Abstract
This article combines research on incentives with nonprofit organization theories to derive three “nonprofit characteristics” that influence the use and effectiveness of incentive mechanisms in nonprofit organizations: the lack of undistorted contractible measures for the organization’s overall performance, the relevance of identified employee motivation and the social relationships between the organization and its stakeholders. Building on research from social psychology, the article argues for a more deliberate use of implicit (i.e., not contractually defined) incentives rather than a shift toward the increased use of performance contracts. Because implicit incentives are often subtle (without the need of formal justification to others) and emergent rather than planned, managers are frequently not aware of these mechanisms, and their deliberate use creates a major challenge.
Nonprofit organizations are now increasingly applying management practices and techniques originally developed for the business sector, for example, (strategic) planning, forecasting and budgeting techniques, market analysis, and performance management (Dart, 2004). The introduction of more businesslike incentive schemes, in particular performance related pay, would seem to be one of the most significant challenges for nonprofits, because they traditionally have tended to avoid paying bonuses, even at the executive level (Steinberg, 1990a, 2010; Young, 1987). However, nonprofit organizations have some specific characteristics that give reason to their nonprofit status, and such specificities may restrict the adoption of incentive mechanisms from the business sector. Indeed, contingency-based research on organizational control suggests that incentive mechanisms which are effective in one organizational context might be ineffective or even counter-productive when applied in other contexts (Chenhall, 2003).
Organizations use performance-based compensation to align the behavior of managers and employees with the organization’s main objectives. Theoretically, performance pay works by increasing effort which, in turn, leads to increases in productivity, that is, “paying on the basis of output will induce workers to supply more output” (Lazear, 2000, p.1346). Moreover, performance pay is a signal for self-selection in the workforce. For example, more able and ambitious workers may choose to seek employment in firms with performance-based remuneration schemes, whereas workers who know that their own “quality” is relatively poor might prefer firms which offer employment contracts with fixed wages (Lazear, 2000). In addition to its impact on productivity, outcome-based performance pay can be understood as a means of communication and decentralization (i.e., delegation of decision rights to agents while controlling their behavior by rewarding desirable outcomes). The definition of relevant outcome dimensions in incentive schemes provides direction for employees and is a powerful signal of those outcomes which are considered important (Merchant, 1981). Hence, incentives serve three main purposes: (a) motivation for increased effort, (b) provision of signals for self-selection in the workforce, and (c) communication of desired work outcomes and channeling of employee attention toward these outcomes.
All organizational incentive mechanisms work by providing rewards for desired behavior or by punishing undesired behavior (Merchant, Van der Stede & Zheng, 2003). But rewards and punishments can come in many forms. In addition to financial bonuses or salary increases, they can include promotions, praise from superiors or coworkers and public recognition, feelings of self-esteem, criticism, extension or loss of autonomy, and even loss of job. When an enforceable (by a third party) contract defines the conditions under which the reward is provided, this is referred to as an explicit incentive. In contrast, implicit incentives are based not on an enforceable contract, but on trust, that is, the agent performs a task and trusts in receiving some form of material or immaterial reward for the effort expended. Implicit incentives can thus be interpreted as “relational contracts,” that is, informal agreements secured by reputation and trust. 2
With such a broad notion of incentives, it becomes clear that incentives are present even in those nonprofit organizations which do not employ any formal measurement and bonus mechanisms. When nonprofits consider introducing a more businesslike incentive structure, they are typically looking at a shift from a predominantly implicit incentive system toward an increased use of explicit incentives, particularly pay-for-performance schemes.
This article extends existing research on incentives in nonprofit organizations (Roomkin & Weisbrod, 1999; Steinberg, 2010) in several ways. First, it deploys the economic theories of nonprofit organizations to derive a set of “nonprofit characteristics” that affect the use and effectiveness of incentives. Second, it contributes to the literature on the motivational characteristics of nonprofit workers, arguing for the importance of distinguishing intrinsic from identified motivation, the latter being managed mainly by implicit incentives. Third, building on Steinberg (2010), the article elaborates on the interaction between explicit and implicit incentives and draws attention to the purposes of incentives beyond their effort-increasing effects, thus contributing to a comprehensive view of “incentive packages.”
Theoretical Foundations of Incentives
Traditionally, research on incentives has focused on explicit incentives and relied largely on agency theory (see Prendergast, 1999). Only recently, insights from behavioral sciences, primarily psychology and sociology, have been used to extend the agency theory based view (Merchant, Van der Stede, & Zheng, 2003).
