Abstract Advances in economic theory help us rethink the traditional public administration concern for accountability and performance in government. Reforms in government have concentrated on organizational designs that flow from piece-rate approaches to employee compensation, but they have largely ignored the prospects for incentive-compatibility within traditional personnel systems. There are important reasons to believe that competitive tournaments in public organization hierarchies, perhaps implemented in promotion systems, could be more effective than the pay- for-performance systems often called for in traditional principal-agent approaches, and therefore can be a useful component of the design of bureaucracies. More importantly, knowledge about tournaments in organizations helps us reconsider key institutional features of public bureaucracies. Introduction Even in the early 1980s, limited attention was paid to the role of incentives in the public sector (Atkinson and Stiglitz 1980), but of course “pay-for-performance” is now central to debates about public personnel compensation. Scholars often point out the need for and usefulness of high-powered rewards for assessing and rewarding public employees (e.g., National Academy of Public Administration 2004; Partnership for Public Service 2005; Risher and Fay 1997; Risher 2004). The roots of this movement, reflected in reform movements such as the Government Performance and Results Act, are found in the economics literature on the firm (often described as the “principal-agency” model of supervision). Principal-agency theories usually center on a core dilemma of how to get an employee or subordinate (agent) to act in the best interests of the employer (principal) when the employee has an informational advantage over and different interests from the principal (Sappington 1991). For bureaucracies, this focus has been on getting principals to control agents through approaches based on the logic of pay-for-performance systems (such as bonuses based in the historical use of “piece rates”) (Moe 1984). For many people in both academia and policy circles, the gold standard of “incentive compatibility” is having specific, countable rewards for measurable increments in government performance. As an example, some have called for making compensation for teachers a linear function of student achievement (e.g., Lazear 2003). Public debates have centered around the appropriateness of pay-for-performance schemes for public sector settings. There is good evidence that political and ideological motives have been very important—perhaps more important than technical concerns—in implementing such pay systems (Kellough, Nigro, and Brewer 2010). Indeed, despite expectations that pay-for-performance systems would improve overall organizational performance, most evaluations of merit-based pay in the public sector suggest less-than-stellar results (Perry, Mesch, and Paarlberg 2006; Perry, Engbers, and Jun 2009). Indeed, although there is little systematic evidence on the effects of pay-for-performance at the individual level in the federal sector, one study indicates that such systems reduce individual-level satisfaction (especially with the organization itself) (Choi and Whitford 2017). In this essay, I depart from that tradition by suggesting that pay-for-performance based on piece rates is not a first-best solution in many specific work settings. Instead, many public personnel systems can allocate rewards through “rank-order tournaments.” Specifically, the rewards an employee receives in a rank-order tournament depend only on how her performance compares to that of others (her ordinal rank). Under fairly general conditions, rank-order tournaments are incentive compatible: people do what the mechanism designer thought they would do (Lazear and Rosen 1981). More importantly for the study of bureaucracies, tournament theory can help us better understand a number of key aspects of public agencies and how they implement public policy. Students of bureaucracies understand the political benefits of delegation, yet even the technical literature largely ignores the incentive-compatibility aspects of traditional personnel systems. High-powered incentives like private-sector bonuses based on share price performance are rarely seen in the public sector, so people try to better monitor or select bureaucrats to make them “principled agents” (Besley 2006; Brehm and Gates 1999; DiIulio 1994). Governments sometimes use monetary incentives, as in the case of the Senior Executive Service, but most designers see problems in their application: bureaucrats have too many bosses, must pursue too many goals, or face complicated “team production functions” (Dixit 2002). Seeing problems with piece rates, we often throw out all of the economics of the firm (including the theory of tournaments), and instead turn to finding agents who have “public service motivation” (PSM) (Perry and Wise 1990; Rainey 1982; Romzek 1990) (although it is becoming increasingly clear that PSM is not a “silver bullet” solution to the problem of motivating public employees; see Kjeldsen and Andersen 2013, Schott and Ritz 2017). Promotion tournaments can be powerful complements to the pay-for-performance systems based on piece rates that compensation experts have proposed in government reforms. Some bureaucracies already have built-in technologies that can allow for tournaments, including public personnel systems that rely on promotion, career ladders, and internal job markets. Evidence on tournaments supports their consideration (Brown et al. 1996; Chevalier and Ellison 1997; Ehrenberg and Bognanno 1990a; Eriksson 1999). Yet, their performance depends on the specific features of individual bureaucracies and the conditions those agencies face when implementing policy, and these fine distinctions have important implications for our understanding of those agencies’ design and performance. The theory of tournaments has both normative and positive features: it describes optimal incentive-compatibility mechanisms, and it explains the motives behind the mechanisms people put in place in real organizations. Normative features help describe the prospects for tournaments as alternatives to systems built on piece rates. Positive features help us better understand the incentive-compatibility features of public personnel systems already in place. Likewise, tournaments come in many flavors. Promotion in a hierarchy may be a useful example of a rank-order tournament; it is not the only one. Promotions can be seen through the lens of tournaments, but since others exist, it is important to not equate promotions with competitive tournaments. The essay proceeds as follows. The next section offers a survey of tournaments as a solution to incentive-compatibility problems in organizations. The third section describes two key dilemmas: accounting for heterogeneity among participants, and constructing “incentives for losers.” I then review empirical evidence for the power of tournaments in experimental and other settings. The fifth section extends these lessons to the case of the public sector. Last, I offer a series of specific implications of a tournament-oriented analysis of motivation for our understanding of the public sector and the implementation of public policy. Incentives in Tournaments In this section I summarize several key analytic results from economic theories of compensation systems. First, I describe various ways to compare piece rates and rank-order tournaments; essentially, they solve performance measurement problems in different ways. After that, I turn to two key dilemmas encountered when making the theory of tournaments more realistic. The general problem is always to make the theory more realistic while retaining tractability to account for the different ways tournaments can address incentives problems. In the theory of the economics of the firm, the principal’s goal is to pay a worker a wage equal to the value of the worker’s marginal product. Piece rates align the worker’s incentives with the organization’s goals (Mirrlees 1976; Stiglitz 1975). Specifically, the amount paid depends on the number of pieces produced so that the agent is automatically rewarded (or punished) for changes in performance. In this typical design, piece rates cause the worker to act as the principal prefers even though the agent has better information about her own ability (the condition of asymmetric information). When worker effort is partially unobservable, the incentive contract causes agents to reveal their effort (this is a “revelation mechanism”). Traditional organization theory noted the power of piece-rate systems. Traditional manufacturers such as Lincoln Electric, Safelite, and Nucor credibly committed to the use of piece rates, and piece rates delivered impressive productivity gains (Miller 1992; Passell 1996). For example, Shearer (2004) estimates that the average productivity gain of using piece rates over fixed wages is roughly 22%. Over time, firms increased their use of incentives-based contracts for senior management, with the result that both compensation and sensitivity of compensation to performance have increased (Hall and Liebman 1998). Yet, firms often use piece rates less than expected (Baker et al. 1988; Coughlan and Narasimhan 1992; Eisenhardt 1989; Miller 1992). As Passell noted in 1996, “With such shining examples to recommend the practice, the $64 billion question is why the pressure of competition and management’s inclination to tinker have not led more employers to try it. Roughly 15 percent of the labor force now has a portion of its pay connected to individual productivity, but most piecework is concentrated in a handful of low-paying industries – notably apparel – or is confined to sales personnel, individual contractors and other workers who, one way or another, see themselves as separate from shop floor workers” (Passell 1996, A1). There has also been less evidence than expected for high-powered incentives even among chief executive officers; less than 5% of the variance in compensation is related to firm performance, and most of it depends on firm size instead (Tosi et al. 2000).1 The link between executive pay and performance is particularly weak in “managerially controlled firms,” with their dispersed ownership (Hambrick and Finkelstein 1995; Tosi et al. 1997). A cynical reader might point to the recent compensation package for Yahoo CEO Marissa Mayer as evidence of weak incentives. This happens for many reasons. A worker might not perform well if the amount paid per piece is too low (the problem of “goldbricking”) or if the employer might lower future payments after seeing the worker’s capabilities (quota restrictions). A risk-averse agent may demand an “irrationally” high bonus (from the principal’s perspective) if he has very low marginal efficacy or if his marginal effort cost is very high (Miller and Whitford 2002). Even if agents do what they are supposed to do given the contract, a budget-constrained principal might balk at the payment necessary to obtain compliance. As Miller (1992) points out, the problem for a principal is not just deciding the rate, but committing to paying the rate after observing effort. The principal faces his own form of “moral hazard” if he might renege on the agreement made when setting up the original compensation scheme (Miller and Whitford 2007). It makes sense to directly link pay and performance when monitoring is inexpensive. The most important reason why pay-for-performance schemes are less prevalent than expected, though, is that workers are smart to shirk when monitoring is expensive and faulty. High monitoring and measurement costs are central problems when we use cardinal performance measures. In those situations, firms may not want to use piece rates. This observation leads to the innovation of a tournament-based approach to the principal-agency problem. Tournaments in Theory and Practice Consider what many have argued is a natural arena for deploying incentive-compatibility designs that are rooted in principal-agency theory: for motivating those who run public schools. Rather than enumerating all of the ways such thinking has influenced policy designs like those brought about in the wake of the No Child Left Behind Act (Whitford and Jaiani 2011), here are a few examples of how incentives have played out in some important school districts. The Houston Independent School District (HISD) is one of the country’s largest. In 2014, the HISD trustees awarded the superintendent a bonus of $116,000. The maximum he could have received was $130,000, and his 2013 bonus was $125,000. But his contract guaranteed him a bonus of $45,000 as long as his annual evaluation was “satisfactory.” This bonus was on top of his annual base salary of $300,000, plus perquisites. Teachers and school principals are also eligible for bonuses, although those payouts are based entirely on test scores whereas the superintendent’s bonus depended on a blend of items such as student enrollments and the takeover of another ISD (Mellon 2014). It is important to note that in that case it is vague how those individual items were weighted. The 2016 hiring of a new superintendent created opportunities to change that contract. Indeed, the base salary increased to $345,000 (Mellon 2016), plus bonuses, and that number was lower than other comparable districts in the area (Oberg and Seibert 2016). The