members. The provider of household services is the residual claimant of services consumed. Services consumed by other family members deliver first- best incentives when altruism by the provider is sufficient to make him or her view services consumed by family members as equal in value (on the margin) to those

8 For example, see Book I, Chapter X of Adam Smith’s (1776) Wealth of Nations: “The five following are the principal circumstances which, so far as I have been able to observe, make up for a small pecu- niary gain in some employments, and counter-balance a great one in others: first, the agreeableness or disagreeableness of the employments themselves; secondly, the easiness and cheapness, or the difficulty and expence of learning them; thirdly, the constancy or inconstancy of employment in them; fourthly, the small or great trust which must be reposed in those who exercise them; and fifthly, the probability or improbability of success in them.”

208 Journal of Economic Perspectives

consumed directly. Relatedly, some consumption might be thought of as work. For example, personal gardening produces output that is not counted in GDP. It is possible that gardening as a consumption activity or cooking for enjoyment fails to be counted in cited estimates of household work. If that is the case, then the estimates (for example, BEA Blog 2012) understate the true share of GDP that is unpaid.

Is intrinsic motivation of broad importance for standard labor market settings? An influential paper by Deci (1972) argued that monetary incentives can in some cases crowd out intrinsic motivation, where people take actions because those actions are personally fulfilling. One famous example is that blood donations declined when the collecting agency moved from voluntary to paid contributions (Titmuss 1970). More germane to the subject at hand is the work of Bruno Frey and coauthors, who have critiqued monetary incentives as a mechanism for enhancing worker productivity, and pointed to some cases in which making explicit the reward has adverse effects on work effort (for example, Frey and Oberholzer-Gee 1997; Frey 1997).

Work that is motivated intrinsically, say by altruism as in the case of work to benefit family members, clearly exists. Nonetheless, most work is done to obtain the direct pay that results from it. At least on the margin, it is necessary to compensate individuals to get them to put forth effort. The reason is straightforward. If effort has value in the market or specifically to a particular employer, then the employer will be willing to pay for that additional effort. That will push the worker to the point where intrinsic motivation has run out and where it is necessary to pay workers to elicit additional effort, a point made as early as Marshall’s (1890) landmark textbook.

For example, many people volunteer some hours, but there are only so many hours that are supplied without pay. Even socially oriented individuals must be paid to exceed that limit. Some salespeople might work a few hours just for the intrinsic joy of selling, but it is unimaginable that much sales work would be done absent explicit compensation. The studies of piece rates and how they motivate workers provide evidence to this effect.

Indeed, the role of compensation and management in general is to induce workers to take those actions that are of value to the firm but would not be provided voluntarily. Much of management is about inducing workers to perform those tasks that have value but are not intrinsically rewarding. Volunteer work and effort provided through intrinsic motivation is consistent with standard economic theory. Economists understand equilibrium and on the margin, a dollar spent raising productivity one way must be as effective as a dollar spent raising it any other way. A firm that has exhausted its profitable means of raising productivity through changes in the compensation structure might find that it has opportunities to raise produc- tivity even further by appealing to or changing worker tastes. The converse holds as well. A firm that has the ability to create a positive culture in the firm must balance the cost of improving culture with simply paying workers to put forth additional effort. An optimizing firm allocates expenditures such that the marginal dollar spent on each form of compensation produces the same worker satisfaction.

Edward P. Lazear 209

Although economists have little to add to psychological theories of prefer- ences, economic theory does provide a guide as to when this kind of psychological manipulation will more plausibly have effects on incentives and pay—and when not.

Altering worker preferences to further align their incentives with those of the firm is likely to be concentrated on those higher-level and higher-paid workers who have stronger attachment to the firm. Job churn in the United States totals about 60 million jobs per year, which is about two-fifths of the entire worker force, according to the Bureau of Labor Statistics (2017) Job Openings and Labor Turnover Survey (JOLTS). Many workers have tenuous connection to their firm, so the firm will have little incentive to invest in ways that would affect their preferences.