The Classic Agency Model
According to Ross (1973), “an agency relationship has arisen between two (or more) parties when one, designated as the agent, acts for, on behalf of, or as representative for the other, designated the principal, in a particular domain of decision problems” (p.134). Typically, the principal and the agent are assumed to have conflicting interests, with the agent being risk averse whereas the principal is risk neutral (or at least less risk averse). The “classic form” of the agency model is as follows: the agent’s action (which cannot be directly observed by the principal) affects the principal’s utility. The outcome of the agent’s action can be observed by the principal. But it depends on both the agent’s action and some other random variable outside the agent’s control (“environmental uncertainty”). Therefore, actions cannot be fully inferred from the observed outcomes. The possibility to reduce effort without being detected gives the agent an incentive to cheat. One possible solution to this problem is for the principal to monitor the agent’s behavior. However, this can be costly, and total supervision, if even possible, would render delegation meaningless. A second solution is to offer the agent an incentive contract whereby the agent’s payment depends on the value of the outcome for the principal (Prendergast, 1999).
The second main class of agency models focuses not on incentives for effort but on incentives for revealing information to the principal. Here, the agent has access to private information that affects the principal’s payoff and might benefit from misrepresenting this information. Again, only the outcome is observable to the principal. The principal requires the agent’s information to make a maximal utility decision. For example, the principal has to decide which agent to assign to a given task or has to allocate resources to business unit managers according to their estimated benefit of resource use (here agents acquire new information after the principal and the agent enter into a contractual relationship). The principal can only make the optimal decision if all available information is reported truthfully by the agents. Hence, the principal seeks to design an incentive mechanism that encourages agents to tell the truth.
The Economic Theory of Incentives Beyond the Classic Agency Model
The shortcomings of the traditional view of incentives have already been addressed by Kerr (1975) in his article “On the Folly of Rewarding A, While Hoping for B.” According to Kerr, the two main causes of distorted incentives are (a) “fascination with an ‘objective’ criterion, (where) individuals seek to establish simple, quantifiable standards against which to measure and reward performance” and (b) “overemphasis on highly visible behaviors (when) some parts of the task are highly visible whereas others are not” (pp. 779-780).
Explicit incentives: Measurability of outcomes and multitask theory
The incentive systems discussed in traditional agency literature assume that the outcome of the agent’s task is measurable, observable, and quantifiable. The violation of this assumption can lead to distorted incentives (Gibbons, 1998). When the relevant outcomes of an agent’s task are multidimensional, that is, the agent carries out multiple activities that cannot be wholly encapsulated by an overall performance measure, the so-called multitask problem occurs. If measurement and compensation are restricted to a subset of the relevant dimensions of performance, agents typically reallocate activities toward those tasks that are measured and rewarded rather than toward other aspects of performance that might be more important but are less easily measured (Holmstrom & Milgrom, 1991). Moreover, in such a multitask setting, agents can be induced to devote time and effort to taking maximum advantage or manipulating the shape of the incentive scheme instead of devoting their time to productive activities. This may distort the performance measures and hence destroy their information content for planning purposes: when used as measures in an incentive contract (the first class of agency models described above), numbers that are important for planning may be manipulated and lose their value for planning purposes or even induce principals to make wrong decisions (the second class of agency models described above). Moreover, when rewards are linked to outcomes that are specified before task completion, agents will have few incentives to be innovative and find new ways (not foreseen by the initial performance contract) to satisfy stakeholders (Baker, Jensen, & Murphy, 1988, p.598).
In their seminal article, Holmstrom and Milgrom (1991) propose that a reduction or elimination of explicit incentives can be an optimal solution to the multitask problem and suggest a job design approach. Grouping tasks based on the measurability of their outcomes can mitigate some of the distorting effects. Tasks with highly measurable outcomes could be grouped together and assigned to employees with outcome-based pay, whereas their nonmeasurable counterparts are grouped together and assigned to employees with no performance-based pay. Moreover, tasks with negatively correlated outcomes should be separated from tasks with positively correlated outcomes with no explicit incentives for employees performing the former and substantial incentives for those performing the latter.
Implicit incentives: Subjective evaluation, career concerns and tournament theory
Economic theory has identified subjective evaluations of employee performance as an important task of superiors (Baker, Gibbons, & Murphy, 1994). An explicit performance contract can only be based on observable and measurable outcome dimensions that are defined prior to task completion and in which the relative value of the different outcome dimensions is specified ex ante when the contract is designed (Kreps, 1997). Implicit incentives based on the subjective assessment of a superior can take account of unforeseen, innovative contributions to the overall mission, dysfunctional (gaming) behavior on the part of an agent as well as negative outcomes that are not caused by agent behavior and thus eliminate noise and distortion.
Although implicit incentives based on subjective evaluations are more flexible and can take account of the bigger picture, they nonetheless have their own shortcomings. Managers’ evaluations of their subordinates tend to be quite positive (leniency bias) and egalitarian (centrality bias). “Tournaments” in which only one candidate can win the prize—a typical career path principle—can help avoid this. However, they can also lead to demotivation on the part of those who do not win the prize (do not get promoted) and, like all subjective mechanisms, can lead to favoritism, a focus on “signaling activities” and even corruption.