most important aspect of that negotiation, in comparison to the previous contract and other comparable contracts in the area is that “There doesn’t appear to be to be a boilerplate method to determine a superintendent’s salary” in any of these cases (Oberg and Seibert 2016). A local lawyer who helped negotiate many of those contracts noted at the time “It gets more into a personal formula. You want this person? This is what it’s going to take to bring them” (Oberg and Seibert 2016). These examples show how difficult it can be to deploy the kinds of bonuses based on strict attention to metrics that are fundamental to the pay-for-performance paradigm given its roots in piece rates. The contracts themselves often depend on outside market forces—a beauty contest, of a sort. The metrics themselves are often conflated with other possible ways of measuring “performance,” and the weights across different factors are mostly left-unstated. Although it is possible for even weak systems to have weakly positive results (e.g., Meier and O’Toole 2002), there is always the possibility that political principals, once the contract is in place, have incentives to undercut the contract in ways that create credibility problems for the “symmetric verifiability” of contract items (Miller and Whitford 2007; Miller and Whitford 2016). In New Jersey, there are useful examples of bonus systems put in place as a way of capping other compensation for superintendents (Izzo 2013). At a time when base salaries were increasing at a rate that politicians found unpalatable, efforts were made to replace base with bonus-based compensation—though it is an open question of whether those bonuses are based on objective metrics or something else. The discussions surrounding the systems often focus on whether bonuses are set so that the performance is actually part of the minimum expectations, where those levels are attainable in most all cases. There has also been discussion of whether metrics were chosen so that the superintendent could set the conditions of success (e.g., a metric to lower the suspension rate when the superintendent is solely in charge of suspensions). One indication about the nature of how difficult it is in practice to get the details right: a teacher’s representative noted that the system “gave the fox the key to the chicken coop” (Izzo 2013). These examples show some of the difficulties of setting bonuses in real settings when the goal is to align cardinal measures of performance and compensation. In contrast, principals are better off making rank-order (or ordinal) comparisons of workers because it changes the costs of measurement and the risk workers bear (Lazear and Rosen 1981). In an optimal compensation contract, rewards depend on a variable that is informative (even incrementally) about the agent’s effort (the “informativeness principle”) (Holmström 1982). Ordinal rankings can serve as that variable instead of cardinal measures of worker performance. Traditional examples of ordinal rankings outside workplace settings include IQ scores and class ranks. The difference between a cardinal measure and an ordinal measure of performance is fairly intuitive. Think back to an early research methods class. The four basic types of measurement scales are nominal, ordinal, interval, and ratio. Nominal scales are like labels: “did this” or “did that”. Ordinal scales require a bit more information. In this case, we can rank “did this” and “did that” in terms of their importance, value, usefulness, or some other ordering that addresses underlying organizational goals. Information needs are greater in the case of interval and ratio measures. For the first, we know not only differences and ordering but also exact differences between the points on the scale that underlies our observed measurements. This is probably what many think underpins the use of test scores as a measure of student performance: that when someone gets a test score that is 10 points higher than an earlier score, that the person has demonstrated improvement in their understanding of material—and also that the improvement is useful for accounting for the quantity of improvement (It is striking, though, that in most cases we take an ostensibly interval-level measurement of score and turn it into an ordinal one—such as when a 92 becomes an “A” whereas an 88 becomes a “B”.). Ratio scores require even more information because in this case there is ordering information, relative differences information, and there is also an absolute zero on the scale that have meaning. Measurements can be added, subtracted, divided, and multiplied. In canonical pay-for-performance based on piece rates, ratio scales are the numbers of widgets produced. In traditional time-motion studies, they are the time required to complete a task. In scientific management, ratio scales are built into our understanding of the number of arm movements it takes to move a product from a raw to a finished state. But unfortunately, although ratio scales as measures provide exactly the information we need to implement piece-rate systems, they also require substantially more information than we often have in real work settings—and this is especially the case in the production settings that we often see in real public agencies. In contrast, the winners of a tournament receive prizes based on their relative position; essentially, competitors are judged against a standard that is a random variable (Green and Stokey 1983). Both tournaments and piece rates, along with other schemes, are Pareto optimal (no individual can be made better off without another being made worse off) when workers are indifferent to risk (risk-neutral) (Lazear and Rosen 1981). Because both are optimal; what really matters is how much each costs to gather information and measure performance. The main advantage, then, is that tournaments and their ordinal rankings are more efficient than pay-for-performance systems that require cardinal measurements because cardinal scales require more information. Moreover, the designer can pick from many different tournament schemes (paired comparisons, round robins, knockout tournaments with single and double elimination, etc.) because they are all good enough (Lazear and Rosen 1981). A rank-order tournament may be more efficient than a piece-rate mechanism when it is difficult to measure individual performance. If we had one universally-accepted cardinal measure of performance, then traditional performance systems based on incentives would be much easier to implement. However, cardinal measures might be especially difficult in the public sector, so tournaments might solve the manager’s problem of relating rewards to the measured performance of public sector employees (Ammons 1992; Behn 2003; Kaplan and Norton 1992; Kravchuk and Schack 1996). To be clear: it is hard to find good performance measures. We might instead have multiple indicators with none that fully characterizes an employee’s output, but then the problem is aggregating the indicators into a meaningful cardinal measure. Even if we could agree on a set of imperfect underlying measures, none of which fully encapsulates public worker performance, the problem of identifying an agreed-upon set of weights. This is a well-known problem in decision theory, as exemplified in multiple-attribute utility theory where the weights can play the most important role in determining the specific overall ranking that emerges from a given set of underlying indicators (e.g., Keeney and Raiffa 1993). The overall point here is that having multiple measures does not solve either the problem of imperfect cardinal measures or perfect cardinal measures that are excessively expensive. We need ordinal measurements for tournaments for public sector employees, itself a tough problem, but we are more likely to have those measures than single or multiple cardinal indicators. The discussion above about incentives in schools reveals the cost and difficulty in obtaining useful cardinal measures. In the case of superintendents, the measures are multiple, mostly ordinal (“successfully solving a problem with a building”), or rely on uncertain or unstated weightings; in this case, there is little comparative performance information because boards rarely employ two competing superintendents. The superintendent is awarded largely by competing against an unknown random variable, as discussed above. Perhaps with teachers the problem is easier? Teachers could be compensated on the basis of students who pass the course, who pass an externally-administered end-of-course test, who graduate, who can read at a given level, who have made substantial improvement in a defined period of time, who master a specific technical competency with proven benefits in labor markets, or perhaps even whether they care enough to attend class regularly. Each of these could be forced into a cardinal measurement of productivity based on thinking drawn from ratio variables. Debates rage about the relative importance of each of these measures, largely uninformed about the individual measures’ relative efficacy for improving public education; even evaluation practitioners are loath to say that one measure has twice or three times the impact of another on the overall goal of “public education.” As such the weights are effectively ordinal statements about contribution (Even college professors are known for “teaching excellence” largely on the basis of ordinal statements of quality rather than a ratio-based scale.). Relative ranks also have other key advantages for setting incentives (Carmichael 1983; Green and Stokey 1983; Lazear and Rosen 1981; Malcomson 1984; O’Keefe et al. 1984). For instance, tournaments can be handy for sorting more able workers on ability and still providing incentives (Rosen 1982; Sattinger 1993). Also, in tournaments everyone is similarly affected by common productivity shocks (Prendergast 1999). Some have argued that a third advantage is that tournaments appear in the real world since firms often use promotion as an incentive or key reward (Green and Stokey 1983; Prendergast 1999). It is worthwhile to note a key distinction between tournaments and pay-for-performance systems like piece rates. Rank-order tournaments can lead to highly skewed prize structures.2 There are many views on skewed wage distributions in hierarchies (Qian 1994; Waldman 1984). Piece rates are linear transformations of output: each additional unit of performance should translate into an additional unit of reward. In contrast, tournaments are highly nonlinear: an extra unit of performance does not necessarily translate into an extra unit of reward because the units are essentially unobservable. Instead, rewards come only when an individual’s ordinal performance ranks highly compared to other individual performances. In practice, real piece rates often have caps or floors, which make them discontinuous or nonlinear in rewards, but those wrinkles usually just cause trouble if one wants strong incentives. Both nonlinear tournaments and piece rates are exactly the same from the perspective of the principal because they cause the same outcome on the part of the agent. The approaches have identical expected utility and investments by participants; the schemes have the same statistical “first moments” in the distributions of their rewards and effort. I want to emphasize four key types of problems with piece rates that tournaments help alleviate. Specifically, tournaments can credibly commit principals to the task of rewarding productive agents. In a typical pay-for-performance system, cardinal measures of performance must be verifiable (symmetrically available) to the principal and the agent so that contracts are legally enforceable (Grossman and Hart 1983); pay-for-performance is not credible if the agent cannot verify the principal’s observation, which often is judged subjectively (Malcomson 1984; Williamson et al. 1975). Instead, a principal could credibly commit to a tournament and prize structure, along with relying on ordinal rankings (Carmichael 1983). For example, a principal might always have an “employee of the month” award, with accompanying benefits. There might be differences of opinion about who should win, but the principal might at least commit to always awarding the prize. A second way that tournaments mitigate problems with piece rates is by sorting workers into those who are more and less able instead of just concentrating wages around a mean; in tournaments, principals distinguish between workers at the same time they determine wage differentials. Of course, promotion usually brings responsibility and higher wages. Essentially, tournaments can help discourage more able workers from the rent seeking and influence activities that they often engage in under a linear piece rate system (Fairburn and Malcomson 2001). Although there are always opportunities for rent-seeking, the prizes awarded in tournaments (e.g., becoming supervisor) might be high and disproportionate enough (i.e., “skewed”) that it increases the costs of awarding the prize on the basis of rent-seeking alone. Third, tournaments are built on relative performance evaluations, which are inherent in assessments of subjective performance. These also happen when evaluators use a forced mean or a curve to evaluate members of a group. Relative evaluations can filter out risk or unfairness that results when a tough evaluator punishes a worker whereas