Additionally, one should always consider the salary level in question. A manager who earns $300,000 per year may prefer a $5,000 improvement in working condi- tions over $5,000 in pay. A worker who earns $25,000 per year is less likely to prefer $5,000 in more appealing working conditions to $5,000 in pay. Working condi- tions are a normal and likely a luxury good, which implies that extrapolation of responses from one group to another is dangerous. This would also cut across occu- pations. Sorting into occupations is not random, and some occupations are chosen by persons who have an inclination toward nonmonetary rewards. Obvious cases involve missionary work, which implies significant hardship and very low compensa- tion. Those who choose those occupations value social rewards over monetary ones to a greater extent than, say, investment bankers. Less extreme cases are abundant in the labor market.

Beyond these issues, firm specificity matters. In order for a firm to be willing to take an action that affects a worker’s productivity, in this case bearing costs to change preferences, the firm must be able to capture the returns from that invest- ment, which is only possible if the value produced is firm specific. Were the worker not partial to the current firm, any investment that made the worker more valuable, either because skills were enhanced or because the worker was willing to supply more effort at a given wage, could not survive competition. The worker might be willing to supply more effort at the same price, but if that preference had value to other firms in the marketplace, they would bid up the wage commensurate with the additional effort and the firm that made the initial investment would not recoup any of it. As a consequence, firms can only be expected to undertake investments that change worker preferences when those workers are already earning rents at the current firm relative to others or when the change in preferences from the invest- ment creates a stronger attachment to the current firm directly.

Many of the critiques of compensation incentive theory highlight anoma- lies, frequently arising from results in laboratory studies. My personal view of this literature, discussed in Lazear (2000a), is that criticism should be welcome. It is dangerous to be excessively smug about our own theories and ideas, and those who force us to reexamine our analyses and predictions provide a useful service. But that does not imply that the discipline and framework provided by standard economic theory should be rejected. The goal of science is to make sense out of nonsense, not the reverse, and it is incumbent on scholars to think deeply enough so as to incorporate into or reformulate theory to provide a unified picture. Scattered facts

210 Journal of Economic Perspectives

and inconsistencies do not form a scientific literature, but they may help stimulate science to be better.

Sometimes, results from laboratory experiments mislead because they lack a well-posed question. Consider studies that purport to compare piece rates with a tournament pay scheme. To determine which is more effective (holding the envi- ronment constant), it is necessary to compare an appropriately designed piece rate against an appropriately designed tournament. It is always possible to make a piece rate more effective than a tournament by increasing the piece rate and reducing the spread between the winning and losing prizes in a tournament. The reverse is also true. Changes in environment can affect the outcome systematically, too. If experi- ments are to offer useful lessons, they must take considerable care to follow the scientific method, as used by economists and physical scientists alike. Lavy (2002) offers a fine example of how this can be done. He examines a given amount of money spent on teachers and compares that to a similar amount spent on more general educational resources. Lavy looks at whether a dollar spent on teacher pay is more effective on student outcomes than a dollar spent on books and other resources that a school has, but which do not go to teachers directly. Holding costs constant in making comparisons between compensation schemes, while seemingly obvious, is an essential ingredient of experimental design.

Markets and the Amount of Compensation: The Case of Managerial Pay The literature on pay structures suggests that relative compensation often

provides positive motivation. But many are shocked by the high levels of managerial compensation, particularly for CEOs, prevalent in the United States. Are these high levels of pay justified? Do they provide appropriate incentives and attract needed talent?

There is an economic literature that criticizes the pay of CEOs and other high- level executives, claiming that it is too high. Overcompensation is attributed to the CEOs ability to capture the board of directors or to social pressure to pay salaries commensurate with some norm. It is argued that compensation at the top of the firm is pushed beyond its efficient level, creating too large a spread in the tournament prize structure. Stated analytically, the argument is that even if the high compensa- tion results in additional effort or better talent, the compensation premium does not produce enough additional value to cover its cost.

Research on the relation of managerial compensation to social factors attracted attention in the late 1980s. O’Reilly, Main, and Crystal (1988) presented evidence that the pay of chief executive offers was correlated with the compensation of direc- tors. Their favored interpretation was that the norms for compensation come at least in part from introspection, and directors use their own pay as a frame of refer- ence when setting compensation for a top corporate officer. The implication is that some chief executive officers may be overpaid simply because they are fortunate in having a board with highly paid directors. Another interpretation, however, is that this correlation is created by sorting. Top companies recruit top directors who are highly paid, both in their regular jobs and as board members. The same compa- nies also recruit the most talented chief executive officers, and the market for such

Compensation and Incentives in the Workplace 211

talent gives them high pay as well. If this is true, the pay of chief executive officers and director at a firm would be correlated, but that correlation could be appro- priate and profit-enhancing.