Insights From Social Psychology Research
In the following we will focus on three important contributions to incentive theory made by social psychology research. First, this research elaborates on the importance of intrinsic motivation and the possible impacts of incentives on intrinsic motivation. Second, it extends the notion of implicit incentives. Whereas economic theory focuses on implicit incentives based on subjective superior evaluations, social psychology points at the role of implicit social incentives and self-approval incentives and elaborates on how these interact with explicit incentives. Third, it draws attention to agents’ perception of the work relationship as social or economic and the impact of explicit incentives on this perception.
Explicit incentives and intrinsic motivation
According to Kreps (1997), in the classic agency theory model, “without extrinsic incentives, effort is necessarily at the lowest possible level” (p. 360). However, in their review of the literature on incentives, Fehr & Falk (2002) conclude that “there is no doubt that people engage in many tasks and activities because they enjoy them. Tasks that are inherently satisfying create an intrinsic reward for those performing them.” (p. 713) There is an important branch of literature on social psychology which suggests that the introduction of external forms of motivation for performing a task may decrease the effort put into the task “for its own sake,” particularly in the long run (Frey, 1997). The theoretical arguments in favor of this claim are based either on self-perception theory (Bem, 1967) or cognitive evaluation theory (Deci & Ryan, 1985). According to self-perception theory, people do not have perfect knowledge of their own reasons for performing a task and infer their motives from the circumstances under which they do so. If a task is performed without external rewards, people are likely to attribute their behavior to intrinsic motivation. Conversely, if there is an external reward, they are likely to attribute their behavior to this reward. Because intrinsic motives are fairly uncertain and fuzzy, whereas external rewards—particularly contractually defined, outcome-based rewards (explicit incentives)—are specified clearly, people will attribute their behavior to an external reward even in cases where there is a combination of both intrinsic and external motivation. Introducing external rewards then might crowd out intrinsic motives. Cognitive evaluation theory proposes that people have a psychological need for self-determination, autonomy, and competence. If rewards are perceived as controls that reduce a person’s autonomy, this in turn reduces or eliminates their intrinsic motivation. However, if rewards are perceived as positive feedback on the person’s competence, this may even serve to enhance intrinsic motivation. Throughout the 1970s and 1980s, dozens of experiments investigated the effects of external rewards on intrinsic motivation, producing mixed empirical findings (Cameron & Pierce, 1994; Wiersma, 1992). More recently, it has been argued that it is important to distinguish more specifically between different forms of external motivation to better understand the possible side effects on intrinsic motivation. Whereas contract-based incentives are likely to be perceived as coercive and constraining, incentives that work through the identification with values or moral/social norms (“identified motivation”) are considered compatible and even synergistic with intrinsic motivation (Adler, 1993, Bonner, Hastie, Sprinkle & Young, 2000, Gagne & Deci, 2005).
Explicit incentives and identified/image motivation
Besides the (intrinsic) desire to work on interesting and enjoyable tasks, certain social factors also influence human behavior. These factors play an important role as implicit incentives. Fehr & Falk (2002, p.710) point to the importance of self- and social approval and the interaction effects between a person’s desire for self-approval and social approval on the one hand and the desire for monetary and other “tangible” rewards on the other. When people perform tasks because they want to do good (i.e., are motivated by their identification with a norm), this motivation is typically enforced both by the desire for social approval (to be the objects of other people’s admiration) and by the desire for self-approval (the feeling of doing good/deserving the admiration of others). Ariely, Bracha & Meier (2009, p.544) call this desire to be liked and respected by others and by oneself “image motivation.” Providing explicit rewards for an activity may destroy both self-approval and social approval incentives. In particular, rewarding people in monetary form for obeying social norms may lead to a gradual erosion of norm-guided behavior (Fehr & Falk, 2002, p.711). This might, for example, explain a classical observation made by Titmuss (1971), who claimed that paying people to donate blood can actually reduce the number of donors (Mellstrom & Johannesson, 2008). Fehr and Falk (2002, p.708) make a similar point with regard to disapproval incentives and illustrate this by referring to an experiment by Gneezy & Rustichini (2000a). In this experiment, the introduction of a (small) fixed fine for picking up children too late from daycare did not induce parents to collect their children on time but instead led to a steep increase in the number of latecomers. Giving norm violators the opportunity to free themselves from social disapproval by making them pay for this violation may be seen as revelation of the price for being late. If this price is lower than the nonpecuniary cost of being late (social disapproval), parents will choose to come late. Once the violation of a morally legitimated rule is transferred to a commodity with a price, it is difficult to undo this transformation by removing the fine. Ariely et al. (2009) give a “signalling value” explanation for possible interaction effects between explicit incentives and implicit approval/image motivation. When an individual engages in prosocial behavior in the hope of social approval, the addition of explicit (financial) incentives makes it more difficult (for the individual and others) to answer the question: “Is the individual behaving prosocially to do good or to do well?” Ariely et al. illustrate this point using the example of the purchase of an environmentally friendly hybrid car which is more expensive than an equivalent car without such technology. Whether driving such a hybrid car will really improve one’s positive image (in a community that values environmentally friendly behavior) depends on whether there is a large tax benefit for the buyer or not. A tax benefit that makes environmentally friendly behavior pay off may be detrimental not only to social approval, but also to self-approval. The results of the laboratory experiment by Ariely et al. suggest that people want to be seen by others to be doing good. With explicit (monetary) incentives, the signal of a prosocial act is diluted. However, the impact of monetary incentives is shown to depend crucially on visibility. If others cannot see that an individual is doing good and doing well at the same time, there may be no decrease in image value. Explicit incentives, by definition, are based on a contract that is verifiable by a third party. This high visibility of explicit incentives makes them most likely to dilute prosocial and norm-guided behavior.