another worker gains from having a lenient evaluator. Relying on an evaluation distribution that is at least partially relative can reduce the problem of varying evaluator stringency. Likewise, ranking can also reduce the chance that evaluators will give everyone the same evaluation because of “centrality” or “leniency” biases (Gunderson 2001; Prendergast 1999). Finally, tournaments are relatively more powerful than individual contracts when a single risk-neutral principal interacts with several risk-averse agents. Managers can use independent contracts when there are no team production externalities, but multi-agent settings make it hard to attribute performance, effort, and rewards (Holmström 1982; Miller 1992, 2000; Nalebuff and Stiglitz 1983). Specifically, a tournament can relay information about the common elements of each individual’s performance while eliminating noise common to all team members’ production (Green and Stokey 1983), although as a consequence it may increase the randomness in any one agent’s compensation. Tournaments can work better than independent contracts when the principal oversees many agents. In sum, tournaments are a different way to use incentives when it is difficult to monitor agents. They rank-order participants to pick winners; in practice, relative performance is a key signal that employees can provide to their supervisors. This reinvigorates traditional debates about promotion systems, career ladders, and other internal labor markets in both the private and public sectors because the theory of rank-order tournaments, dating to Lazear and Rosen (1981), shows that these compensation systems may be incentive-compatible. Two Dilemmas: Heterogeneity and Incentives for Losers Yet, it is important to consider two dilemmas in applying the theory of tournaments to real-world phenomena like promotion, career ladders, other internal labor markets, and other competitive ranking systems in the public sector. The first dilemma comes when applying tournaments to groups of people who are different from one another—or in technical language, heterogeneous—in their abilities, tastes for risk, or utility valuations. The second dilemma is located after the tournament occurs, and it comes in two flavors: in how to shape incentives for winners after they win the tournament, and in shaping incentives for losers after they lose it. How well does a tournaments-based approach for motivating public employees handle these dilemmas? The first problem is the heterogeneity of contestants (in their ability, risk aversion, or disutility of effort). This problem is so fundamental for the theory of tournaments that theorists have struggled for almost 30 years with it, and yet solutions remain elusive. Specifically, in theory, heterogeneity undermines incentives for most types of workers. The upshot is that tournaments may not provide incentives unless the organization can run the competition within a homogeneous group. For instance, this has been described as the aggressive sabotage of the organization by less aggressive workers (Lazear 1989). Essentially, when there are two types of people in the organization, the relative comparisons built into tournaments give a specific type of people incentives to spend time making high-quality workers look bad. The result is a degree of wage compression, which has the residual effect of limiting the power of the tournament to incentivize competitors. In a second example, because agents are often relatively more risk averse than principals, there is an “insurance/incentives tradeoff” in setting incentives. Risk-aversion can cause an agent to reject the bonus principal-agency theory says should be sufficient to get them to do what a principal wants (Laffont and Martimort 2002; Miller and Whitford 2002). Generally speaking, the performance of tournaments depends on what agents want in terms of their goals (their utility functions) and their expectations about how their individual performance ranks (their likelihood of winning the tournament) (Lazear and Rosen 1981). The problem is that relatively more risk-averse workers may see tournaments as safer than other incentive plans; it may be this is the stereotype of the bureaucrat seeking a steady, but low income over the risks inherent in the market. Can we design a mechanism that accounts for these perverse incentives, but retains the main advantages of a tournaments-based design for public organizations? As noted above, theorists want a mechanism that is tractable, can be implemented in the real world, and accurately reflects the underlying choice problem. In the case of heterogeneity, one mechanism is a “within cohort” tournament, in which workers of a specific type compete, rather than one that gathers together all types in a single competition. Essentially, the idea is to get the different types of workers to provide first-best effort levels and to self-select into ability-specific contracts (Bhattacharya and Guasch 1988). These kinds of evaluations would help the whole organization by pushing “positive production externalities” downward in the hierarchy, causing workers to group and compete; yet, these kinds of evaluations are hard to make because noise is inherent when grading workers. Most importantly, this mechanism shows the problem of designing a blanket solution. The Bhattacharya and Guasch paper is complex and it is unclear how one should implement their method in real organizations, including firms. We might supplement it with “across cohort” tournaments, such as between work groups, since these help reduce suboptimal helping and free-riding (Drago et al. 1996); an example is the common ranking and rewarding of schools based on average test scores. Yet, this is a second-best solution. A second possible design comes from the common concern in real work settings that risk-averse or low-quality workers will select into the tournament system. In essence, the growing number of contestants reduces each person’s effort when people are free to enter the tournament (Taylor 1995). In fact, having only two competitors may be best in some tournaments if they are heterogeneous (Fullerton and McAfee 1999). Since a tournament is really just a kind of horse race, one could have knockout rounds of pairwise competition, just as one sees in sports (e.g., the NCAA tournaments). The principal’s problem is different in that case: it is to select the best pairing of contestants. Competitive handicapping helps if a low-quality worker tries to infiltrate a high-quality organization, and in fact this is used in the case of basketball. There are few examples of pairwise knockout competitions in firms or other work settings, although Tong and Leung (2002) provide evidence that these kinds of “dynamic tournaments” work well at inducing effort. This also leads to an impasse: since homogeneity requires sorting (in the form of handicapping), this strategy runs headlong into another main purpose of tournaments: sorting, as in “promoting the best performers into jobs of greater importance.” Below I describe evidence about how heterogeneity affects tournaments, but it is important to recognize that one cannot simply assume away the issue in practice. The theoretical problem of designing for heterogeneous agents is central to the development of a tournaments-based approach to incentives. The second dilemma is equally important for applying the theory of tournaments to the real world of the public sector. The two parts relate to the two kinds of people (winners and losers) that emerge from the tournament when it is used to motivate. First, can tournaments continue to induce quality effort levels for winners (those workers who have been rewarded)? Specifically, the fear is that winning contestants will “rest on their laurels”. Mitigating the problem for winners is related to the wage skews described above—by giving successive prizes disproportionate weight (Rosen 1986). The disproportionate weight for those higher prizes at higher levels of the organization is the same natural wage skew seen in most of these kinds of tournaments (Lazear and Rosen 1981). For instance, a tournament with an elimination design has to provide extra awards so that those who survive into the next round still face the necessary incentives (and that next round of awards is not connected to past performance and accomplishments). Again, in incentive systems with tournaments (e.g., promotion systems, internal markets, and career ladders), skew is both natural and desirable. The second part is mitigating the serious problem with tournaments of motivating those passed over for reward, and this is related to the bigger problem of heterogeneity. We can call it the “incentives for losers” problem (Baker et al. 1988). Just as in the previous situations where we saw secondary tournaments as solutions, likewise additional layers of tournaments offer possible second-best solutions. For example, principals might use a series of tournaments, so that first-round losers have opportunities to succeed (assuming improved performance) in the second round. A sports analogy from soccer is when early losers in the Champions League in European club soccer then can play in a second tournament, competing against teams of lower ability. In the public sector, it is common for organizations to have tournaments of different sizes: promotions coexist alongside employee-of-the-month awards. Before turning to a description of some of the evidence that has been assembled on tournaments in experiments, firms, and other settings, it pays to emphasize one theme in this section: the use of layered mechanisms to improve the power of incentives. In the case of heterogeneity, the power of tournaments is lower when people are very different, but we might fill gaps with other strategies. The same goes in the cases of incentives for winners and losers. The main point is that in real world applications it is likely that the choice of an incentive mechanism is not an either/or choice—that we will observe layers of organizational mechanisms that interact in part because of the kinds of dilemmas identified in this section. The types of mechanisms the literature in public administration studies these days—such as public service motivation (Perry and Wise 1990; Rainey 1982; Romzek 1990)—is probably most interesting when simple tournaments need a boost due to heterogeneity, or finding “incentives for losers.” Tournaments in Experiments and Firms Tournaments are potentially first-best incentives systems. In addition, tournaments reflect traditional features of labor arrangements in large organizations (Malcomson 1984). Organizations generally have (1) hierarchical wage structures and (2) internal promotion; (3) wages that rise with seniority and promotion (and not necessarily rising production); (4) the variance of earnings that increases with experience; and, (5) wage rates that attach mostly to jobs and not workers. This last point is particularly important in the public sector, with its dependence on civil service regulations for determining wages. Unlike some criticism of “weak incentives” in the public sector, the incentives theory of tournaments is not as troubled by wage differentials being determined mostly by administrative procedures and not by an external market; it just depends on the procedures. Lazear and Rosen (1981) is the basic approach to designing mechanisms that address two basic problems. First, with piece-rate type solutions, managers suffer fundamental information problems—at a minimum, the problem of calculating cardinal performance measurements. Second, because piece-rate mechanisms are not predominant for rewards, the realism of the approach is constantly called into question; if piece rates are the only way to get the incentives right, why are they so rarely used? Tournaments are a fine solution: they are theoretically efficient and can be applied to many types of organizations, including public agencies and nonprofits. For organization theory, the tournaments approach has both desirable normative and empirical properties as an organizing frame of reference. Does the theory of tournaments for incentives have empirical support? Evidence comes from three primary sources. This section first reviews the experimental evidence for tournaments as incentive mechanisms. Then I review nonexperimental evidence in competitive settings and firms. After that, in the following section, I offer empirical observations about the historical and continuing use of tournament-like mechanisms as motivators in the public sector. In a number of important ways, experimental evidence is key for assessing formal theories of incentives and individual rationality (Friedman and Sunder 1994). Early experiments found mean effort levels for both tournaments and piece rates that fit with the theoretical predictions, but they also found that effort levels varied more in tournaments than in piece rate experiments (Bull et al. 1987). Two-thirds of the variance comes from the fact that participants make choices in a “game” environment—that responding to a tournament is not a straightforward maximization problem like that for piece rates; the remaining variance comes from computational difficulty (Drago and Heywood 1989). In tournament experiments without the game framework, so that people played against an unknown cardinal measure, participants balanced the rewards they expected from additional calculation against the