A large literature connects pay of chief executive officers to firm size, and most of that literature argues that the connection can be justified. The most talented manager is more effectively utilized when running a multi-billion dollar company than when running the local hardware store. If a very able chief executive officer increases the value of a company by 10 percent more than that of an average alter- native, this will have a larger effect in companies that have more capital (Rosen 1976; Gabaix and Landier 2008, and the references therein).

In yet another critique, Bebchuk, Cremer, and Peyer (2011) find evidence that the larger the share of chief executive officer pay among the top five executives in the firm, the lower the profit and efficiency that result. Their preferred interpreta- tion is that the excessive share going to the chief executive officers does not enhance shareholder nor worker value. Perhaps, but this pattern could also mean that other executives are underpaid. Alternatively, the pattern could arise from a form of survi- vorship bias in the data. Economic theory suggests that those firms that give their chief executive officers either too high or too low a share will have lower profits than those who pay the optimum share. Undercompensated chief executive officers are more likely to be lured away by the competition. Overcompensated chief executive officers lower the profitability of the firm, but that factor alone is unlikely to drive a firm out of business because compensation for chief executive officers is generally a small part of total costs to the firm. But because of this survivorship bias, there will be fewer data points below the optimum than above, even if there is no built-in bias toward overcompensation.

Setting pay for corporate executives is an inexact science applied in a continu- ally evolving corporate landscape, so it is unsurprising that some chief executive officers will appear overpaid or underpaid, especially when the case is viewed in hindsight. But the evidence that chief executive officers receive pay that exceeds that which is consistent with profit maximization is neither unequivocal nor compelling. At the same time, the literature that presents evidence that rationalizes the high pay of chief executive officer and other managers, while of high quality, remains limited.

Conclusion

Compensation and its structure have profound effects on worker motivation. The theory of incentives laid out decades ago and refined in recent years continues to garner support as more firm-based data have become available.

■ I am grateful to Erielle Davidson for her extensive work reviewing the literature that is discussed in this essay.

212 Journal of Economic Perspectives

References

Bandiera, Oriana, Iwan Barankay, and Imran Rasul. 2005. “Social Preferences and the Response to Incentives: Evidence from Personnel Data.” Quarterly Journal of Economics 120(3): 917–62.

Bandiera, Oriana, Iwan Barankay, and Imran Rasul. 2007. “Incentives for Managers and Inequality among Workers: Evidence from a Firm- Level Experiment.” Quarterly Journal of Economics 122(2): 729–70.

Bartel, Ann P., Brianna Cardiff-Hicks, and Kathryn Shaw. 2017. “Incentives for Lawyers— Moving Away from ‘Eat What You Kill.’ ” Industrial and Labor Relations Review 70(2): 336–58.

BEA Blog. 2012. “What is the Value of Household Work?” June 11. https://blog.bea. gov/2012/06/11/household-work/.

Bebchuk, Lucian A., K. J. Martijn Cremers, and Urs C. Peyer. 2011. “The CEO Pay Slice.” Journal of Financial Economics 102(1): 199–221.

Belzil, Christian, and Michael Bognanno. 2008. “Promotions, Demotions, Halo Effects, and the Earnings Dynamics of American Executives.” Journal of Labor Economics 26(2): 287–310.

Bergson, Abram. 1944. The Structure of Soviet Wages: A Study in Socialist Economics. Cambridge, MA: Harvard University Press.

Bergson, Abram. 1978. “Managerial Risks and Rewards in Public Enterprise.” Journal of Compara- tive Economics 2(3): 211–25.

Booth, Alison L., and Jeff Frank. 1999. “Earn- ings, Productivity, and Performance-Related Pay.” Journal of Labor Economics 17(3): 447–63.

Bridgman, Benjamin, Andrew Dugan, Mikhael Lal, Matthew Osborne, and Shaunda Villones. 2012. “Accounting for Household Production in the National Accounts, 1965–2010.” Survey of Current Business 92(5): 23–36.