Explicit incentives and the perception of the work relationship
Gneezy and Rustichini (2000b) suggest that the terms of payment within employment relationships play a role in the way these relationships are perceived. An adequate fixed payment for a task may be perceived as an implicit agreement that embodies a moral obligation (based on social norms of cooperation) on the part of employees to provide (extra) effort. When explicit incentives are introduced, the perception may be that extra effort should be rewarded in line with the performance contract and that effort will be minimal without such rewards. Benabou and Tirole (2003) argue that contractually defined external rewards may be interpreted as a signal of the principal’s distrust in the agent’s abilities or as a signal that the task is considered unattractive by the principal. In contrast, subjective ex post rewards provided by the principal (i.e., implicit incentives) may even boost the agent’s self-esteem or intrinsic motivation because he/she interprets such rewards as a signal that the task was considered difficult and that his/her performance is appreciated. Similarly, Falk and Kosfeld (2006) argue that agents seem to interpret ex ante control as a signal of a lack of trust on the part of the principal in the agent’s abilities and offer empirical evidence for this claim. Deckop, Mangel, and Circa (1999) provide evidence that the motivational effects of performance pay depend on whether the relationship is perceived as economic or social. Heyman and Ariely (2004) also argue and provide evidence that the effect of rewards on effort depends on whether a relationship is perceived as a social relationship or a market relationship. Whereas effort in the former is shaped by altruism and social concerns, in the latter it will depend directly on payment (p. 788). Interestingly, they find evidence that the form of payment plays an important role. Whereas (subjective ex post) rewards in the form of gifts do not seem to transform a social relationship into a market relationship, monetary rewards (or even mentioning the monetary value of a gift) can transform the social relationship into a market relationship where individuals exert effort according to reciprocity and only do what they are paid to do.
Nonprofit Characteristics and the Effectiveness of Incentives
The defining characteristic of a nonprofit organization is the nondistribution constraint, that is, a nonprofit organization is not allowed to distribute its surplus resources in financial form to those who control the organization (Hansmann, 1980). Economic theories of nonprofit organizations (Steinberg, 2006) analyze which distinctive characteristics of an organization and its role make such a constraint economically meaningful and, hence, set a nonprofit organization apart.
Demand-Side Theories: Collective Good and Contract Failure/Trust-Related Theories
Collective good theories explain the existence of nonprofit organizations by the specific properties of the goods and services they provide (Weisbrod, 1988). Economic theory shows that markets are not able to provide adequate quantities of collective goods (“market failure”). Government provides collective goods in quantities and qualities that reflect a “sufficient political consensus” yet leave a dissatisfied minority (“government failure”). Those who want higher levels than those provided by government support nonprofits. Typically, collective goods involve multiple (positive) externalities, that is, the provision of the good or service generates a multitude of valuable outcomes for a variety of (ill-defined) individuals who take the role of “principals/stakeholders,” but often lack both the incentives and power to actively monitor the relationship with the nonprofit organization (“common agency problem,” Steinberg, 2010).
Whereas collective good theories point out that governmental undersupply of collective goods gives rise to private organizations whose main purpose is to provide such goods (and not the maximization of profits from selling goods), contract failure or trust-related theories help to explain why certain characteristics of goods and services make a nondistribution constraint meaningful for the organization that provides them. When the quantity and quality of services is difficult to verify, the owners of profit-seeking firms have strong incentives to take advantage of such information asymmetries. Organizations that cannot distribute profits to those who control the organization have weaker incentives to do so. More generally, the nondistribution constraint credibly signals the trustworthiness of an organization to stakeholders who have entered into a difficult-to-monitor relationship with the said organization and do not control it (Hansmann, 1980). Hence, such stakeholder relationships can be characterized as trust-based social relationships.