possible gains. Although a tournament is a game, figuring out the best response is about as difficult as in a complicated maximization problem. Experiments show tournaments systematically outperform target-based schemes for eliciting effort (Nalbantian and Schotter 1997). One “hard test” shows that participants did not make suboptimal choices in following rounds when they lost in early rounds (Drago and Heywood 1991). This is different than in experiments on traditional incentives (such as contingent wages) that show principals usually do not impose incentives that the theory says are sufficient to motivate a high effort from a rational, self-interested agent (Bottom et al. 2006; Conlon and McLean Parks 1990; Fehr et al. 1997; Fehr and Fischbacher 2003; McLean Parks and Conlon 1995). Those experiments on traditional incentives have driven some to argue that parties in principal-agent negotiations enforce contracts through reciprocity—not through principal-agent type reasoning. The contrast is that in experiments with tournaments people seem to do what the theory says. The second major question for experimentalists is how people act when there are differences—when they are heterogeneous—given the debates reviewed above about how to apply the theory in real-world settings. One important result is that when participants had different abilities, the effort of advantaged participants was as high as predicted by theory, and the effort of disadvantaged participants was higher than predicted (Bull et al. 1987).3 Experiments where differences are controlled by design have shown that tournaments can have some power when participants are heterogeneous. Advantaging the disadvantaged (through handicapping) seems to cause both greater effort by advantaged participants and greater collective effort (Schotter and Weigelt 1992). These studies provide evidence that tournaments can be robust ways to get the incentives right in organizations. However, the evidence is mixed at best on heterogeneity in tournaments. The laboratory experiment referred to above suggests that tournament behavior under heterogeneity does not display the problems theory suggests will result. One lab experiment is not enough to reject the theory. For instance, one can think of cases where contestants “slack off” in a contest when the match is uneven. A better example comes from a study of the conditions under which women and men select into different competitive environments. Niederle and Vesterlund (2007) first “matched” men and women on the basis of ability. Participants in a laboratory experiment were then asked to solve a real task under two different schemes: a noncompetitive piece rate, and after that, a competitive tournament. Niederle and Vesterlund found no differences in performance that were associated with gender. However, they found that men were twice as likely as women to select the tournament as their compensation scheme when asked to do so after experiencing the piece rate scheme. Specifically, 73% of the men in the study selected the tournament scheme. In contrast, 35% of the women selected the scheme. They could not explain the gender gap they discovered in tournament entry by examining key aspects of individual participation and behavior, such as performance. They could not explain the gap by reference to differences in relative risk aversion, a traditional concern in the theoretical literature on heterogeneity. What Niederle and Vesterlund found was that the differences in entry to tournaments were caused by gender-based differences in overconfidence and tastes for performing in a competition. As they argue, “the result is that women shy away from competition and men embrace it” (Niederle and Vesterlund 2007, 1067). This shows how the evidence is mixed at best on heterogeneity in tournaments.4 Although several laboratory experiments suggest that heterogeneity does not cause problems in tournament behavior, other experiments show that participation in tournaments itself is affected by heterogeneity within the set of possible participants. One lab experiment is not enough to reject—or confirm—the theory. Are tournaments effective in real competitions? The first systematic nonexperimental evidence for the power of tournaments came from data from professional golf tournaments; the data show that tournaments cause participants to take costly effort in response to the level and structure of prizes (Ehrenberg and Bognanno 1990a, 1990b). Similar findings are found in other tournament settings including professional tennis (Coate and Robbins 2001), basketball (McClure and Spector 1997), and golf (again) (Melton and Zorn 2000; Moy and Liaw 1998). But the most interesting nonexperimental evidence about tournaments comes from firms themselves. As noted, scholars have been drawn to the theoretical analysis of tournaments because of the widespread use of promotion in firms (Malcomson 1984). Of course, scholars recognized this widespread use and the power of promotion and other competitive forms well before the incentives revolution in compensation research. As Drucker puts it, “proper spirit and performance therefore require a rational promotion system” (Drucker 1954, 154); irrational promotion systems—inefficient ones—are “wasteful of a scarce and expensive resource” (Drucker 1954, 155). Although promotion from within could be the norm, it does not have to be exclusive; in either case, it is a powerful motivator: promotion is a “life and death” decision for a manager (Drucker 1954, 155). However, although the power of promotion schemes for organizational performance has long been recognized, scholars have usually been more skeptical about the effect of losing in tournaments (more so than the above experiments) (March and Simon 1958; Stone 1952). Other studies suggest generally positive effects of tournaments in business settings (e.g., Baker et al. 1994; Bognanno 2001; Conyon et al. 2001). Other studies in organization behavior and compensation theory are also optimistic about the power of tournaments. The motivational role of upward mobility in achievement is widely recognized (Kantner 1993; McClelland 1966; Rosenbaum 1979, 1984). More studies specifically find support on the relevance of promotion systems as tournaments. For instance, a panel study of over 200 firms and 2,000 executives finds systems to be consistent with the operation of tournaments (Main et al. 1993). Another study reports that the probability of promotion strongly affects the ratio of executive salary to lower-level employee salary—executive salaries and bonuses serve as tournament prizes and as managerial incentives (Xu 1997). The predictions of tournament theory are seen in differential pay between job levels, pay dispersion, and firm performance (Eriksson 1999). The exponential growth of large law firms shows that promotion tournaments can be effective mechanisms for obtaining high organizational performance (Galanter and Palay 1991). However, we should temper our acceptance of the evidence for the effects of tournaments in real organizations because other mechanisms often predict the same outcomes also ascribed to tournaments. For instance, Prendergast (1998, 348) points out that the increasing returns to promotion that Eriksson (1999) says are consistent with tournament theory fail “to distinguish tournament theory from the plausible alternative.” In tournaments, pay rises in a convex fashion with a person’s level in the firm; yet, this is also consistent with the sorting of workers over time, through multiple levels of the hierarchy (Rosen 1982); pay may rise with promotion as labor market becomes more certain about a worker’s unknown ability (Waldman 1984). The problem is to identify the “option value” to workers of the prospect of future promotion (Prendergast 1998, 348). One paper on the use of tournaments finds that promotion systems blend tournaments and the use of exogenous standards (Coupé et al. 2006). However, De Varo (2006) shows that promotions are determined by performance, that firms set wage spreads to induce that performance, and that worker effort increases with expected wage spread (Prendergast’s expected option value). This discussion shows how the theory of tournaments has both normative and positive features. On one hand, the theory describes optimal incentive-compatibility mechanisms. On the other hand, scholars use it to explain how people put in place mechanisms in real organizations. The tests described here contribute to both purposes. The normative features of tournament theory describe optimal mechanism design. The tests assess whether the technology accomplishes the outcomes theorists say it will accomplish. At the same time, the tests illuminate different features of real work environments so that we better understand the composite of features already in place. When examining real organizations, most tests turn to promotion as a good example of a rank-order tournament, in part because of data availability. Since others exist, we should be careful not to equate tests that rely on data from promotions with definitive tests of the theory of competitive tournaments. Before turning to an extended discussion of the potential role of tournaments for understanding incentives in government, it is important to restate the true nature of incentives in firms and other private-sector organizations. Rather than a dichotomous choice, the actual use of all mechanisms for gaining compliance is really one of shades of gray. Tournaments (of many types) are blended with piece rates; piece rates are blended with fixed wages; all of these are blended with other mechanisms that do not center on monetary compensation. Ichniowski et al. (1997) estimate the impact of the incentives paradigm in American firms to show that as those firms sought to match the performance of Japanese manufacturers in the 1980s and 1990s, they turned to multiple innovative human resource practices. Rather than choosing one practice, they adopted blends. They first relied on traditional practices such as pay not connected to performance, but then adopted packages of innovations that included problem-solving teams, worker rotation, screening and selection mechanisms, training, and incentive pay in multiple forms. Specifically, in practice we see tournaments that coexist with both fixed pay and incentives. Special Aspects of Tournaments in the Public Sector Historically, promotion has played an important role in the public sector. The nonexperimental evidence reviews above suggests promotion tournaments help induce high effort in firms. This section describes how tournaments might enhance incentive-compatibility in public organization hierarchies. Tournaments can complement the pay-for-performance systems compensation experts have suggested in government reform efforts. The advent in the twentieth century of merit-based competitive hiring and promotion systems fundamentally shifted hierarchical control and compensation in government. The shift from patronage to merit-based personnel systems altered how career public personnel make and implement public policy (Kernell and McDonald 1999; Knott and Miller 1987). Of course, patronage just encompassed a different kind of competition; now, a fundamentally different metric underlies the comparisons and rank ordering of personnel in our modern civil service. The modern civil service changed who wins promotion and why. Promotion always includes both economic and prestige rewards (Barnard 1938; Simon 1945); the modern civil service rewards different people than does patronage because the competition is different in terms of the metric underlying rank orders and the requirements for entry. However, it is important to note the great variety of systems within the United States and across countries. Some points made in this section are even more relevant for non-US federal systems given how the US federal system can be seen as rigid in comparison (e.g., see Choi and Whitford 2011). Throughout this discussion, though, the focus is on mechanisms that have been historically important in the public sector because it is that long tradition that tells us something about efficacy (A critic might point to the history of compensation via “baksheesh,” which can mean corruption and bribery in the public sector, but that discussion is beyond the scope of this essay for obvious reasons.). The significant opportunities for promotion in the public service are the most important evidence of how tournaments might operate in the public sector. There are usually twice as many promotions as recruitments each year in the US federal civil service (Fesler and Kettl 1996, 158). This ratio will surely grow with the impending retirement of many workers over the next 10 years. Agencies have substantial discretion in making these promotions. External candidates are not required. Not all employees must participate. Employees often participate in other groups’ promotion processes when they change job ladders (DiPrete and Soule 1988). In general, promotion in the federal civil service is built on four foundations (Chase 1991; Fesler and Kettl 1996). First, designers of civil service human resource systems believe they are able to get different and better recruits if recruits see the prospects of long careers. Second, because some characteristics cannot be judged by formal examination, performance in the job (and subsequent promotion) is useful for screening. Third, by rewarding