Bull, Clive, Andrew Schotter, and Keith Weigelt. 1987. “Tournaments and Piece Rates: An Experi- mental Study.” Journal of Political Economy 95(1): 1–33.

Bureau of Labor Statistics.  2017. Job Openings and Labor Turnover Survey. Washington, D.C.: United States Department of Labor, Bureau of Labor Statistics. https://www.bls.gov/jlt/.

Deci, Edward L. 1972. “Intrinsic Motivation, Extrinsic Reinforcement, and Inequality.” Journal of Personality and Social Psychology 22(1): 113–20.

Drago, Robert, and Gerald T. Garvey. 1998. “Incentives for Helping on the Job: Theory and Evidence.” Journal of Labor Economics 16(1): 1–25.

Ehrenberg, Ronald G., and Michael L. Bognanno. 1990. “Do Tournaments Have Incentive Effects?” Journal of Political Economy 98(6): 1307–23.

Eissa, Nada, and Hilary Williamson Hoynes.

2004. “Taxes and the Labor Market Participation of Married Couples: The Earned Income Tax Credit.” Journal of Public Economics 88(9–10): 1931–58.

Eriksson, Tor. 1999. “Executive Compensation and Tournament Theory: Empirical Tests on Danish Data.” Journal of Labor Economics 17(2): 262–80.

Eriksson, Tor, Sabrina Teyssier, and Marie-Claire Villeval. 2009. “Self-Selection and the Efficiency of Tournaments.” Economic Inquiry 47(3): 530–48.

Falk, Armin, Ernst Fehr, and David Huffman. 2008. “The Power and Limits of Tournament Incentives.” March 28, http://citeseerx.ist.psu. edu/viewdoc/download;jsessionid=F93C0F6D484 8B10E29CA25D037602B3F?doi=10.1.1.724.7243& rep=rep1&type=pdf.

Fernie, Sue, and David Metcalf. 1998. “(Not) Hanging on the Telephone: Payment Systems in the New Sweatshops.” CEP Discussion Paper 390, Centre for Economic Performance, London School of Economics and Political Science.

Fernie, Sue, and David Metcalf. 1999. “It’s Not What You Pay it’s the Way that You Pay it and that’s What Gets Results: Jockeys’ Pay and Performance.” Labour 13(2): 385–411.

Florida, Richard. 2002. The Rise of the Creative Class: And How It’s Transforming Work, Leisure, Community and Everyday Life. New York: Basic Books.

Frey, Bruno S., and Felix Oberholzer-Gee. 1997. “The Cost of Price Incentives: An Empirical Analysis of Motivation Crowding-Out.” American Economic Review 87(4): 746–55.

Frey, Bruno S. 1997. Not Just for the Money: An Economic Theory of Personal Motivation. Cheltenham, UK: Edward Elgar.

Gabaix, Xavier, and Augustin Landier. 2008. “Why has CEO Pay Increased So Much?” Quarterly Journal of Economics 123(1): 49–100.

Gaynor, Martin, James B. Rebitzer, and Lowell J. Taylor. 2004. “Physician Incentives in Health Maintenance Organizations.” Journal of Political Economy 122(4): 915–31.

Giving USA. 2017. “Giving USA 2017: Total Charitable Donations Rise to New High of $390.05 Billion.” Press Release. Giving USA, June 12. https://givingusa.org/giving-usa-2017-total-chari- table-donations-rise-to-new-high-of-390-05-billion/.

Gneezy, Uri, Muriel Niederle, and Aldo Rustichini. 2003. “Performance in Competitive Environments: Gender Differences.” Quarterly Journal of Economics 118(3): 1049–74.

Goldin, Claudia. 1986. “Monitoring Costs and Occupational Segregation by Sex: A Historical Analysis.” Journal of Labor Economics 4(1): 1–27.

Edward P. Lazear 213

Green, Jerry R., and Nancy L. Stokey. 1983. “A Comparison of Tournaments and Contracts.” Journal of Political Economy 91(3): 349–64.

Gruber, Jonathan. 2000. “Disability Insurance Benefits and Labor Supply.” Journal of Political Economy 108(6): 1162–83.

Hamilton, Barton H., Jack A. Nickerson, and Hideo Owan. 2003. “Team Incentives and Worker Heterogenity: An Empirical Analysis of the Impact of Teams on Productivity and Participation.” Journal of Political Economy 111(3): 465–97.

Hass, Lars Helge, Maximilian A. Müller, and Skrålan Vergauwe. 2015. “Tournament Incentives and Corporate Fraud.” Journal of Corporate Finance 34: 251–67.

Heyman, Fredrik. 2005. “Pay Inequality and Firm Performance: Evidence from Matched Employer-Employee Data.” Applied Economics 37(11): 1313–27.

Hölmstrom, Bengt. 1979. “Moral Hazard and Observability.” Bell Journal of Economics 10(1): 74–91.

Hölmstrom, Bengt. 1982. “Moral Hazards in Teams.” Bell Journal of Economics 13(2): 324–40.

Hrywna, Mark. 2017. “Volunteer Time Value Pegged at $193.0 Billion.” NonProfit Times, April 20. http://www.thenonprofittimes.com/news- articles/volunteer-time-value-pegged-193-billion/.

Huck, Steffen, Andrew J. Seltzer, and Brian Wallace. 2011. “Deferred Compensation in Multi- period Labor Contracts: An Experimental Test of Lazear’s Model.” American Economic Review 101(2): 819–43.

Hutchens, Robert M. 1987. “A Test of Lazear’s Theory of Delayed Payment Contracts.” Journal of Labor Economics 5(4): S153–S170.

Jackson, Matthew O., and Edward P. Lazear. 1991. “Stock, Options, and Deferred Compensa- tion” In Research in Labor Economics, vol. 12, edited by Ronald G. Ehrenberg, 41–62. Greenwich, CT: JAI Press.

Jirjahn, Uwe, and Kornelius Kraft. 2007. “Intra- Firm Wage Dispersion and Firm Performance: Is There a Uniform Relationship?” Kyklos 60(2): 231–53.

Kale, Jayant R., Ebru Reis, and Anand Venkateswaran. 2009. “Rank-Order Tournaments and Incentive Alignment: The Effect on Firm Performance.” Journal of Finance 64(3): 1479–1512.

Kandel, Eugene, and Edward P. Lazear. 1992. “Peer Pressure and Partnerships.” Journal of Political Economy 100(4): 801–817.

Kaufman, Alan S., and John L. Horn. 1996. “Age Changes on Tests of Fluid and Crystallized Ability for Women and Men on the Kaufman Adolescent and Adult Intelligence Test (KAIT) at Ages 17–94 Years.” Archives of Clinical Neuropsychology 11(2):

97–121. Kini, Omesh, and Ryan Williams. 2012. “Tour-

nament Incentives, Firm Risk, and Corporate Policies.” Journal of Financial Economics 103(2): 350–76.

Krueger, Alan, and Alexander Mas. 2004. “Strikes, Scabs, and Tread Separations: Labor Strife and the Production of Defective Bridgestone/ Firestone Tires.” Journal of Political Economy 112(2): 253–89.

Kruse, Douglas. 2016. “Does Employee Owner- ship Improve Performance?” IZA World of Labor. December. https://wol.iza.org/articles/does- employee-ownership-improve-performance/long.

Lavy, Victor. 2002. “Evaluating the Effects of Teachers’ Group Performance Incentives on Pupil Achievement.” Journal of Political Economy 110(6): 1286–1317.

Lazear, Edward P. 1979. “Why Is There Manda- tory Retirement?” Journal of Political Economy 87(6): 1261–84.

Lazear, Edward P. 1981. “Agency, Earnings Profiles, Productivity, and Hours Restrictions.” American Economic Review 71(4): 606–20.

Lazear, Edward P. 1986. “Salaries and Piece Rates.” Journal of Business 59(3): 405–431.

Lazear, Edward P. 1989. “Pay Equality and Industrial Politics.” Journal of Political Economy 97(3): 561–80.

Lazear, Edward P. 1999. “Personnel Economics: Past Lessons and Future Directions.” Presidential Address to the Society of Labor Economists, deliv- ered in San Francisco on May 1, 1998. Journal of Labor Economics 17(2): 199–236.

Lazear, Edward P. 2000a. “Economic Impe- rialism.” Quarterly Journal of Economics 115(1): 99–146.

Lazear, Edward P. 2000b. “Performance Pay and Productivity.” American Economic Review 90(5): 1346–61.

Lazear, Edward P. 2000c. “The Power of Incen- tives.” American Economic Review 90(2): 410–14.

Lazear, Edward P. 2017. “The Incredible Shrinking Workforce.” Wall Street Journal, December 7. https://www.wsj.com/articles/the- incredible-shrinking-workforce-1512692004.

Lazear, Edward P., and Sherwin Rosen. 1981. “Rank-Order Tournaments as Optimum Labor Contracts.” Journal of Political Economy 89(5): 841–64.

Lazear, Edward P., Kathryn Shaw, and Chris Stanton. 2015. “The Value of Bosses.” Journal of Political Economy 33(4): 823–61.

Marshall, Alfred. 1890. Principles of Economics. New York: Macmillan and Co.

Mas, Alexandre. 2006. “Pay, Reference Points, and Police Performance.” Quarterly Journal of

214 Journal of Economic Perspectives

Economics 121(3): 783–821. Mas, Alexandre, and Enrico Moretti. 2009.

“Peers at Work.” American Economic Review 99(1): 112–45.

Mobbs, Shawn, and Charu G. Raheja. 2011. “Internal Managerial Promotions: Insider Incen- tives and CEO Succession.” Journal of Corporate Finance 18(5): 1337–53.

Moretti, Enrico, and Jeffrey M. Perloff. 2002. “Efficiency Wages, Deferred Payments, and Direct Incentives in Agriculture.” American Journal of Agricultural Economics 84(4): 1144–55.

Nalebuff, Barry J., and Joseph E. Stiglitz. 1983. “Prizes and Incentives: Towards a General Theory of Compensation and Competition.” Bell Journal of Economics 14(1): 21–43.

Neumark, David, and Wendy Stock. 1999. “Age Discrimination Laws and Labor Market Efficiency.” Journal of Political Economy 107(5): 1081–1125.

O’Reilly, Charles A., III, Brian G. Main, and Graef S. Crystal. 1988. “CEO Compensation as Tournament and Social Comparison: A Tale of Two Theories.” Administrative Science Quarterly 33(2): 257–74.

Rosen, Harvey S. 1976. “Taxes in a Labor Supply Model with Joint Wage-Hours Determina- tion.” Econometrica 44(3): 485–507.

Rosen, Sherwin. 1974. “Hedonic Prices and Implicit Markets: Product Differentiation in Pure Competition.” Journal of Political Economy 82(1): 34–55.

Ross, Stephen A. 1973. “The Economic Theory of Agency: The Principal’s Problem.” American Economic Review 63(2): 134–39.

Roy, Donald. 1952. “Quota Restriction and Goldbricking in a Machine Shop. American Journal of Sociology 57(5): 427–42.

Roy, Donald. 1954. “Efficiency and ‘The Fix’: Informal Intergroup Relations in a Piecework Machine Shop.” American Journal of Sociology 60(3): 255–66.

Ruth, Jan-Erik, and James E. Birren. 1985. “Creativity in Adulthood and Old Age: Relations to Intelligence, Sex, and Mode of Testing.” Inter- national Journal of Behavioral Development 8(1): 99–109.

Ryan, Joseph J., Jerome M. Sattler, and Shane J. Lopez. 2000. “Age Effects on Wechsler Adult Intelligence Scale-III Subtests.” Archives of Clinical Neuropsychology 15(4): 311–17.

Shearer, Bruce. 2004. “Piece Rates, Fixed Wages, and Incentives: Evidence from a Field Experiment.” Review of Economic Studies 71(2): 513–34.

Smith, Adam. 1776. The Wealth of Nations. Scot- land: W. Strahan and T. Cadell.

Titmuss, Richard M. 1970. The Gift Relationship: From Human Blood to Social Policy. London: Unwin & Allen.

Weitzman, Martin L. 1984. The Share Economy: Conquering Stagflation. Cambridge, MA: Harvard University Press.

Williamson, Oliver E. 1975. Markets and Hierar- chies: Analysis and Antitrust Implications. New York: Free Press.

Williamson, Oliver E. 1985. The Economic Institu- tion of Capitalism. New York: Free Press.

Zinsmeister, Karl. 2016. The Almanac of American Philanthropy. Washington, DC; The Philanthropy Roundtable.