Collective good theories point out in particular that nonprofit organizations often have stakeholders who receive goods and services from the organization but have no power or incentive to control the quality of such services (this is typically the case for recipients of the services of “donative nonprofits”). As indicated by Hansmann (1987), even controlling stakeholders of donative nonprofits (e.g., donors or volunteer workers) often lack information on whether their (marginal) contribution leads to a better attainment of the organization’s purposes (p. 30). In a similar vein, contract failure/trust-related theories explain that stakeholders may well have the power and incentive to monitor the services they receive from the organization, but lack the information and knowledge to assess its performance (this is typically the case for customers of “commercial nonprofits”).
Taken together, these demand-side theories point at two main “characteristics of nonprofits” that are relevant when it comes to the use of incentives. First, when an organization’s primary purpose is to provide collective goods, it is very difficult to transform this into a consistent and measurable organizational goal system. When the attainment of the primary organizational goals depends on the benefits generated for multiple ill-defined beneficiaries, who lack both the power and the incentive to articulate their preferences and monitor the services provided, such goals tend to be multidimensional and somewhat vague. Typically, activities that increase the output for one collective good (benefits for certain beneficiaries) simultaneously decrease output for other collective goods (benefits for other beneficiaries) and it is then unclear, whether such activities lead to overall value increases. In other words, nonprofit organizations who provide collective goods often face severe tradeoffs, yet at the same time have little information on how to deal with them. Without a well-defined and consistent overall organizational goal system that can be decomposed and cascaded down into subgoals at the business unit, project or process level, it is difficult to define performance measures and draw up performance contracts (Weisbrod, 1988). Second, nonprofit organizations are characterized by trust-based relationships between the organization and its stakeholders which are facilitated and strengthened by the nondistribution constraint. The trust-based nature of these relationships calls for a trust-based style of control toward the employees (who contribute to the provision of goods and services for trusting stakeholders). For example, offering nurses explicit incentives which reward the number of patients served or tasks completed in a given period of time, but not the nonvisible (noncontractible) quality of services, might motivate them to focus on the rewarded aspects of performance while cheating on quality (the multitask problem mentioned above)—exactly the type of incentive the nondistribution constraint seeks to prevent. Although dimensions of performance that are noncontractible vis-à-vis customers will hardly be contractible between superiors and employees, it does seem possible that some aspects of employee performance which are not observable by customers can be assessed by superiors ex post, making the use of implicit incentives based on subjective superior assessment applicable.
In contrast to nonprofits, firms whose primary purpose is to maximize profit 3 have a simple, one-dimensional primary goal that is relatively easy to measure and apply in pay-for-performance contracts at the top management level. Moreover, “profit” provides a common language across different projects, products, business units and even firms, allowing the use of relatively standardized (though not perfect) internal and external financial information and control systems whose main purpose is to assess (for internal managers as well as external stakeholders) whether each dollar of investment into product quality, production process design, employee training or a particular business unit or project yields more than one dollar in return (see e.g., Speckbacher, 2003; Speckbacher, Bischof & Pfeiffer, 2003). Owners of profit-seeking firms have much better information on the outcome/return on their investment than the primary stakeholders in nonprofits. When financial success is the main reason for the existence of an organization, a focus on financial incentives is also in line with the overall organizational purpose. Moreover, financial measures (profits, revenues) can be used by principals to measure agent performance, although all decisions on how to achieve this performance (which products to offer to whom, how to produce them) are delegated to the agents. In contrast to nonprofits, whose ultimate purpose is to provide particular services and products, profit-maximizing firms view the provision of products and product quality only as a means to an end (profit maximization). Hence, they only invest in product quality if the return on this investment exceeds the investment and do not produce products that are not profitable, regardless of their “social value.”