competence, designers can enhance supervision, morale, and thus organizational performance. Fourth, each promotion creates a vacancy chain, so one promotion diffuses throughout the system. Each of these foundations shows how promotion tournaments can have incentive consequences in the public sector. As noted in the introduction of this essay, designers have sought to shape the behavior of public sector workers by many different means. Pay-for-performance is only a recent entrant into this contest for “most effective intervention.” PSM is equally if not even more important in this discussion. All such approaches start from an underlying model of what motivates workers in this sector. It would be simple to split approaches into two categories: those where workers care about personal outcomes like pay, comfort, lifetime earnings, or slack; and those where workers care about broader themes such as policy outcomes, public welfare, or “doing the public’s work.” It is likely that if this answer has not yet been discovered, that even if it is discovered soon, that there still will be uncertainty about whether that answer applies to future workers. For the purposes of this essay, answering that question is not particularly important. Imagine a worker that competes to improve their human capital so that they might win the contest of better external options in the private sector. As long as those improvements are consistent with views in the organization about “what matters most,” then winning that lottery is incentive-compatible. Similarly, a worker who wants to improve the chances that their policy views are represented at the highest levels might strive to win the tournament with an eye toward promotions that increase their leverage in the organization. Such outcomes, as long as the worker does what is asked by the organization in the ordinal rankings, are also incentive-compatible. In this sense tournaments are simply mechanisms for rewarding behavior that is consistent with the demands of those who design the tournament (albeit with unique features described in the previous sections). This is similarly true for pay-for-performance or any other means of providing feedback from performance appraisals. If a perfect PSM detection engine is ever discovered, then adverse selection probably falls away—but even so moral hazard remains unless we can predict actions continuously based on those deep PSM attributes. In this way, tournaments are just mechanisms for offering feedback. The following discussion centers on promotion because scholars of the civil service have long recognized its power for motivation. Promotion ranks as one key extrinsic reward along with pay raises and health benefits (Crewson 1997; Rainey 1976). Historically, studies of government employees have shown that promotion is perhaps less important for public sector than private sector (Crewson 1997), or at least that government employees are less satisfied with the way in which promotions are handled (Rainey 1976). Other studies show no significant differences between the attitudes of public and private employees in terms of the perceived effectiveness of promotion systems, or in its relative ranking compared to other extrinsic and intrinsic motivators (Garbis and Simo 1995; Wittmer 1991). Of course, as in the private sector, a history of gender and race-based discrimination in public sector promotion policies taints the incentive-compatible effects of such policies and procedures (e.g., DiPrete and Soule 1988). These basic attributes of the civil service are important because they resonate with core aspects of the theory of tournaments (Lazear and Rosen 1981; Malcomson 1984). It is important to remember that the theory says that no one tournament is most efficient: many designs are efficient and acceptable for introducing incentives. For example, as Mosher (1968) noted in his classic Democracy and the Public Service, a key distinction in civil service systems is that “rank inheres in the job,” whereas in career systems “rank inheres in the person” (Mosher 1968). This fine distinction is useful but not terribly important. In a civil service system, step-level pay gradations and nontrivial promotion probabilities can create the kind of mobility implied by tournament theory (even if they are not always—or currently are—used for that purpose). In career systems, competition between all of the members within a level of an agency creates the pressure implied by tournament theory, even if there is limited exposure to outside competition. Either system would be consistent with the general theory of rank-order tournaments. Public administration scholars historically have criticized the federal promotion system, noting the lack of clarity of Office of Personnel Management guidelines, as well as the possibility of favoritism or even ossification (e.g., Baron and Cook 1992; Diprete and Soule 1988; Johnson and Sink 1986; Young et al. 1998). Directly linking pay and performance may be even more problematic since “many managers also dislike certain aspects of performance-based compensation. Though they would prefer to promote subordinates on the basis of merit, distributing pay on that basis is risky. It means making hard-to-defend distinctions among many people with whom they must continue to work” (Wilson 1989, 144). Promotion often depends on seniority even in business; it costs supervisors to advocate for position reclassification or direct promotion. Yet, these are costly decisions because winners value the prize of promotion; not all competitors win. In important ways, the American states have become laboratories for using bonuses, skill-based pay, or competency-based pay to compensate workers (e.g., Kearney 2006; Rainey 2006). Even though managers may be more likely to use bonuses and other contingent pay schemes, though, their impact may be limited because designers need cardinal measures of performance. For example, designers in Georgia found it difficult to construct even ordinal measures of performance that have face validity for both supervisors and subordinates (Nigro and Kellough 2006). Most likely such incentives will have low power because of little connection between the system’s rewards and clear, cardinal measures. As noted above, incentive systems rarely replace existing compensation systems but are usually layered on top of older systems (Grandori 1999). In South Carolina, designers wanted the Employee Performance Management System (EPMS) to develop evaluation instruments for rating employees, comparing ratings, and then rewarding high performers with pay increases (Hays et al. 2006), a system like a rank-order tournament. The Accountability Act also created other recognitions allocated through competitive tournaments, for example, plaques, meals, parking. But designers limited the incentives in the second category by limiting the rewards to less than $50 cash, so only rarely would civil servants take the incentives seriously. In both cases, the core decision for the principal is to select the number of employees to gain the rewards; the theory of tournaments tells us that the more that receive, the weaker the incentives. Other systems in Texas, Arizona, and Florida have allowed bonuses or other monetary awards but similarly limited their usefulness. One important way in which governments have changed incentives is the innovation of broadbanding. Specifically, broadbanding is the practice of consolidating salary grades into several broader pay ranges so that the range is wider but there is low overlap. Broadbanding is part of the movement to flatten organizational hierarchies and push decision-making to the “street level” (e.g., Lipsky 1980). At least 16 states have tried full or limited broadbanding, mainly to trim the number of job classifications and accomplish pay for performance (Whalen and Guy 2008). Recall that incentives based on a linear dependence of pay on performance always require strong cardinal performance measures. Production in the public sector makes it hard enough to find such measures at the individual level; they are even harder when individuals work in groups with positive production externalities (e.g., when what one teacher does determines another teacher’s success rate). Broadbanding is meant to provide financial flexibility. In piece-rate reasoning, broadbanding helps differentiate between employee’s performances (measured cardinally). In theory of tournaments, it all depends on the step size between the bands and whether managers can move employees across them. More importantly, broadbanding limits the power of promotions because there are fewer opportunities for upward movement in the hierarchy. Flat, “street-level” organizations have few promotion opportunities because there are relatively few supervisory positions given large numbers of street-level bureaucrats. Flattening organizations and enhancing the power of individual caseworkers will reduce the power of traditional, promotion-based incentives (along with inevitably increasing the supervisors’ required span of control).5 Of course, the federal government’s structure for promotions and advancement predates the theory of tournaments, although the connections are enticing. The pay grades used to move regular civil service employees forward through the system, from the General Schedule GS-1 level through GS-15, provide for movements between and within grades (through steps). From GS-1 Step 1 to GS-15 Step 10, employees face opportunities for significant increases in extrinsic rewards. This does not mean that government always uses the advancement system as the theory of tournaments suggests. As Wilson notes, “Theoretically, moving up these steps requires obtaining a good performance rating, but in practice almost everybody gets a fully satisfactory rating and so moving up the steps tends to reflect the passage of time” (Wilson 1989, 143). However, the prospects for its use for incentive-compatibility purposes remain. As Wilson notes, moving up a grade requires “getting the personnel office to approve a higher classification for that person, and that brings us back to the vague (but lengthy) language of the position classification standards” so that “managers spend much of their time memorizing verbal abstractions, honing their skills at drafting flowery memos, recommending promotions, and arguing with position classifiers” (Wilson 1989, 143). Presumably that amount of effort on the part of a manager only happens when a worker has shown herself to rank higher than other competing workers. It is fair to say that the rules that govern the civil service (especially the Federal GS system) probably make it difficult to implement the kinds of tournaments described in the theoretical literature. The transition of workers through the Steps in each GS Grade has become rigid over time. In some ways, civil service regulations protect government workers from personnel practices that are used in a range of private sector firms, although the current blends of practices documented by Ichniowski et al. (1997) are widely represented across different levels of government. Most importantly, public administration scholars have noted the “deinstitutionalization” of the civil service over the last decades, most notably at agencies like the Department of Homeland Security and the Department of Defense (DOD) (Thompson 2006). As a result of this deinstitutionalization in 2006, fewer than 20% of employees in the Executive Branch remained covered under the traditional civil service system (Schuh 2006). Essentially, the collective bargaining agreements vary across sets of federal employees; as noted above, adding in state and local employees makes the situation more complex. Moreover, there is even greater variation across the range of civil service systems seen around the world. Two general points can be made about the prospects for altering these frameworks to look more like the tournaments seen in theory and experimental analysis. First, the key question in applying tournaments is how to cause managers to make forced relative rankings. This issue reflects the fundamental difference in tournaments: the use of ordinal ranks instead of the kinds of cardinal performance measures seen in piece rates. As noted above, forced relative rankings reduce centrality and leniency bias (Landy and Farr 1980; Murphy and Cleveland 1991). They cause managers to stretch the distribution of grades away from a single, modal answer for each employee. They force managers to punish under-employing employees by giving them poor marks. We have little empirical knowledge of how individual managers transmit implicit forced relative ranks into the claims they make for employees who are seeking rewards. In some ways, this type of research is the “dirty work” for those studying public bureaucracies because it requires visiting and understanding how managers employ the tools of sanction and reward. Of course, this type of work is done in some areas of organizational behavior research, but much of that area remains underrepresented in public administration. The second point is that promotion is not the only reward managers wield when deciding how and when to reward employees, so just looking at the GS system pulls research attention away from a vast array of other mechanisms in place for someone wanting to implement a competitive tournament. Non-promotion oriented tournaments such as the use of other bonuses