Supply-Side Theories: Social Entrepreneurship, Self-selection, and Social Incentives
Although collective good and contract failure theories explain the demand for nonprofit organizations and their important role in the economy, they do not explain why people found a nonprofit organization. This question is addressed by supply-side theories of nonprofit organizations. Although the right to appropriate financial profits is a main incentive for founding a profit-seeking firm, the nondistribution constraint explicitly excludes the use of this incentive for nonprofits. Supply side-theories of nonprofit organizations point out that the lack of such explicit financial incentives may serve as a self-selection mechanism to attract those entrepreneurs whose interest depends not (only) on extrinsic financial incentives but more on their intrinsic motivation or identification with a particular mission. Moreover, the nondistribution constraint prevents the founders of nonprofits from withdrawing their initial investment once other individuals have been put in place to support the organization and therefore functions as a signal of their strong and enduring belief in the importance of the organization’s mission (Steinberg, 2006). The motivation and values of an organization’s founder are considered to have a strong influence on organizational culture and workforce values (Badelt, 1997; Young, 1983). Existing literature also contends that intrinsically motivated employees at the middle or lower management and grassroots levels may well opt for nonprofit organizations. Job applicants who compare for-profit and nonprofit organizations may tend to choose a nonprofit organization when its mission corresponds to their own values and goals (Ben-Ner, Ren & Paulson, 2011; Callen & Falk, 1993; Handy & Katz, 1998; Steinberg, 1990b;Young, 1983). Because the people in control of a nonprofit organization have fewer incentives to take (financial) advantage of the altruistic motivation of (powerless) stakeholders to contribute to the organization’s purpose, the nonprofit character functions as a signal of trust. Nonprofit organizations can therefore be more successful at attracting intrinsically motivated workers (Hansmann, 1980; Roomkin & Weisbrod, 1999; Rose-Ackerman, 1996). A recent stream of economics-based literature builds on these ideas and analyzes the motivational characteristics of workers in public services organizations (Francois, 2003, 2007). According to this literature, the nondistribution constraint ensures employees that their prosocial behavior really will increase the quality of services and not be turned into owner profit. For example, when workers in a (for-profit) nursing home are prepared to work overtime in a prosocial attempt to increase service quality (which is unobservable by patients), the owners may anticipate this behavior and understaff the respective departments—turning the intended quality improvements into cost savings and, ultimately, financial profit. The nondistribution constraint lowers incentives for those in control of the organization to exploit prosocial workers. According to Besley and Ghatak (2003, 2005), prosocially motivated employees will be attracted by organizations whose primary goal and mission correspond to their own values and altruistic preferences.
Unfortunately, empirical evidence on whether nonprofits can rely on a better motivated workforce is extremely thin. Some authors interpret evidence on the reduced use of bonuses and performance pay or lower wages as an indication of higher levels of intrinsic motivation. However, one has to be careful with such an interpretation, because there is a lack of knowledge on the exact relationship between the nonprofit character of organizations, pay structures/levels and employee characteristics (for a recent discussion of this topic see Ben-Ner et al., 2011; see also the model provided by Slivinski, 2002). Moreover, studies on differences in pay structures/levels between for-profit and nonprofit organizations are mostly limited to industries where both types of organization coexist (e.g., hospitals, nursing homes) and evidence on the reduced use of performance pay (Roomkin & Weisbrod, 1999) and lower wage levels (Ballou & Weisbrod, 2003; Preyra & Pink, 2001; Roomkin & Weisbrod, 1999) is largely restricted to the executive level or managerial and highly skilled nonprofit employees. In a comprehensive cross-industry study, Leete (2001) finds that there is no simple, economy-wide nonprofit effect on compensation (in the sense that nonprofits pay lower wages and/or bonuses in general).
On the whole, the view that nonprofit managers and workers are more altruistic and intrinsically motivated lacks persuasive theoretical foundation and systematic empirical evidence, although it is widespread in the literature (Ben-Ner et al., 2011; Caers et al., 2006; Schepers et al., 2005). Only very few studies offer evidence in direct support of higher intrinsic motivation on the part of employees in the nonprofit sector (Borzaga & Tortia, 2006; Tippet & Kluvers, 2009).
This lack of evidence may be due in part to an unclear theoretical specification of the expected differences in motivation on the part of nonprofit workers. In a strict sense, intrinsic motivation means that people choose to perform an activity of their own free will, simply because they enjoy it, have fun doing so or find it interesting. But such motivation may not be most typical for nonprofits. Research in psychology (e.g., Burton, Lydon, D’Alessandro, & Koestner, 2006) points at the importance of distinguishing intrinsic motivation from identified motivation, which is characterized by an individual’s perception of the task being important, meaningful, valuable, or worthwhile (the individual identifies with the task or its intended outcomes). Prosocial motivation in the form described by the above cited research on public services organizations is a form of identified motivation. When their identification with norms and values drives people to perform an activity, this does not necessarily imply that they enjoy what they do or are fascinated by the task. Instead, they may even perform the activity if it involves painful effort, drawbacks or frustration (Burton et al., 2006). Identified motivation depends on what a person believes in and considers a legitimate norm. As argued in the above section on social psychology research, a person’s identification with and willingness to act in accordance with specific norms and goals (even when this might involve pain) is reinforced by the desire for social approval and for self-approval. Although some nonprofit workers may be motivated by the fact that their tasks are enjoyable, exciting or interesting, the above cited literature on the enduring belief of nonprofit entrepreneurs in the importance of the organization’s mission and the influence of their values and visions on subordinate employees and the self-selection of employees who support this mission emphasizes identified motivation, not intrinsic motivation. In particular, when comparing nonprofit and for-profit organizations operating in the same industry, it is unclear why a task which is enjoyed by workers in a nonprofit organization is not enjoyed when the same task increases a firm owner’s profit. However, seeing someone else making a profit from one’s own strenuous efforts to serve a good purpose (identified motivation) may be discouraging, in particular when the increase in profit comes at the expense of the promotion of this purpose (see the above example on prosocial behavior).
Hence, the role of the nondistribution constraint in enabling or supporting intrinsic motivation may be quite limited. But whether the primary and ultimate purpose of an organization is to make a profit for its owners or whether it is to provide “valuable” goods and services can make a significant difference with respect to identified motivation. Of course, identified motivation can also be found among workers in for-profit organizations. Engineers, for example, can be motivated by their identification with customer needs or the goal of producing technically outstanding products (Markus, Manville & Agres, 2000). However, their superiors or the firm’s financial controllers will continuously remind them that the firm is not there to make customers happy or produce technical gimmicks, it is there to make profit (a goal conflict between employees and owners). Drucker (1989) adds to these observations with his assertions that it is (a) more difficult for people to identify with an abstract goal like profit maximization than with a mission like delivering beneficial services to customers and that (b) profit maximization involves obvious conflict between stakeholders, whereas mission-orientation avoids such conflict to a far greater extent.
The Design of Incentive Packages in Nonprofits
The above discussion of existing nonprofit organization theories highlights three “nonprofit characteristics” that have been argued to affect the applicability and effectiveness of incentives: (a) the lack of undistorted contractible measures to assess the organization’s overall performance and derive a coherent system of subgoals at the business unit, project, and process level (cascading down measures and assessing tradeoffs between subgoals), (b) the motivational characteristics of employees working in nonprofits, in particular the importance of identified motivation, and (c) the trust-based, social character of the relationships between the organization and its stakeholders. The extent to which each of these characteristics is present may well vary across different nonprofit organizations and also across tasks within such organizations. Generally, the literature discussed in the section on the theoretical foundations of incentives indicates that point (a) negatively affects the application of explicit (contract-based) measures, whereas points (b) and (c) increase the role of implicit incentives. Because employees are regularly motivated by diverse sources simultaneously, including intrinsic and identified motivation, approval incentives and performance-based rewards (see e.g., Markus, Manville, & Agres, 2000), each particular nonprofit organization faces the challenge to design “incentive packages” which take adequate account of the importance of (a), (b), and (c) for particular tasks and organizational subunits.
When an organization is able to attract employees who autonomously choose to perform well because they are intrinsically motivated or identify with the task and its outcomes, explicit or implicit incentives seem unnecessary. However, sole reliance on intrinsic motivation can be problematic. According to Prendergast (2008), low contractibility of outputs may lead nonprofit organizations to hire agents who have a strong intrinsic interest in some aspects of their tasks, yet ignore others. When a social worker produces a noncontractible output for clients, a natural response to this output noncontractibility would be to hire agents who are highly intrinsically motivated to serve clients. However, this is likely to come at the cost of their ignoring aspects like cost control. To take account of the latter, a cost controller with excess interest in cost control may then be hired. However, a strategy of hiring biased agents, whose biases move in opposite directions, generates a conflict of interest. Organizations cannot normally attract employees whose intrinsic motivations are a perfect match for all relevant dimensions of their tasks, that is, workers will be more intrinsically motivated by some aspects of their jobs than others. Building teams of agents with diverse intrinsic motivations in which each agent specializes on his/her preferred task dimension(s) (i.e., a job design approach) helps to align individual motivations with overall objectives at the team level, but also generates conflict and additional coordination costs within the team.
In contrast to intrinsic motivation (the most autonomous form and difficult to control using incentives), identified motivation can be influenced directly using incentives. In addition to affirmation of mission-congruent behavior by superiors, even individual behavior not observable by superiors can be rewarded by coworkers (e.g., through praise or social recognition and esteem), whereas behavior only noticeable by the individual in question can be regulated by self-approval incentives. This, however, presupposes that coworkers (and the individual) have internalized mission-related behavior as a moral principle or norm. In theory at least, identified motivation can avoid the conflicts that arise from diverse intrinsic motives. Identified motivation can lead employees to perform any task if they believe it contributes to the purpose they identify with and value—whether they like the task (in terms of intrinsic motivation) or not. When workers identify with organization’s objective to make maximum use of funds to serve clients in need, they will also be motivated both to serve those clients and be cost efficient. This in turn presupposes that workers not only share a common mission, but also understand (and agree) how different tasks contribute to promoting the overall mission. Whereas the former calls for mechanisms to enhance goal commitment (e.g., mission statements), the latter calls for strategic performance management systems that clarify and communicate each task’s contribution to overall mission achievement (Speckbacher, 2003). In fact, using implicit instead of explicit incentives does not imply that performance measurement becomes less relevant. Whereas explicit incentives link rewards to performance measures, implicit incentives make use of performance measures mainly for information purposes. Superiors use such information to assess the performance of their subordinates (e.g., in promotion decisions), coworkers use it to evaluate an individual’s contribution to mission achievement (e.g., social/image incentives), and the individual uses it for self-regulation purposes (e.g., self-approval incentives). Because implicit incentives are softer and less precise than explicit incentives, they provide less incentive for opportunistic manipulation of these performance measures. This is particularly important for tasks or projects where outcomes are difficult to measure and there is no clear bottom line, because this makes the measures relatively easy to manipulate. Linking incentives to incomplete performance measures may not only lead to opportunistic maximization of rewards (you literally get what you pay for, that is, better numbers, but not better overall outcomes for the organization; see Baker, 2002), it can also distort the information content of these measures, thereby undermining the efforts to improve the measurement of performance in projects and programs.