and rewards are normal and allowed even in the traditional civil service. Historically, the US Office of Personnel Management (OPM) offered specific rules that allow for a range of financial and other motivational awards. Incentive awards have been described in Chapter 451 of Title 5 in the Code of Federal Regulations. There have been four general types of awards: lump-sum cash awards, honorary awards, informal recognition awards, and time off awards. Most agencies have been able to offer cash awards up to $10,000 without OPM approval and the president must approve amounts over $25,000; interestingly, the DOD and the Internal Revenue Service have offered awards up to $25,000 without OPM approval. Agencies have been able to use honorary and informal recognition awards at will. Time-off rewards have been normal and governed by relatively few regulations. OPM rules have even allowed for “referral bonuses” as incentive awards for recruitment of new employees, essentially allowing for headhunting (OPM 2010). One advantage of having a range of rewards outside traditional promotion systems is that it allows for greater treatment of the heterogeneity problem. Managers need homogenous pools in order to separate the sorting and incentivization attributes of tournaments. Homogeneity is achieved, to a degree, by testing requirements built into most of the different types of civil service systems. With rewards beyond promotion, managers can carve out subpopulations of employees for consideration of special treatment. They could employ within-cohort competitions like those proposed by Bhattacharya and Guasch, or even pairwise competitions. The general point is that there is now substantial flexibility within the federal civil service because of the range of systems in place, the possibility for relative ranking, and the availability of alternative reward systems outside traditional promotion. Promotion remains a potent tool for implementing tournaments, but it would exist as part of the blend of tools now present in all modern organizations. Implications The primary claim here is that public sector reforms have emphasized pay-for-performance systems that are based in the theory of piece rates over approaches that take advantage of the logic of tournaments. In this section I offer implications drawn from the tournaments approach to incentives for our understanding of government practices. First, rank-order tournaments are powerful because they rely on ordinal rankings instead of cardinal performance measurements. At a minimum, then, you can use tournaments when data on ordinal ranks are available and cardinal data are not. Of course, every cardinal scale contains an ordinal one (Stevens 1968), so cardinal data collection is useful for building a rank-order promotion tournament. Employees in firms have long known how to game piece rates, in part because those systems emphasize quantity over quality measures. As in other settings, tournament designers must decide the “fitness criterion”—the content of the scale on which everyone is graded. A good example of the importance of this choice from the political science literature is the choice of the fitness criterion in Axelrod’s classic The Evolution of Cooperation, which is itself a kind of tournament (e.g., Axelrod 1984). Is the fitness criterion in government efficiency, representativeness, enforcement, or some other measure? Should it be one, all, or a combination? If combined, what “social weights” should we use? Of course, all pay-for-performance systems suffer from this constraint as well, but tournaments place fewer constraints on how we use this complex combination of a variety of individual measures—they must only scale. Second, past reform efforts in government show the influence of the institutions that govern promotion decisions. As Mosher notes, the civil and career services, and now the Senior Executive Service (SES), track workers into various promotion tournaments, each with unique benchmarking opportunities for determining competition. Kaufman’s classic The Forest Ranger reflects this concern. In that study, promotion systems are both a specific organizational choice—intended in part to induce compliance by widely-dispersed rangers—and a ladder describing the position of individual rangers in their career life cycle. Traditional business distinctions between “functional” and “product” lines determine how downward externalities spread in hierarchical firms. Tournament theory is probably inadequate for “dealing with the richness and complexity of hierarchical organizations and to the importance of the remaining tasks in integrating the theories of incentive schemes and hierarchies” (Bhattacharya and Guasch 1988, 880). Yet, the variety of systems in place shows the need to consider tournaments when examining bureaucracies; a simple approach driven by piece rates is not flexible. The institutional structure of promotion is rarely recognized in current debates about public agencies. A side benefit of the theory of tournaments is that it helps us reinterpret past studies of public agencies. For example, Eisner’s study of change at the Department of Justice Antitrust Division showed the effects of the intrusion of economists into a traditionally lawyer-dominated workplace (Eisner 1991). Heterogeneous worker pools—with structured interaction through promotion tournaments—shifted both individual efforts and organizational outcomes. Eisner’s careful work gives us an empirical lens on these outcomes, whereas tournament theory shows a workplace-level mechanism with formal theoretical legs. Third, even though, as Mosher points out, the career and civil service are two core attributes of public organizations, political appointees play an important, although not always positive, continuing role in politics and the public sector (e.g., Lewis 2008). Tournaments offer three lenses on appointments. When choosing appointees, presidents cull lists of prospective appointees for those that best match explicit and implicit criteria in an implicit tournament. Although in power, appointees balance presidential priorities and organizational maintenance when they lead agencies; they do so with a view to their fitness in post-appointment competitive markets, which may be based on reputation effects or on external options. Finally, appointees in power vie with one another in implicit tournaments for presidential favor. Some are advantaged through Cabinet status; in tournaments, individuals who are disadvantaged may in fact give greater effort (Schotter and Weigelt 1992). The allotment of ministerial positions in parliamentary systems is a unique type of rank-order tournament. Fourth, through the lens of tournaments, the American “representative bureaucracy” movement looks almost efficient (e.g., Meier 1975). Experimentalists use tournaments to test whether affirmative action decreases organizational efficiency through the mechanism of promotion (Schotter and Weigelt 1992). They find the opposite: representation actually increases the effort levels of those previously advantaged; the total organization is better off. Specifically, this article shows that “unfair tournaments,” those in which some participants are “advantaged” produce higher-than-theorized effort levels for all participants. “There is no tradeoff between equity and efficiency” in these cases: “firms should prohibit discrimination in the workplace because it is in their best interest” by improving overall performance (Schotter and Weigelt 1992, 538). Fifth, a typical attribute of modern bureaucracies, smaller pay differentials at higher hierarchical levels, may make promotion-seeking employees less competitive. Salaries in firms have high skew at higher levels of a pyramidal hierarchy (Rosen 1986). This gives an additional backbone to expanding remuneration schemes in the SES; it also fits with how other countries pay higher-level civil service employees “market wages” (e.g., Singapore) (Van Rijckeghem and Weder 2001). Naturally, bonuses can be budget-breaking (Miller and Whitford 2002). But fewer workers must be rewarded to align incentives in the hierarchy in tournaments than in pay-for-performance systems. Note also that the credibility of political choices to pay bureaucrats such piece-rate bonuses remains an open question; decisions to not pay agreed-on bonuses bring into question the credibility of the labor contract for payment of members of the SES (Newell 2010). Sixth, the whole reform effort to flatten the federal bureaucracy and expand the span of control of supervisors can actually make organizations less efficient (Theobald and Nicholson-Crotty 2005). Subordinates competing in promotion tournaments in flatter organizations with fewer supervisors have fewer individual opportunities to advance; effort levels may fall in response and so may organizational performance. Few participants should participate in many types of tournaments (Fullerton and McAfee 1999). Having a broad level within a hierarchy makes it difficult to motivate workers in rank-order tournaments because of the low promotion probability (Brown 1998). Firms that face this problem (e.g., one with a large number of retail outlets and a small regional supervision staff) often franchise outlets to motivate managers by giving them a residual claim to local profits. Likewise, national governments delegate some federal policies to state-level agencies instead of federal field offices. Having many federal field offices reduces promotion probabilities and makes it harder to motivate. This is especially true given the difficulty (or impossibility) of pay-for-performance compensation systems. Over time, agencies like the Department of Agriculture have found it important to reduce the number of federal field offices. Last, perhaps the hardest problem that flows from civil service arrangements as they are experienced in the United States is the “incentives for losers” problem. What if a class of workers develops that cannot get promoted? What if pay-for-performance offers little respite in terms of motivating behavior? What if employee-of-the-month awards are not sufficient? If these workers cannot be selected out via a PSM-flavored pre-test, what options are left? “Cannot be fired” is a substantial institutional limit on tournaments because they can create that class of losers who are not motivated by recurring tournaments. Some might argue that the answer lies in “exit tournaments.” Imagine a setting in which reductions-in-force are applied based on performance rather than seniority or other protections. Private sector experience supports the idea that exit tournaments can also be incentive-compatible, even at low probabilities. At a minimum, probationary periods exist and provide opportunities for early termination (Barrett and Greene 2015). Moreover, termination does happen, although perhaps at lower levels than in the private sector (Yoder 2015). Although it is clear that termination could be abused, further elaboration of the prospects of such mechanisms would be useful. Discussion This essay integrates a body of economic theory on rank-order tournaments as compensation systems with the traditional public administration concern for compliance and accountability in government. Although the returns to delegation are well-known, studies have largely ignored the incentive-compatibility aspects of traditional personnel systems, in part because of a preoccupation with specific mechanism designs that flow from piece-rate approaches to employee compensation. Calls for reform have embraced the economic literature on principal-agency without considering subsequent research on the motivational power of rank-order tournaments in hierarchy. The central contribution of this essay is to advance the claim that promotion tournaments in public organization hierarchies can act as incentive-compatibility mechanisms; indeed, such mechanisms might perhaps be more efficient than pay-for-performance systems rooted in the theory of piece rates. This observation is bolstered by empirical support showing the power of tournaments in experiments and firms. It is also bolstered by the prospects for promotion- and reward-based tournaments in numerous government settings. Along with this normative proposition, I make the claim that the lens of tournament theory offers significant opportunities for empirical investigation in public management. If tournaments appear at least theoretically efficient, what explains their use in some settings and not others? What explains the use of one form over another? Clearly, these are the realms of organizational design and organizational behavior. Institutional form in government, just as in business, is a paramount choice. Mechanism design centers on tournament structures, the integration of measurement systems and rankings, and even the collective choices organizations make in personnel decisions like promotion and other awards. The effects of tournaments on institutional form may even extend to the relative decentralization of the nation-state. If anything, future research should recognize and investigate a multiplicity of efficient mechanism designs beyond traditional principal-agency structures. References Ammons, David. 1992. Productivity barriers in the public sector. In Public productivity handbook , ed. Marc Holzer. New York, NY: Marcel Dekker. Atkinson, Anthony B., and Jospeh E. Stiglitz. 