Obviously, the tasks to be performed within nonprofits often greatly differ to those in for-profit organizations in terms of their characteristics and, in particular, the contractibility of relevant outcomes. Although Prendergast’s example assumes noncontractibility of services for clients, cost-related goals might be well contractible, that is, the cost controller could also be motivated by explicit incentives. Building on Holmstrom and Milgrom’s (1991) job design approach and using Weisbrod and Schlesinger’s (1986) distinction between “type 1 dimensions” of quality (those which are relatively easy to monitor and assess) and “type 2 dimensions” (those which are difficult to monitor), one could decompose tasks into “type 1 jobs” and “type 2 jobs” using explicit incentives for the former (without undermining the nondistribution constraint) and intrinsic or identified forms of motivation for the latter.
The evidence on incentives in Job Training Partnership Act (JTPA) centers provided by Heckman, Heinrich and Smith (1997) and Heckman, Smith and Taber (1996) even suggests that a mix of intrinsic/identified motivation and motivation through explicit incentives may be possible and useful (see also Baker, 2002). JTPA centers are mandated to provide job training for the disadvantaged. Their funding depends on their performance in successfully placing their clients. However, this performance measure does not adequately reflect the “real” ultimate purpose of increasing the employability of the disadvantaged. Instead, it gives JTPA managers incentives to play the system by only accepting into the program those clients most likely to place well and strategically terminating clients from the program to maximize the reported numbers (“cream skimming”). According to data provided by Heckman et al., less disadvantaged clients may benefit more from the offered job training than the most disadvantaged. Employees who work at the centers, however, show (identified) motivation to help the most disadvantaged, that is, they want to serve those who may get less out of the program (a case of identified motivation that is noncongruent with the overall mission which might also be resolved by explication, communication, and explanation of the organizational mission). Heckman et al. argue that the performance standards, which encourage cream skimming, may actually increase the value added of the program by giving JTPA center employees an incentive to attract clients who will benefit more. Hence “distorted” explicit incentives may be used beneficially to “adjust” employees’ intrinsic or identified motivation.
However, one has to be extremely careful when combining performance related pay (in particular contractual, i.e., explicit, incentives) with intrinsic or identified forms of motivation. Besides the widely discussed crowding-out effects, explicit incentives might affect implicit social approval and self-approval incentives, which serve to reinforce identified motivation. Because explicit incentives are strong signals of what really counts, they will affect the established system of shared norms and goals which underlies identified motivation. Moreover, employees who do not share these norms and goals, but who know that they perform well in the rewarded dimensions, may self-select into the organization, thereby undermining the established system of norms. Eventually, this process may lead to a significant increase in performance in the rewarded dimensions, but a serious erosion of performance in the nonrewarded dimensions.
Explicit incentives also affect perceptions of the work relationship. When nonprofits change their incentive packages and make more use of explicit incentives, this may change the type of employment relationship from social to market (shaped by “quid pro quo considerations”), thus undermining the functioning of (implicit) social incentives. Introducing market elements (contract-based incentives) into a social relationship may turn the whole relationship into a market relationship, and the advantages of the social nature of relationships between superiors and subordinates (trust and the feeling of obligation) may disappear. This suggests that providing explicit rewards for some performance dimensions while relying on social incentives in others may not work, because the frame of a relationship is either social or economic, but rarely both. Spillover effects across different workers may also be found. The use of explicit incentives for type 1 job workers (e.g., cost controllers) might have negative effects on type 2 job workers (social workers) in the same organization, because it may become difficult for superiors to maintain social relationships with some employees, yet at the same time work with contract-based rewards for others. In such cases, the decoupling of type 1 and type 2 tasks into different organizations may appear to be a solution, although this can increase the interorganizational coordination costs required to optimize overall mission achievement (see e.g., the Harvard University example provided by Baker, 2002).
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Notes
Author Biography
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