1980. Lectures in public economics . New York, NY: McGraw-Hill. Axelrod, Robert. 1984. The evolution of cooperation . New York, NY: Basic Books. Baker, George P., Michael C. Jensen, and K. J. Murphy. 1988. Compensation and incentives: Practice vs. theory. Journal of Finance 43: 593– 616. Google Scholar CrossRef Search ADS Baker, George P., Michael Gibbs, and Bengt Holmström. 1994. The wage policy of a firm. Quarterly Journal of Economics 109: 921– 55. Google Scholar CrossRef Search ADS Barber, Brad M., and Terrance Odean. 2001. Boys will be boys: Gender, overconfidence, and common stock investment. Quarterly Journal of Economics 116: 261– 92. Google Scholar CrossRef Search ADS Barnard, Chester I. 1938. The functions of the executive . Cambridge, MA: Harvard Univ. Press. Baron, James N., and Karen S. Cook, ed. 1992. Process and outcome: Perspective on distribution of rewards in organizations. Special Issue. Administrative Science Quarterly 37: 198– 361. Google Scholar CrossRef Search ADS Barrett, Katherine, and Richard Greene. 2015. ‘You’re Fired’: Ways to get rid of bad government workers. Governing . September 7. http://www.governing.com/columns/smart-mgmt/gov-firing-public-employees.html (accessed January 10, 2018). Behn, Robert D. 2003. Why measure performance? Different purposes require different measures. Public Administration Review 63: 586– 606. Google Scholar CrossRef Search ADS Besley, Timothy. 2006. Principled agents? Oxford, England: Oxford Univ. Press. Bhattacharya, Sudipto, and J. Luis Guasch. 1988. Heterogeneity, tournaments, and hierarchies. Journal of Political Economy 96: 867– 81. Google Scholar CrossRef Search ADS Bognanno, Michael L. 2001. Corporate tournaments. Journal of Labor Economics 19: 290– 315. Google Scholar CrossRef Search ADS Bottom, William P., Jim Holloway, Gary J. Miller, Alexandra Mislin, and Andrew B. Whitford. 2006. Building a pathway to cooperation: Negotiation and social exchange between principal and agent. Administrative Science Quarterly 51: 29– 58. Google Scholar CrossRef Search ADS Brehm, John and Scott Gates. 1999. Working, shirking, and sabotage . Ann Arbor, MI: Univ. of Michigan Press. Brown, K.C., W.V. Harlow, and L.T. Starks. 1996. Of tournaments and temptations: An analysis of managerial incentives in the Mutual Fund Industry. Journal of Finance 51: 85– 110. Google Scholar CrossRef Search ADS Brown, William O. 1998. Transaction costs, corporate hierarchies, and the theory of franchising. Journal of Economic Behavior and Organizations 36: 319– 29. Google Scholar CrossRef Search ADS Bull, Clive, Andrew Schotter, and Keith Weigelt. 1987. Tournaments and piece rates: An experimental study. Journal of Political Economy 95: 1– 33. Google Scholar CrossRef Search ADS Carmichael, H. Lorne. 1983. The agent-agents problem: Payment by relative output. Journal of Labor Economics 1: 50– 65. Google Scholar CrossRef Search ADS Chase, Ivan D. 1991. Vacancy chains. Annual Review of Sociology 17: 133– 54. Google Scholar CrossRef Search ADS Chevalier, J., and G. Ellison. 1997. Risk taking by mutual funds as a response to incentives. Journal of Political Economy 105: 1167– 200. Google Scholar CrossRef Search ADS Choi, Sungjoo, and Andrew B. Whitford. 2011. A topography of civil service laws. International Public Management Journal 14: 106– 30. Google Scholar CrossRef Search ADS ———. 2017. Employee satisfaction in agencies with merit-based pay: Differential effects for three measures. International Public Management Journal 20: 442– 66. CrossRef Search ADS Coate, Douglas, and Donijo Robbins. 2001. The tournament careers of top-ranked men and women tennis professionals: Are the gentlemen more committed than the ladies? Journal of Labor Research 22: 185– 93. Google Scholar CrossRef Search ADS Conlon, E. J., and J. M. Parks. 1990. Effects of monitoring and tradition on compensation arrangements: An experiment with principal-agent dyads. Academy of Management Journal 33: 603– 22. Google Scholar CrossRef Search ADS Conyon, Martin J., Simon I. Peck, and Graham V. Sadler. 2001. Corporate tournaments and executive compensation: Evidence from the UK. Strategic Management Journal 22: 805– 15. Google Scholar CrossRef Search ADS Coughlan, A. T., and C. Narasimhan. 1992. An empirical analysis of sales-force compensation plans. Journal of Business 65: 93– 121. Google Scholar CrossRef Search ADS Coupé, Tom, Valérie Smeets, and Frédéric Warzynski. 2006. Incentives, sorting and productivity along the career: Evidence from a sample of top economists. Journal of Law, Economics, and Organization 22: 137– 67. Google Scholar CrossRef Search ADS Crewson, Philip E. 1997. Public-service motivation: Building empirical evidence of incidence and effect. Journal of Public Administration Research and Theory 7: 499– 518. Google Scholar CrossRef Search ADS De Varo, Jed. 2006. Internal promotion competitions in firms. Rand Journal of Economics 37: 521– 42. Google Scholar CrossRef Search ADS DiIulio, John D. Jr. 1994. Principled agents: The cultural bases of behavior in a federal government bureaucracy. Journal of Public Administration Research and Theory 4: 277– 318. DiPrete, Thomas A., and Whitman T. Soule. 1988. Gender and promotion in segmented job ladder systems. American Sociological Review 53: 26– 40. Google Scholar CrossRef Search ADS Dixit, Avinash K. 2002. Incentives and organizations in the public sector: An interpretative review. Journal of Human Resources 37: 696– 727. Google Scholar CrossRef Search ADS Drago, Robert, and John S. Heywood. 1989. Tournaments, piece rates, and the shape of the payoff function. Journal of Political Economy 97: 992– 8. Google Scholar CrossRef Search ADS ——— 1991. Subject response to loss in an experimental tournament. Economics Letters 36: 371– 4. CrossRef Search ADS Drago, Robert, Gerald T. Garvey, and Geoffrey K. Turnbull. 1996. A collective tournament. Economics Letters 50: 223– 7. Google Scholar CrossRef Search ADS Drucker, Peter F. 1954. The practice of management . New York, NY: Harper and Row. Ehrenberg, Ronald G., and Michael L. Bognanno. 1990a. Do tournaments have incentive effects? Journal of Political Economy 98: 1307– 24. Google Scholar CrossRef Search ADS ——— 1990b. The incentive effects of tournaments revisited: Evidence from the European PGA tour. Industrial and Labor Relations Review 43: 74S– 88S. CrossRef Search ADS Eisenhardt, K. M. 1989. Agency theory: An assessment and review. Academy of Management Review 14: 57– 74. Eisner, Marc Allen. 1991. Antitrust and the triumph of economics . Chapel Hill, NC: Univ. of North Carolina Press. Eriksson, Tor. 1999. Executive compensation and tournament theory: Empirical tests on Danish data. Journal of Labor Economics 17: 262– 80. Google Scholar CrossRef Search ADS Eriksson, Tor, Sabrina Teyssier, and Marie-Claire Villeval. 2006. Self-selection and the efficiency of tournaments. IZA Discussion Paper No. 1983. Fairburn, James, and James Malcomson. 2001. Performance, promotion, and the Peter principle. Review of Economic Studies 68: 45– 66. Google Scholar CrossRef Search ADS Fehr, Ernst, Simon Gächter, and G. Kirchsteiger. 1997. Reciprocity as a contract enforcement device. Econometrica 65: 833– 60. Google Scholar CrossRef Search ADS Fehr, Ernst, and Urs Fischbacher. 2003. The nature of human altruism. Nature . 425: 785– 91. Google Scholar CrossRef Search ADS PubMed Fesler, James W., and Donald F. Kettl. 1996. The politics of the administrative process , 2nd ed. Chatham, NJ: Chatham House Publishers. Friedman, Daniel, and Shyam Sunder. 1994. Experimental methods: A primer for economists . Cambridge, England: Cambridge Univ. Press. Google Scholar CrossRef Search ADS Fullerton, Richard L., and R. Preston McAfee. 1999. Auctioning entry into tournaments. Journal of Political Economy 107: 573– 605. Google Scholar CrossRef Search ADS Galanter, Marc, and Thomas Palay. 1991. Tournament of lawyers: The transformation of the big law firm . Chicago, IL: Univ. of Chicago Press. Grandori, Anna. 1999. Organizations and economic behavior . London, UK: Routledge. Green, Jerry R., and Nancy L. Stokey. 1983. A comparison of tournaments and contracts. Journal of Political Economy 91: 349– 64. Google Scholar CrossRef Search ADS Grossman, Sanford J., and Oliver D. Hart. 1983. An analysis of the principal-agent problem. Econometrica 51: 7– 45. Google Scholar CrossRef Search ADS Gunderson, Morley. 2001. Economics of personnel and human resource management. Human Resource Management Review 11: 431– 53. Google Scholar CrossRef Search ADS Hall, Brian J., and Jeffrey B. Liebman. 1998. Are CEOS really paid like bureaucrats? Quarterly Journal of Economics 113: 653– 91. Google Scholar CrossRef Search ADS Hambrick, D.C., and S. Finkelstein. 1995. The effects of ownership structure on conditions at the top: The case of CEO pay raises. Strategic Management Journal 16: 175– 93. Google Scholar CrossRef Search ADS Hays, Steven W., Chris Byrd, and Samuel L. Wilkins. 2006. South Carolina’s human resource management system: The model for states with decentralized personnel structures. In Civil service reform in the states , ed. J. Edward Kellough and Lloyd G. Nigro, 171–202. Albany, NY: State Univ. of New York Press. Holmström, Bengt. 1982. Moral hazard in teams. Bell Journal of Economics 13: 324– 40. Google Scholar CrossRef Search ADS Ichniowski, Casey, Kathryn Shaw, Giovanna Prennushi. 1997. The effects of human resource management practices on productivity. American Economic Review 86: 291– 313. Izzo, Christina. 2013. Despite salary cap, school districts can use state-authorized merit bonuses to try and retain superintendents. The Times of Trenton , August 11. http://www.nj.com/mercer/index.ssf/2013/08/despite_superintendent_salary_cap_districts_can_use_state authorized_merit_bonuses_to_try_and_make_u.html (accessed December 17, 2017). Johnson, V. E., and D. Sink. 1986. Personnel reform in consolidated metropolitan governments: Executive responsibility and neutral competence. Public Personnel Management 15: 11– 21. Google Scholar CrossRef Search ADS Kantner, Rosebeth Moss. 1993. Men and women of the corporation , 2nd edition. New York, NY: Basic. Kaplan, Robert S., and David P. Norton. 1992. The balanced scorecard - measures that drive performance. Harvard Business Review 70: 71– 91. Google Scholar PubMed Kearney, Richard C. 2006. The labor perspective on civil service reform in the states. In Civil service reform in the states , ed. J. Edward Kellough and Lloyd G. Nigro, 77–94. Albany, NY: State Univ. of New York Press. Keeney, Ralph L., and Howard Raiffa. 1993. Decisions with multiple objectives: Preferences and value tradeoffs . New York, NY: Cambridge Univ. Press. Google Scholar CrossRef Search ADS Kellough, J. Edward, Lloyd G. Nigro, and Gene A. Brewer. 2010. Civil service reform under George W. Bush: Ideology, politics, and public personnel administration. Review of Public Personnel Administration 30: 404– 22. Google Scholar CrossRef Search ADS Kernell, Samuel, and Michael P. McDonald. 1999. Congress and America’s Political Development: The transformation of the post office from patronage to service. American Journal of Political Science 43: 792– 811. Google Scholar CrossRef Search ADS Kjeldsen, Anne Mette, and Lotte Boghe Andersen. 2013. How pro-social motivation affects job satisfaction: An international analysis of countries with different welfare state regimes. Scandinavian Political Studies 36: 153– 76. Google Scholar CrossRef Search ADS Knott, Jack H., and Gary J. Miller. 1987. Reforming bureaucracy: The politics of institutional choice . Englewood Cliffs, NJ: Prentice-Hall. Kravchuk, Robert S., and Ronald W. Schack. 1996. Designing effective performance measurement systems under the government performance and results act of 1993. Public Administration Review 56: 348– 58. Google Scholar CrossRef Search ADS Laffont, J., and D. Martimort. 2002. The theory of incentives . Princeton, NJ: Princeton Univ. Press. Landy, Frank J., and James L. Farr. 1980. Performance rating. Psychological Bulletin 87: 72– 107. Google Scholar CrossRef Search ADS Lazear, Edward P. 1989. Pay equality and industrial politics. Journal of Political Economy 97: 561– 80. Google Scholar CrossRef Search ADS ——— 2003. Teacher incentives. Swedish Economic Policy Review 10: 179– 214. Lazear, Edward P., and Sherwin Rosen. 1981. Rank-order tournaments as optimum labor contracts. Journal of Political Economy 89: 841– 64. Google Scholar CrossRef Search ADS Levitt, Steven D., and Sudhir Alladi Venkatesh. 2000. An economic analysis of a drug-selling gang’s finances. Quarterly Journal of Economics 115: 755– 89. Google Scholar CrossRef Search ADS Lewis, David E. 2008. The politics of presidential appointments: Political control and bureaucratic performance . Princeton, NJ: Princeton Univ. Press. Google Scholar CrossRef Search ADS Lipsky, Michael. 1980. Street-level bureaucracy . New York, NY: Russell Sage Foundation. Main, Brian G.M., Charles A. O’ReillyIII, and James Wade. 1993. Top executive pay: Tournament or teamwork? Journal of Labor Economics 11: 606– 28. Google Scholar CrossRef Search ADS Malcomson, James M. 1984. Work incentives, hierarchy, and internal labor markets. Journal of Political Economy 92: 486– 507. Google Scholar CrossRef Search ADS March, James, and Herbert A. Simon. 1958. Organizations . New York, NY: Wiley. McClelland, David C. 1966. That urge to achieve. Think : 82– 9. McClure, James E., and Lee C. Spector. 1997. Tournament performance and ‘Agency’ problems: An empirical investigation of ‘March Madness.’ Journal of Economics and Finance 21: 61– 8. Google Scholar CrossRef Search ADS McLean Parks, J., and E. J. Conlon. 1995. Compensation contracts: Do agency theory assumptions predict negotiated agreements? Academy of Management Journal . 38: 821– 38. Google Scholar CrossRef Search ADS Meier, Kenneth J. 1975. Representative bureaucracy: An empirical analysis. American Political Science Review 69: 526– 42. Google Scholar CrossRef Search ADS Meier, Kenneth J., and Larry J. O’Toole, Jr. 2002. Public management and organizational performance: The effect of managerial quality. Journal of Policy Analysis and Management 21: 629– 43. Google Scholar CrossRef Search ADS Mellon, Ericka. 2014. HISD trustees again give superintendent a bonus. Houston Chronicle , December 11. http://www.houstonchronicle.com/news/article/HISD Supt Grier s bonus 116 600 5951875.php (accessed December 17, 2017). ——— 2016. New Houston ISD superintendent richard carranza getting $345,000 annual salary in 3-year cONTRACT. Houston Chronicle , August 18. http://www.chron.com/news/education/article/HISD officially hires new superintendent from San 9171272.php (accessed December 17, 2017). Melton, Michael, and Thomas S. Zorn. 2000. An empirical test of tournament theory: The senior PGA tour. Managerial Finance 26: 16– 32. Google Scholar CrossRef Search ADS Miller, Gary J. 1992. Managerial dilemmas . New York, NY: Cambridge Univ. Press. Google Scholar CrossRef Search ADS ——— 2000. Above politics: Credible commitment and efficiency in the design of public agencies. Journal of Public Administration Research and Theory 10: 289– 328. CrossRef Search ADS Miller, Gary J., and Andrew B. Whitford. 2002. Trust and incentives in principal-agent negotiations: The ‘Insurance/Incentive Trade-off’. Journal of Theoretical Politics 14: 231– 67. Google Scholar CrossRef Search ADS ——— 2007. The principal’s moral hazard: Constraints on the use of incentives in hierarchy. Journal of Public Administration Research and Theory 17: 213– 33. ——— 2016. Above politics: Bureaucratic discretion and credible commitment . New York, NY: Cambridge Univ. Press. Mirrlees, James A. 1976. The optimal structure of incentives and authority within an organization. Bell Journal of Economics 7: 105– 31. Google Scholar CrossRef Search ADS Moe, Terry M. 1984. The new economics of organization. American Journal of Political Science 28: 739– 77. Google Scholar CrossRef Search ADS Mosher, Frederick C. 1968. Democracy and the public service . New York, NY: Oxford Univ. Press. Moy, Ronald L., and Thomas Liaw. 1998. Determinants of professional golf tournament earnings. American Economist 42: 65– 70. Google Scholar CrossRef Search ADS Murphy, Kevin R., and Jeanette Cleveland. 1991. Performance appraisal: An organizational perspective . Boston, MA: Allyn and Bacon. Nalbantian, Haig R., and Andrew Schotter. 1997. Productivity under group incentives: An experimental study. American Economic Review 87: 314– 41. Nalebuff, Barry J., and Joseph E. Stiglitz. 1983. Prizes and incentives: Towards a general theory of compensation and competition. Bell Journal of Economics 14: 21– 43. Google Scholar CrossRef Search ADS National Academy of Public Administration. 2004. Recommending performance-based federal pay . Washington, DC: National Academy of Public Administration. Newell, Elizabeth. 2010. Freeze on bonuses could have trickledown effect. GovExec.com. http://www.govexec.com/story_page.cfm?articleid= 45842&sid=2 (accessed December 17, 2017). Niederle, Muriel, and Lise Vesterlund. 2007. Do women shy away from competition? Do men compete too much? Quarterly Journal of Economics 122: 1067– 101. Google Scholar CrossRef Search ADS Nigro, Lloyd G., and J. Edward Kellough. 2006. Civil service reform in Georgia: A view from the trenches. In Civil service reform in the states , ed. J. Edward Kellough and Lloyd G. Nigro, 117–143. Albany, NY: State Univ. of New York Press. Google Scholar CrossRef Search ADS Oberg, Ted, and Trent Seibert. 2016. New HISD superintendent likely to get super salary. ABC13.com, August 17. http://abc13.com/news/new hisd superintendent likely to get super salary/1473742/ (accessed December 17, 2017). Office of Personnel Management. 2010. Performance management: General awards. http://www.opm.gov/perform/faqs/genawrds.asp (accessed September 15, 2014). O’Keefe, Mary, W. Kip Viscusi, and Richard J. Zeckhauser. 1984. Economic contests: Comparative reward schemes. Journal of Labor Economics 2: 27– 56. Google Scholar CrossRef Search ADS Partnership for Public Service. 2005. Pay for performance. Issue Brief PPS-05-01 . Washington, DC: Partnership for Public Service. Passell, Peter. 1996. Earning it; Paid by the widget, and proud. New York Times , June 16. P. A1. Perry, James L., and Lois R. Wise. 1990. The motivational bases of public service. Public Administration Review 50: 367– 73. Google Scholar CrossRef Search ADS Perry, James L., Debra Mesch, and Laurie Paarlberg. 2006. Motivating employees in a new governance era: The performance paradigm revisited. Public Administration Review 66: 505– 14. Google Scholar CrossRef Search ADS Perry, James L., Trent Engbers, and So Yun Jun. 2009. Back to the future? Performance-related pay, empirical research, and the perils of persistence. Public Administration Review 69 ( 1): 1– 31. Google Scholar CrossRef Search ADS Prendergast, Canice. 1998. What happens within firms. In Labor Statistics Measurement Issues , ed. John Haltiwanger, Robert Topel and Marilyn Manser, 329–56. Chicago: University of Chicago Press. Prendergast, Canice. 1999. The provision of incentives in firms. Journal of Economic Literature 38: 7– 63. Google Scholar CrossRef Search ADS Qian, Yingyi. 1994. Incentives and loss of control in an optimal hierarchy. Review of Economic Studies 61: 527– 44. Google Scholar CrossRef Search ADS Rainey, Hal G. 1976. Perceptions of incentives in business and government: Implications for civil service reform. Public Administration Review 39: 440– 8. Google Scholar CrossRef Search ADS ——— 1982. Reward preferences among public and private managers: In search of the service ethic. American Review of Public Administration 16: 288– 302. CrossRef Search ADS ——— 2006. Reform trends at the federal level with implications for the states: The pursuit of flexibility and the Human Capital Movement. In Civil service reform in the states , ed. J. Edward Kellough and Lloyd G. Nigro, 33–58. Albany, NY: State Univ. of New York Press. Risher, Howard, and Charles Fay. 1997. New strategies for public pay . Jossey-Bass. Risher, Howard. 2004. Pay for performance: A guide for federal managers . Washington, DC: IBM Center for the Business of Government. Romzek, Barbara. 1990. Employee investment and commitment. Public Administration Review 50: 374– 82. Google Scholar CrossRef Search ADS Rosen, Sherwin. 1982. Authority, control, and the distribution of earnings. Bell Journal of Economics 13: 311– 23. Google Scholar CrossRef Search ADS ——— 1986. Prizes and incentives in elimination tournaments. American Economic Review 76: 701– 15. Rosenbaum, James E. 1979. Tournament mobility: Career patterns in a corporation. Administrative Science Quarterly 24: 220– 41. Google Scholar CrossRef Search ADS ——— 1984. Career mobility in a corporate hierarchy . New York, NY: Academic. Sappington, D. 1991. Incentives in principal agent relationships. Journal of Economic Perspectives 3: 45– 66. Google Scholar CrossRef Search ADS Sattinger, Michael. 1993. Assignment models and the distribution of earnings. Journal of Economic Literature 31: 831– 80. Schott, Carina, and Adrian Ritz. 2017. The dark sides of public service motivation: A multi-level theoretical framework . Perspectives on Public Management and Governance. https://doi.org/10.1093/ppmgov/gvx011 (accessed December 17, 2017). Schotter, Andrew, and Keith Weigelt. 1992. Asymmetric tournaments, equal opportunity laws, and affirmative action: Some experimental results. Quarterly Journal of Economics 107: 511– 39. Google Scholar CrossRef Search ADS Schuh, Anna Marie. 2006. Institutional values: The Foundation for Civil Service Change. Public Personnel Management 35: 49– 69. Google Scholar CrossRef Search ADS Shearer, Bruce. 2004. Piece rates, fixed wages and incentives: Evidence from a field experiment. Review of Economic Studies 71: 513– 34. Google Scholar CrossRef Search ADS Simon, Herbert A. 1945. Administrative behavior . New York, NY: Free Press. Stevens, S.S. 1968. Measurement, statistics, and the schemapiric view. Science 161: 849– 56. Google Scholar CrossRef Search ADS PubMed Stiglitz, Joseph E. 1975. Incentives, risk, and information: Notes toward a theory of hierarchy. Bell Journal of Economics and Management Science 6: 552– 79. Google Scholar CrossRef Search ADS Stone, R.C. 1952. Mobility factors as they affect workers’ attitudes and conduct toward incentive systems. American Sociological Review 18: 28– 35. Google Scholar CrossRef Search ADS Taylor, Curtis R. 1995. Digging for golden carrots: An analysis of research tournaments. American Economic Review 85: 872– 90. Theobald, Nick A., and Sean Nicholson-Crotty. 2005. The many faces of span of control. Administration & Society 36: 648– 60. Google Scholar CrossRef Search ADS Thompson, James R. 2006. The Federal Civil Service: The demise of an institution. Public Administration Review 66: 496– 503. Google Scholar CrossRef Search ADS Tong, Kwok-kit, and Kwok Leung. 2002. Tournament as a motivational strategy: Extension to dynamic situations with uncertain duration. Journal of Economic Psychology 23: 399– 420. Google Scholar CrossRef Search ADS Tosi, H.L., J. P. Katz, and L. R. Gomez-Mejia. 1997. Disaggregating the agency contract: The effects of monitoring, incentive alignment, and term in office on agent decision making. Academy of Management Journal 40: 584– 602. Google Scholar CrossRef Search ADS Tosi, H.L., S. Werner, J. P. Katz, and L. R. Gomez-Mejia. 2000. How much does performance matter? A meta-analysis of CEO pay studies. Journal of Management 26: 301– 39. Google Scholar CrossRef Search ADS Van Rijckeghem, Caroline, and Beatrice Weder. 2001. Bureaucratic corruption and the rate of temptation: Do wages in the civil service affect corruption, and by how much? Journal of Development Economics 65: 307– 31. Google Scholar CrossRef Search ADS Waldman, Michael. 1984. Worker allocation, hierarchies, and wage distribution. Review of Economic Studies 51: 95– 109. Google Scholar CrossRef Search ADS Whalen, Cortney, and Mary E. Guy. 2008. Broadbanding trends in the states. Review of Public Personnel Administration 28: 349– 66. Google Scholar CrossRef Search ADS Whitford, Andrew B., and Vasil Jaiani. 2011. Policy windows, public opinion, and policy ideas: The evolution of no child left behind. Quality Assurance in Education 19: 8– 27. Google Scholar CrossRef Search ADS Williamson, Oliver E., Michael L. Wachter, and Jeffrey E. Harris. 1975. Understanding the employment relation: The analysis of idiosyncratic exchange. Bell Journal of Economics 6: 250– 78. Google Scholar CrossRef Search ADS Wilson, James Q. 1989. Bureaucracy: What government agencies do and why they do it . New York, NY: Basic. Wittmer, Dennis. 1991. Serving die people or serving for pay: Reward preferences among government, hybrid sector, and business managers. Public Productivity and Management Review 14: 369– 83. Google Scholar CrossRef Search ADS Xu, Peng. 1997. Executive salaries as tournament prizes and executive bonuses as managerial incentives in Japan. Journal of the Japanese and International Economies 11: 319– 46. Google Scholar CrossRef Search ADS Yoder, Eric. 2015. You can’t fire a federal employee? Don’t tell that to the agency that hears their appeals. Washington Post , May 11. https://www.washingtonpost.com/news/federal-eye/wp/2015/05/11/you-cant-fire-a-federal-employee-dont-tell-that-to-the-agency-that-hears-their-appeals (accessed December 17, 2017). Young, Brian S., Stephen Worchel, and David J. Woehr. 1998. Organizational commitment among public service employees. Public Personnel Management . 27: 339– 48. Google Scholar CrossRef Search ADS 1 Tosi et al. (2000) did not cover the growing use of stock options for executive compensation during the 1990s, but note that options were used less after the passage of Sarbanes-Oxley. 2 These prize distributions have the characteristic skew of extreme value distributions. 3 The greater variance in effort that tournaments exhibit over piece rate settings is probably due to the experimental design since participants usually are not allowed to choose between compensation schemes, which occurs in real work settings (Eriksson et al. 2006). 4 It is useful to note that there is a broader literature on gender and risk aversion, which shows generally these effects in stock market holdings and trading behavior (e.g., see Barber and Odean 2001). 5 A useful analogue is the story of gangs that are able to recruit low-level drug dealers (whose compensation is quite low) only if there is a significant probability of future riches via promotion; such gangs have highly-skewed compensation systems (Levitt and Venkatesh 2000). © The Author(s) 2018. Published by Oxford University Press on behalf of the Public Management Research Association. All rights reserved. For permissions, please e-mail: firstname.lastname@example.org.
Perspectives on Public Management and Governance – Oxford University Press
Published: Jan 29, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera