Vkunz_2002_accounting,-organizations-and-society

Accounting, Organizations and Society 27 (2002) 275–295 www.elsevier.com/locate/aos

Agency theory, performance evaluation, and the hypothetical construct of intrinsic motivation Alexis H. Kunza,*, Dieter Pfaffb a

University of Applied Sciences Aargau (Northwestern Switzerland), Martinsberg, CH-5401 Baden, Switzerland b Institute for Accounting and Control, University of Zurich, Plattenstrasse 14, CH-8032, Zurich, Switzerland

Abstract Cognitive evaluation theory and its hypothetical construct of intrinsic motivation are enjoying increasing popularity in the fields of business administration and economics. Consequently, intensifying skepticism towards performance incentives and agency theory is postulated. According to cognitive evaluation theory, it is argued that performance pay may undermine an agent’s intrinsic motivation. In contradiction to agency theory, the principal might be worse off when providing an incentive contract to the agent than without doing so. Since the contention is substantiated by empirical evidence, it seems worrying enough for further investigation. Restricting attention to performance pay in business corporations, the scope of this article is to evaluate whether agency theory faces an urgent need to incorporate the construct of intrinsic motivation and its ‘hidden costs of reward’ as postulated by supporters of the concept. The subsequent analysis reveals good and bad news for agency theory. The bad news is that hidden costs of reward do indeed exist. The good news is that the empirical evidence on undermining effects cannot be interpreted as being contradictory to agency theory. In particular, the antecedents for such effects not only seldomly prevail in business corporations, they are also easily avoidable. # 2002 Elsevier Science Ltd. All rights reserved.

There is no doubt that agency theory and its advocated view of the firm as a complex nexus of contracts constitutes one of the major pillars of theoretical accounting. As such it not only helps to understand and explain the behavior of business actors, but also provides a rich fund of practical implications for the design of governmental structures such as transfer prices, the delegation of residual rights or the design of managerial incentive contracts. In particular, the overwhelming empirical preponderance of pay for performance systems has often been taken as circumstantial

* Corresponding author. E-mail address: [email protected] (A.H. Kunz).

evidence for the theory’s contribution to real-life situations (Milgrom & Roberts, 1992). Despite this, standard agency theory has often been criticized because of its presumptions about human behavior. Especially some psychological and sociological concepts refer to disparate assumptions about human behavior and seem to have established a certain rivalry of thoughts between the disciplines. Overall, there exists quite a disagreement about the consequences of statecontingent performance pay. On the one hand, agency theory asserts that increased performance incentives ceteris paribus raise the agent’s productivity when risk considerations are omitted. On the other hand, the psychological concept of intrinsic motivation may predict quite the opposite. In accordance with cognitive evaluation

0361-3682/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved. PII: S0361-3682(01)00031-9

276

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

theory, it has been pointed out that performance incentives through state-contingent extrinsic rewards may diminish an agent’s intrinsic motivation (Ryan & Deci, 2000). Thus, in contradiction to agency theory, a principal can be worse off when providing an incentive contract to his agent than without doing so. Such detrimental consequences of rewards have been termed the corruption effect of extrinsic motivation (Deci, 1975), hidden costs of reward (Lepper & Greene, 1978) or the crowding-out effect of intrinsic motivation (Frey, 1997). At present, the psychological concept of intrinsic motivation enjoys increasing popularity in the fields of business administration and economics. As a consequence skepticism towards performance incentives and formal agency theory has intensified both in literature and in practice (Brittan, 1997; Frey, 1997; Osterloh & Frey, 2000; Tosi, Katz, & Gomez-Mejia, 1997). Supporters of the intrinsic concept argue that economic theory advocates a myopic view of personal motivation which might induce severe shortcomings when it comes to real life applications. Referring to empirical evidence of detrimental effects due to extrinsic rewards, they plead for an urgent need to incorporate intrinsic motivation into economic theory and real life compensation plans (Frey, 1997). Some authors even criticize the dominance of extrinsic-oriented rewards in business compensation plans and request stronger intrinsic-oriented incentive systems which from their point of view might be an as-yet unexplored competitive advantage (Osterloh & Frey, 2000). The categorical assertion that rewards lessen task performance has profound practical and theoretical implications. First, any popularization of such views can foster public attitudes against the use of tangible rewards to promote socially desirable behavior (Eisenberger & Cameron, 1996). Secondly, one is likely to jump to the conclusion that agency theory may map human behavior inadequately and should therefore be modified or be skipped accordingly. So far, only a few economists have turned their attention towards the construct of intrinsic motivation. Kreps (1997, p. 360) talks about a ‘stylized fact’ but he still thinks it may be worthwhile for future research since, ‘there may be nothing to explain,. . .(but) abun-

dant smoke signifies a fire, and the assertion is too strongly rooted in folk wisdom to be entirely hot air.’ Restricting attention to performance evaluation in business corporations, the scope of this article is to clarify whether agency theory faces an urgent need to incorporate the construct of intrinsic motivation. Since this problem addresses both empirical and theoretical aspects, a two fold analysis will be carried out. Reviewing the construct of intrinsic motivation, we first investigate whether its underlying theory may serve as a model for incorporating intrinsic motivation into the agency perspective. The second analysis concentrates on the empirical phenomenon of undermining effects. Reviewing experimental evidence, we investigate whether the effect is of any practical relevance, and if so, under what precise conditions we have to account for it. Our analysis reveals good and bad news for the agency perspective. The bad news is that hidden costs of reward do indeed exist. The good news is that the empirical evidence on undermining effects cannot be interpreted as being contradictory to agency theory. In particular, the antecedents for negative after-effects of reward seldomly prevail in business corporations. In addition they are also easily avoidable. The remainder of this article is structured as follows: the second section focuses on performance evaluation and agency theory from both a theoretical and an empirical point of view. Section 2 and 3 elaborate definitions, conceptualizations and theories of intrinsic motivation. Empirical evidence on undermining effects, particularly on their antecedents, is scrutinized in Section 4. Section 5 draws some conclusions.

1. Agency theory, performance evaluation, and compensation 1.1. The agency theoretic perspective Modern economic organizations are complex team-productions, since their output is jointly produced by several-input owners, e.g. stakeholders, managers and employees (Alchian & Demesetz, 1972). As the team is forced to achieve

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

some level of efficiency, tasks are delegated to specialized production units which act on behalf of others. Specialization inevitably involves the dispersion of knowledge and information; in addition, it often entails goal incongruencies between the participants. Since the combination of both asymmetric information and conflicts of interest may result in an efficiency loss, procedures and mechanisms are needed to mitigate this problem. The function of performance evaluation in such a setting is twofold. First, it aims to control discretionary behavior by aligning interests through the provision of state-contingent incentives. Secondly, it evaluates the contribution of each input-owner to the overall output, thereby administering compensation conditional upon individual performance. Being a theoretical branch of game theory, agency theory represents an economic methodology for analyzing and assessing the efficiency of such mechanisms (Baiman, 1990; Lambert, 2001). The standard agency model involves a (risk-neutral) principal, employing a (risk averse) agent to act on his behalf. The agent possesses private information, e.g. about his effort level, the state of nature etc. that is not costlessly available to the principal. It is supposed that the agent chooses actions to maximize his utility. He is assumed to be work averse in the sense that tempting off-thejob opportunities may instigate him to reallocate his effort to maximize his overall utility resulting from payoffs both, on and off the job. The combination of information asymmetry and the agent’s aversion both to work and risk, steer him away from cooperative behavior. Since the output is conditional upon factors which are exogenous to his effort, like the state of nature, the principal cannot infer the agent’s performance from the overall result. Thus, while designing an incentive contract, the principal must prod the agent’s work effort by forcing him to bear at least some of the production risk. According to standard agency theory, an optimal incentive contract involves a pay-for-performance scheme which ties the agent’s pay-off to production indicators which (partially) correlate with his effort level. The intuition behind this result is the mimicry of a statistical inference problem, making pay-outs contingent upon the

277

likelihood that the desired actions were in fact taken (Holmstrom, 1979). The derivation of the optimal contract results from a mathematical program maximizing the principal’s objective function, subject to several constraints, e.g. production and technology constraints and the agent’s characteristics, e.g. risk and work aversion, information, action space and reservation utility. The principal offers a contract to the agent, who then chooses those actions that maximize his overall utility restricted by contractual constraints. Since the agent simultaneously faces on- and offthe-job-opportunities, providing effort is costly to him. When deciding upon his optimal action, the agent has to trade off effort costs against the expected utility resulting from the monetary (salary) and non-monetary (prestige) consequences of his decision. The precise conditions of an optimal contract are heavily dependent on the prevailing situational characteristics, but in most cases they involve a variable component varying with an indicator of the agent’s effort level. The efficiency loss due to the agent’s self-interested behavior is measured by comparing the effective outcome under asymmetric information with a fictitious outcome under symmetric information. Since the latter serves as an unattainable focal point of the analysis, it is referred to as the first-best solution, whereas the former represents the so-called second-best solution. Although attainability of the first-best solution is impossible in most cases, its approximation through an incentive contract is not. Consequently, agency theory represents a methodology for comparing and assessing such contractual designs as far as their efficiency is concerned. Although a countless number of different types of agency models exist in the literature (for overviews see Mas-Colell, Whinston & Green, 1995; Milgrom & Roberts, 1992), they all share at least two main characteristics: the presumptions about conflicts of interest, and informational asymmetries between the parties. If these necessary conditions were relaxed, the problem would become trivial. On the one hand, symmetric information simply allows the principal to prescribe and control the desired action. On the other hand, aligned interests between principal and agent make motivation needless, since the

278

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

required actions already belong to the very (self)interest of all actors. If both conditions, asymmetric information and conflicts of interest, hold then an optimal incentive contract needs both to partition decision rights and to control discretionary behavior. In addition, it aligns interests between the parties by tying compensation to production indicators which at least partially correlate with the agent’s action (effort level). Although the agency perspective recommends rewarding agents on the basis of their performance while explicitly supposing that increased incentives usually raise productivity, the use of incentive systems is not unrestrictedly advocated. Agency theory not only constitutes a balanced framework for evaluating efficient incentive contracts, it additionally provides a rich fund of insights on their limitations and pitfalls, thus explicitly accounting for the inherent dangers of reward systems. Different models discuss dysfunctional effects of explicit incentives within the concept of agency theory, e.g. due to the agent’s (severe) risk aversion, his limited liability, multi-tasking problems or the case of sufficiently noisy indicators about the agent’s performance level (for overviews see Baker, 1992; Gibbons, 1998; Lazear, 2000). 1.2. Empirical evidence Assessing the empirical status of the theory provides a mixed but optimistic picture; its major predictions concerning the interaction of incentives, risk-bearing properties, and monitoring technologies are in general supported by empirical research (for overviews see Eisenhardt, 1989; Lazear, 2000; Prendergast, 1999). Nevertheless, the empirical picture additionally shows that the overall effect of incentives on subsequent performance might involve large variances in performance outcomes, giving a strong rationale for the importance of contingency factors (Lanen & Larcker, 1992). On the one hand, performance variances of incentives are explicable by arguments that not only are consistent with the theory but also belong to its very core, e.g. (1) the agent’s risk aversion, affecting his sensitivity towards the incentive contract (Haubrich, 1994; Murphy, 1999; Oyer, 2000); (2) individual rent-seeking

activities (Jensen & Smith, 2000); (3) partial substitution of the incentive contract through monitoring technologies (Beatty & Zajac, 1994); or (4) noisy performance indicators (Gibbons, 1998). On the other hand one could just as well argue that performance variances might be caused by variables which are beyond the scope of agency theory. Indeed, experimental economics seems to have identified some behavioral patterns which at present are not sufficiently integrated in the theory’s axiomatic (for overviews see Davis & Holt 1993; Fehr & Gaechter 2000a, b; Kagel & Roth 1995). Examples are, for instance, fairness considerations, interpersonal reciprocity in bilateral bargaining contexts (Davis & Holt 1993; Fehr & Gaechter 2000a), violations of the theory’s rationality assumptions (Camerer, Babcock, Loewenstein & Thaler, 1997; Conlisk 1989) or dysfunctional effects of fines (Gneezy & Rustichini, 2000). At present, there is an ongoing controversy whether some of these effects are primarily a matter of insufficient financial incentives during the experiments or whether they are really challenging the theory (Davis & Holt, 1993). Economists argue that although such anomalies may be troublesome, they need not be necessarily fatal, ‘particularly if there exists no alternative theory that explains both the anomalies and the standard patterns of behavior’ (Davis & Holt, 1993, p. 510). In addition, as ongoing research demonstrates, economists are dealing seriously with these anomalies, trying to reintegrate them into their axiomatic. Initial progress has already been made concerning the modeling of self-perception and social interactions (Benabou & Tirole, 2000), reciprocity (Falk & Fischbacher, 2000; Fehr & Schmidt, 1999; Rabin, 1993), learning (Fudenberg & Levine, 1997) and the impact of social norms on voluntary cooperation (Fudenberg & Levine, 1997; Kandori, 1992; Knack, 1992). A common characteristic of all these models is the attempt to explain voluntary cooperation (or dysfunctional effects of explicit incentives respectively) within the very axiomatic of economic theory, e.g. without referring to the concept of intrinsic motivation. Benabou and Tirole (2000), for example, offer an explanation for hidden cost of rewards in terms of agency theory. In their

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

model, the principal has superior information about the agent’s ability to successfully accomplish a task. In such a setting, increased monetary incentives may backfire when interpreted by the agent as a signal of low competence, diminishing his self-confidence and his overall motivation to work vigorously. Summarizing, agency theory and its related axiomatic proves to be a powerful methodology. The implications of agency theory are empirically testable and have empirical support. Nevertheless, agency theory and game theoretical modifications thereof have to be considered as theories which are still in progress. So far, the permanent interplay between theory and empirical work has induced a process of continuous refinement of the theory’s axiomatic, thereby broadening its explanatory and predictive power rather than exposing any severe limitations. In the next section, we investigate to what extent the theory’s recommendations on the practical use of incentives might be disputed due to the agent’s intrinsic motivation.

2. Intrinsic motivation as a hypothetical construct 2.1. Definitions and conceptualizations Hypothetical constructs relate experiences by inventing a fictitious substance or process in terms of which these experiences can be expressed, thereby involving the supposition of entities or processes which are neither among the observed nor can be measured directly (Benjamin, 1937; MacCorquadale & Meehl, 1948). Intrinsic motivation and the crowding-out effect are hypothetical constructs in the sense that their existence is not empirically proven, but rather results from plausible hypothesization. As imaginary (re)constructions of unobservable psychological processes, they aim to explain behavioral outcomes stemming from a preceding constellation of antecedents. One point is of major importance for the following discussion: it has to be carefully distinguished whether one refers to intrinsic motivation as a hypothetical construct, or whether attention is directed to the empirical evidence of externally influenced performance lessening. The

279

latter is the empirical phenomenon which has to be elucidated (explanandum) whereas the former captures one possible theoretical position for its illumination (explanans). Obviously, the latter need not imply the former. Thus, one might easily explain the very same effects in different theories, e.g. without necessarily referring to the construct of intrinsic motivation. For analytical purposes the following denotations will be employed throughout the paper: the term postreward performance lessening will refer only to the empirical phenomenon, whereas crowding-out will be solely related to the construct of intrinsic motivation, its underlying theory or conceptualizations thereof. Existing definitions of intrinsic motivation vary considerably among scholars and are usually very loose, if not utterly vague (Thierry, 1990, p. 67). Commonly, people are said to act intrinsically motivated if they value activities for their own sake, such as they perform them without being externally induced. Reviewing the literature, Heckhausen (1989, pp. 456–460) identified six different conceptualizations of the construct of intrinsic motivation, which are briefly summarized below. 1. Internal drive without aiming at drive reduction. In accordance with the first qualification, intrinsic motivation originates from an internal drive without explicitly seeking drive reduction. Hence, intrinsic motivation differs from drives that regulate homeostatic crisis of the organism like hunger, thirst, or pain avoidance (Koch, 1956; Woodworth, 1918). 2. Activities without any aim or goal. The second qualification involves all activities which are autotelic, like leisure-time pursuit or children’s play (Klinger, 1971). 3. Optimal operation level. According to the third qualification, intrinsic motivation is seen as behavior which aims at reinstalling the maintenance of an optimal operation level, such as arousal (Hebb, 1955), incongruence of informational input (Berlyne, 1960) or adaptation (Helson, 1964). 4. Personal causation and evaluation theory. The fourth concept defines intrinsically motivated behavior as behavior a person chooses in

280

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

order to feel competent and self-determining (Deci, 1971). 5. Enjoyment or flow. According to the fifth concept, intrinsically motivated behavior is cheerful enjoyment while performing a task, or absolute absorption in an activity. It is ‘a holistic sensation people feel when they act in total involvement’ (Csikszentmihalyi, 2000, p. 10). 6. Endogenity of goal and activity. The last qualification refers to the individual’s subjective perception. Therefore, intrinsic motivation is at stake as long as there is inherent congruity between means (action) and goals (behavioral results) of activities, as far as the individual attribution process is concerned (Heckhausen, 1977; Kruglanski, 1975). A core element of the construct is the attribution of behavioral sources towards internal versus external factors, thereby differentiating between internal versus external motivators. Apart from this rather abstract definition, no general distinction between intrinsic and extrinsic motivation has been elaborated. Nevertheless, almost all conceptualizations of intrinsic motivation share at least one of the three following qualifications: 1. Internal motivation: intrinsic motivation stems from a drive from within the person, e.g. no outside incentives force the individual to act the way he does. In contrast, extrinsically motivated behavior is supposed to be induced through forces outside of the person. 2. Lacking instrumentality: intrinsically motivated behavior aims only at its own fruition, whereas extrinsic motivation implies an instrumental relationship between behavioral results and desired outcomes. 3. Attached feelings: intrinsic motivation consists of the feelings which are attached to or result from the performance of certain activities, so that intrinsically motivated individuals experience enjoyment or satisfaction during the performance of these activities. As shown by different conceptions, researchers clearly do not have the same phenomenon in mind when referring to intrinsic motivation (Bandura, 1986; Dyer & Parker, 1975; Heckhausen, 1989).

Often, different qualifications are used interchangeably (Bandura, 1986). In order to avoid any confusion among scholars and in order to build a valuable foundation for further discussion, we turn to the description of the two predominant theories in this area, cognitive evaluation theory and overjustification research. 2.2. Personal causation and cognitive evaluation theory According to Deci’s cognitive evaluation theory (CET), human beings strive to fulfil two basic needs: the need for competence and the need for self-determination. Intrinsically motivated behavior can therefore be defined as behavior a person chooses in order to feel competent and self-determining (Deci, 1975, p. 61). Concerning the influence of external rewards on intrinsic motivation, Deci formulated three propositions which summarize the early theory (Deci, 1975, p. 139–142): 1. The first process by which intrinsically motivated behavior can be affected is a change in perceived locus of causality from internal to external. As a consequence intrinsic motivation can diminish when someone receives extrinsic rewards for engaging in intrinsically motivated activities. 2. The second process focuses on the person’s emotions as experienced feelings. Intrinsically motivated behavior increases if a person’s feelings of competence and selfdetermination are enhanced; otherwise, if these feelings are diminished, intrinsic motivation will abate. 3. The third process differentiates between two aspects of every reward, a controlling and an informational one. Since rewards usually are administered contingent upon the achievement of some predefined goals, the controlling aspect restricts a person’s action space, thereby separating wanted from unwanted behavior. The informational aspect provides the recipient with information about his competence and self-determination. The relative salience of both aspects determines whether, and if so to what extent, intrinsic

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

motivation will be affected. If the informational aspect prevails, feelings of competence and self-determination will be initiated, whereas if the controlling aspect is more salient, the change in locus of causality described above will be induced. Recent research on intrinsic motivation shows that although intrinsic and extrinsic motivation were initially conceived as a dichotomy, nowadays many researchers no longer hold to this notion (Harter & Jackson, 1992; Rigby, Deci, Patrick, & Ryan 1992; Sansone & Harackiewicz, 2000b). Several studies showed that extrinsic rewards enhance intrinsic motivation (Ryan, 1982; Ryan, Mims, & Koestner, 1983). Therefore, it became increasingly clear that extrinsic rewards need not necessarily be detrimental, but rather can play a complementary role to intrinsic motivation. As a consequence, Deci and his colleagues clearly dissociated themselves from the former misleading dichotomy of intrinsic and extrinsic motivation (Rigby et al., 1992, p. 167). They argue that a person’s relative autonomy of behavior is a better way of describing the motivational basis of one’s performance than the old dichotomy was. In their modified concept, they assume that individuals tend to internalize and integrate the regulations of extrinsically motivated behavior in order to cope effectively with their social world. Therefore, the traditional dichotomy is replaced by a conceptualization of extrinsic motivation as a continuum of internalization and integration. Although extrinsically motivated behavior remains instrumental by definition, external regulations may be perceived as competence-enhancing and self-determining when integrated. Thus, contrary to former theorizing, extrinsic incentives need not be detrimental to intrinsic motivation per se. Rather, they represent opportunities to enhance intrinsic motivation, when administered correctly (Rigby et al., 1992, p. 169–171). 2.3. Self-perception theory and the overjustification effect Following a more cognitive approach, some attribution theorists account for postreward

281

performance lessening on the basis of ‘overjustification of one’s behavior’ (Lepper & Greene, 1978; Lepper, Greene, & Nisbett, 1973). The theoretical background of overjustification research is rooted in the theories of self-perception (Bem, 1967a, 1967b) and self-attribution processes (Kelley, 1967, 1973). According to this paradigm people have no direct knowledge of their motives, but infer them from their actions and contingencies. People are assumed to engage in postbehavioral inferences about their motivational states in order to self-justify or explain their behavior (Bem, 1972; Kelley, 1973). To the extent that external contingencies of reinforcement are salient, unambiguous and psychologically sufficient to explain one’s actions, such self-justification processes are hypothesized to attribute behavior to these external constraints. On the other hand, if external contingencies of reinforcement appear to be absent, or are perceived as weak, unclear or psychologically insufficient, a person is supposed to attribute his behavior to personal dispositions, interest, or desires (Bem, 1972; Lepper & Greene, 1978). As a theoretical proposition the overjustification effect states that if rewards are administered for performing an activity which is already rewarding in itself, people will attribute their behavior to these rewards. As a consequence, if these rewards are withdrawn, performance will ceteris paribus decrease to a level that is lower than the original level, before rewards were administered (Lepper & Greene, 1978; Lepper et al., 1973). The self-perception perspective explains this effect in terms of cognitive reevaluation of an activity’s intrinsic characteristics due to the mis-attribution of one’s causes, leading to changed expectations and attitudes towards the activity. According to the theory, mis-attributions are triggered by the disclosure of an external instrumentality between an activity and salient reinforcement contingencies. Before rewards are administered, a subject attributes his behavior to his own interests, inferring that he likes the activity or that he believes in it. Thus, the imposition of rewards for an activity one would have executed anyway ‘overjustifies’ one’s behavior. Additionally, one’s attitude towards the activity is likely to change.

282

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

Since behavior is now attributed to external rewards, the activity’s intrinsic characteristics were cognitively reevaluated. The subject concludes that since he was performing the activity for some external reward, it must not be as interesting or enjoyable as he thought it would be. The overjustification hypothesis predicts negative after-effects of rewards only if all additional factors besides the imposition and removal of rewards are held constant. Therefore, attribution theorists emphasize that when dealing with applied reward programs, ‘there is nothing in the present line of reasoning or in the present data to suggest that contracting to engage in an activity for an extrinsic reward will always, or even usually, result in a decrement in intrinsic interest in the activity’ (Lepper & Greene, 1976, p. 33). Conversely, it is pointed out that external contingencies of reinforcement will prove to be beneficial: (1) if people expect continued reinforcement in the same or in similar situations; (2) if rewards convey information about ability or competence, leading a person to believe that he is personally responsible for his behavior; or (3) if rewards promote the acquisition of new skills, which are necessary for the perception of intrinsic interest in that activity. Attribution theorists emphasize that the heuristic value of the overjustification perspective does not lie in discouraging the use of rewards systems to promote socially desirable behavior. Rather, it is pointed out that instead of exclusively focusing on immediate effects of incentive schemes, more attention should be directed to the possibility of harmful long-term effects of such systems, particularly in settings where subsequent extrinsic contingencies are likely to be absent (Lepper & Greene, 1976).

3. Theoretical criticism of the construct of intrinsic motivation Despite its popularity, the concept of intrinsic motivation is heavily disputed in the literature; positions tend from (gradual) acceptance to total rejection (for overviews of shortcomings see Bandura, 1986; Bates, 1979; Bernstein, 1990; Dickinson, 1989; Eisenberger & Cameron, 1996; Flora,

1990; Morgan, 1984; Notz, 1975; Thierry, 1990). Most criticism focuses on the different conceptualizations of intrinsic motivation and their implications, without disputing that postreward performance lessening might occur under special conditions. Since empirical evidence has not been considered yet, we first turn to theoretical criticism of the concept. The psychological literature summarizes three major shortcomings of the concept: 1. Confusing terminology: Detractors point out that the concept lacks an unequivocal terminology and a consistent theoretical foundation (see among others Bandura, 1986; Dickinson, 1989; Dyer & Parker, 1975; Heckhausen, 1989; Thierry, 1990). Since no precise distinction between intrinsic and extrinsic motivation has been elaborated, the term inevitably lumps together quite distinct issues, like internal drives (motivation), one’s behavior (being performed for its own sake), behavior of others (recognition) or one’s feelings (satisfaction). Besides the jumbled terminology, the concept also involves some conceptual confusion. The distinction between intrinsic and extrinsic motivations has convinced many people that motivation either originates from within or from outside a person, thereby implying the well-known person-situation subject area, although phrased in terms of a contrast. This view of motivation may be deceptive, because it diverts theoretical reasoning from the complex interaction between personal and situational characteristics as they jointly—and not mutually exclusively determine human motivation (Thierry, 1990). Consequently, it is pointed out that a great deal of conceptual confusion still surrounds the notion of intrinsic motivation (Dickinson, 1989; Notz, 1975); designations are said to be ‘shrouded in ambiguity’ (Bandura, 1986, p. 241) or to be ‘equivocal at best’ (Thierry, 1990, p. 69). 2. Major mechanisms are unclear: The precise mechanisms through which intrinsic and extrinsic motivation interact are still widely unclear (Deci & Ryan, 1985). Newer research even seems to indicate that these interactions

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

might be much more complicated than previously supposed; multi-goal perspectives (Pinrich & Garcia, 1991; Wentzel, 1991), the question whether and under what conditions intrinsic motivation is a trait (Harter & Jackson, 1992), whether it is stable or undergoes continuing changes (Sansone & Morgan, 1992; Sansone, Weir, Harpster & Morgan, 1992) raise more questions than answers. 3. Circularity: Whether a factor is perceived as intrinsic or extrinsic is highly subjective, depending on the individual attribution process. Unfortunately, the theories are of very little help in differentiating under what precise conditions intrinsic or extrinsic motivation are supposedly at stake. The same confusion prevails for the crowding-out effect, regardless of whether cognitive evaluation theory or overjustification research is concerned. According to cognitive evaluation theory, incentives diminish intrinsic motivation when they appear to be controlling, but increase motivation when conveying information about competence. It is criticized that since the theory fails to provide objective criteria for classifying rewards as controlling or informative, empirical and theoretical research in this area rests on intuitive and post hoc classifications of incentive systems (Bandura, 1986, p. 244). In addition, since performance-improving rewards have been redefined as being informative, ‘the circularity of the concepts strips the theory of any predictive value’ (Bandura, 1986, p. 245). A similar problem crops up with the attributional explanation of negative after-effects in terms of overjustification research. Since people have many different possibilities to justify their behavior, it is almost impossible to predict reliably what the target of sufficiency of justification may be. If, for instance, monetary rewards are absent or weak, people are not inevitably forced to attribute their behavior to the interestingness of their task. They may as well think of themselves as being motivated by the social contacts their work provides or by the goals the organization strives for (Pearce, 1983, 1987; Lepper &

283

Greene, 1978). Consequently, it is argued that there is not much prediction in the intrinsic concept since almost anything goes. After reviewing the theoretical concept and its shortcomings, we next turn to its empirical evidence and the effects of postreward performance lessening.

4. Experimental evidence 4.1. Experimental designs, early results and criticism During the 1970s, many experiments were conducted in order to find empirical evidence for the crowding-out effect (for overviews and further references see Deci & Ryan, 1985; Lepper & Greene, 1978; Sansone & Harackiewicz, 2000a). Experimental designs consisted in most cases of three phases: (1) subjects, mostly children or students, were observed while performing a rather attractive task for which no incentive but the activity itself had been apparent; (2) in a second phase external rewards were given for the performance of the same activity; (3) in a third phase, rewards were withdrawn and the subjects were offered a free-choice situation, e.g. they could continue performing the same activity without being rewarded externally or they could do whatever they liked. Different indicators were employed during the third phase in order to measure intrinsic motivation, like (1) resumption or rejection of the activity, (2) amount of time spent, (3) subject ratings of enjoyment, or (4) selfreports of the participants. These data were compared with the results of a control group, which had not been influenced by external interventions during the second phase. In particular research focused on different antecedents of the effect, like (1) monetary incentives vs. verbal or symbolical incentives, (2) contingent vs. non-contingent payment, (3) expected vs. non-expected payment or (4) salience vs. non-salience of the reward. The first results seemed to indicate a preponderance of empirical evidence for postreward performance lessening. Nevertheless, its theoretical

284

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

explanation through the construct of intrinsic motivation was and still is disputed (see among others Bandura, 1986; Boal & Cummings, 1981; Heckhausen, 1989; Locke & Henne, 1986; Thierry, 1990). Critics mainly accept the existence of externally influenced conditions which might be deleterious to performance, but they doubt whether diminution of intrinsic motivation is the proper concept to account for this effect. In order to clarify this point of view the major criticism is briefly summarized. 1. Lacking internal validity, insufficient operationalization of intrinsic motivation: Many researchers claim that cognitive evaluation theory as a whole has not been tested in a single experiment (Bandura, 1986; Boal & Cummings, 1981; Locke & Henne, 1986). First, distinctions between the needs for competence and self-determination and the sheer enjoyment of an activity have rarely been made (Locke & Henne, 1986). Secondly and more seriously, it is advocated that for testing the theory, intrinsic motivation has been insufficiently operationalized for empirical purposes. It is argued that in order to test the theory, drives have to be measured independently of the behavior they supposedly activate. However, most empirical work asserts that intrinsic motivation ‘must’ be at stake when no (observable) external incentives are apparent. Therefore, intrinsic motivation is measured in terms of performance variations after withdrawal of extrinsic rewards. This leads to severe shortcomings, since ‘if variations in task performance are taken as evidence of variations in strength of a competence drive, the circularity strips the theory of predictive value. Unless the strength of a drive is measured separately from its postulated effects the functional properties ascribed to it are not empirically verifiable’ (Bandura, 1986, p. 245). 2. Lacking external validity, lacking generalizability: Extensive reviews of experimental findings concerning postreward performance lessening revealed that many studies suffer from severe methodological inadequacies

and are therefore flawed (for critical overviews see Bandura, 1986; Bates, 1979; Dickinson, 1989). In addition, addressing the external validity of the concept, three major arguments are pointed out. First, the experimental environments which characterize the research cannot be equated with the scope they purport to investigate, no matter whether classrooms (Bates, 1979) or work environments (Staw, 1989) are concerned. Secondly, additional criticism focuses on subject selection, stressing that children of three and four years are hardly capable of accounting for their actions in the complex way attribution theory predicts (Morgan, 1983; Sandelands, Ashford, & Dutton, 1983). Thirdly, it is pointed out that little effort has been made to investigate long-term effects of extrinsic rewards (Bandura, 1986; Bates, 1979). Laboratory experiments and field studies indicate that intrinsic motivation is not static but seems to vary between individuals, activities, surrounding contexts and time (Sansone & Morgan, 1992; Sansone et al., 1992). Currently not a single study exists reliably proving that undermining effects are transitory for adults (Deci, Ryan, & Koestner, 1999a, p. 650). Some studies measuring task interest at several points in time revealed that postreward detrimental effects do not persist over time; subjects showed as much or even more interest in the previously rewarded activity (Feingold & Mahoney, 1975; Sagotsky & Lewis, 1978; Vasta & Stirpe, 1979). Consequently, critics doubt whether postreward performance lessening is an enduring effect, and stress that any generalizations of experimental findings to real-life situations and adult behavior stand on tenuous ground. 3. Different interpretation of experimental results: Detractors often point out that rewards can decrease later performance through a variety of processes without necessarily transforming motivational systems. Therefore, experimental findings can be interpreted differently through other psychological theories and need not refer to intrinsic motivation (Bandura, 1986; Eisenberger &

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

Cameron, 1996; Locke & Henne, 1986; Staw, 1989). For instance, it is suggested that detrimental effects may be caused by satiation or tedium (Bandura, 1986), learned helplessness (Eisenberger & Cameron, 1996) or by violating a situational norm of no payment (Bandura, 1986; Staw, 1989; Staw, Calder, & Hess, 1975). Conflicting explanations through other theories have profound implications for both theory and practice. In the case of learned helplessness for instance, external rewards would improve rather than damage postreward performance. In this case, diminution of intrinsic motivation would be inappropriate to explain the problem at hand. 4. Contradictory findings: Moreover, much experimental evidence exists which does not fit the theory. Newer empirical work points out that the style and the language with which external regulations are administered may influence their effects to quite an extent (Ryan et al., 1983); it is found that autonomy supportive social contexts do enhance intrinsic motivation (Grolnick & Ryan, 1989; Williams, 1991). Given the diverse findings reported in the literature, the empirical evidence for the crowding-out effect proves to be messier than previously supposed. Results and their interpretation are contentious, and overall no general commitment exists about the relevance of the bulk of empirical work. In view of this unsatisfactory situation some psychologists have already signaled serious skepticism: ‘If results continue to indicate that slight changes in methodology dramatically influence performance or subjective judgments of task interest, and that different measures of the same variable respond differently to the same experimental treatments, then one may question both the practical value and the psychological validity of the construct of intrinsic motivation’ (Bates, 1979, p. 574). 4.2. Meta-analyses Since experimental findings vary considerably between studies, one might gain additional insight

285

when conducting meta-analyses. Meta-analyses provide a methodology that allows quantifying effect sizes and combining them across different experiments in order to test whether certain hypotheses constitute behavioral regularities which persist reliably over all studies included in the analysis. Such an approach seems particularly appealing since it allows drawing conclusions about the statistical significance of undermining effects of distinct reward categories across different studies. Nevertheless, the limits of this approach are that results vary dramatically with the researchers’ selection of relevant experiments, as well as with their definition of theoretically meaningful cells (Lepper, Henderlong, & Gingras, 1999). Besides, results must be interpreted with caution because of an expected publication bias in favor of finding detrimental effects.1 A detailed review of all existing meta-analyses is beyond the scope of this article. We therefore turn to the findings of the two newest and most elaborated studies by Eisenberger and Cameron (1996) and Deci et al. (1999a), hereafter denoted EC-96 and DRK-99, respectively. Both analyses assess the subjects’ intrinsic motivation in terms of their selfreported interest in the activity (self-reported interest) as well as the time they spent on the activity during a free choice period subsequent to the experimental phase (free choice behavior). The EC-96 meta-analysis covers 61 studies on free choice behavior and 64 studies on self-reported task interest published in the period 1971 to September 1991, whereas the DRK-99 analysis focuses on 101 experiments on free choice behavior and 84 studies on self-reported task interest from 1971 to August 1997. In contrast to the EC-96 study, the DRK-99 analysis includes unpublished doctoral dissertations. Subjects in both studies are pre-school children and college students. Figs. 1 and 2 show the summary statistics for both meta-analyses. Analyses are arranged hierachically, desegregating the general classification of extrinsic incentives into more specific types of reward. Definitions of the reward categories are given in Table 1. For each analysis, effects-sizes measured in terms of standard deviation units 1

We are grateful to one referee for pointing this out.

286

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

Fig. 1. Summary statistics of the Eisenberger and Cameron (1996) meta-analysis.

were accumulated across individual studies and weighted by the number of participants in each group. The pair of numbers in parentheses refers to the 95%-confidence interval for the composite effect-size d. Since the mean of the control group was subtracted from the mean of the reward group, a negative d-value reflects an undermining effect, while a positive d-value shows an enhancement effect. Statistically reliable effects are given in boldface with an asterisk next to a d-value. The symbol k refers to the number of studies, (i.e. to the number of effect-sizes in each composite effectsize) while outliers and studies with insufficient information are excluded. Of greatest interest are those rewards that are in accord with the agency perspective. As shown, agency theory generally advocates the use of payfor-performance systems, with the exceptions of (severe) risk considerations, limited liability and the case of sufficiently noisy indicators about the agent’s performance. Since experimental designs do not fit one of these exceptions, our main interest lies in those rewards categories that induce performance-contingent incentive effects, i.e. quality-dependent rewards in the EC-96 analysis and

performance-contingent rewards in the DRK-99 study, respectively. The summary statistics of the EC-96 analysis reveal that undermining effects due to quality-dependent rewards prove to be statistically insignificant on the measure of free choice behavior. Quality-dependent rewards even induce a reliable increasing (!) effect on the measure of self-reported interest. In contrast to these findings, Deci et al., (1999a) report a reliable undermining effect of performance-contingent rewards on the measure of free choice behavior. If those studies that are methodologically questionable due to the lack of adequate feedback control, are excluded, the effect proves to be supported by only 10 experiments. Like Eisenberger and Cameron, (1996), Deci et al. (1999a) fail to provide statistical significance for an undermining effect on the measure of self-reported interest. Thus, restricting attention to economically meaningful rewards, (i.e. rewards that induce an incentive effect), the EC-96 analysis clearly contradicts CET, whereas the theory is only partially supported by the DRK-99 study. Supporters of the intrinsic concept heavily criticized the EC-96 study because of methodological

287

Fig. 2. Summary statistics of the Deci et al. (1999a) meta-analysis.

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

288

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

Table 1 Reward categories Reward category Eisenberger and Cameron (1996) Verbal Tangible Unexpected Expected Quality-dependent Completion dependent Performance independent

Deci et al. (1999a) Verbal Tangible Unexpected Expected Task-noncontingent Engagement-contingent Completion-contingent Performance-contingent Positive/negative feedback control No-feedback control Maximum reward Not maximum reward

Description Rewards are given only in terms of positive feedback. Rewards are tangible, i.e. they are nonverbal. Rewards are not announced beforehand, i.e. subjects are performing the activity without any expectation of reward. Rewards are announced beforehand, i.e. subjects are performing the activity with the expectation of being rewarded afterwards. Rewards are contingent on the achievement or surpassing of a pre-defined quality standard. Rewards are administerd upon the completion of a task with no regard to how the activity is performed. Rewards are administered for simply taking part in the activity, no matter whether the activity is completed or whether the pre-defined performance standards are met.

Rewards are only given in terms of positive feedback. Rewards are tangible, i.e. they are nonverbal. Rewards are not announced beforehand, i.e. subjects performe a target activity without any expectation of being rewarded afterwards. Rewards are announced beforehand, i.e. subjects performe the activity with the expectation of being rewarded afterwards. Rewards are administered for something other than engaging in the target activity, such as simply participating in the study. Rewards are administered for simply engaging in the target activity, without the requirement that the activity has to be completed. Rewards are administered upon the completion of the target activity with no regard to how it is performed. Rewards are administered contingent upon the achievement or surpassing of a pre-defined quality standard, set for the target activity. The control group also gets the positive/negative feedback conveyed by the reward. The control group does not get the feedback conveyed by the reward. All participants receive the maximum reward. Participants that perform less than optimal, receive less than the maximum reward.

issues like study selection and cell definition. They pointed out that the DRK-99 analysis is more accurate in mapping CET (Deci, Ryan, & Koestner, 1999b; Lepper, Henderlong, & Gingras, 1999). However, even this analysis provides only weak support for undermining effects when referring to economically meaningful incentive systems. Its failure to prove detrimental effects on the selfreport measure, despite an expected publication bias in favor of finding such effects, evidently calls their relevance into question. Deci et al., (1999a, p. 655f.) try to explain this fact in that subjects might have been confusing their interest in the activity with their enjoyment of receiving a reward. Such

rationalizations are not very convincing. Particularly not since the same authors have suggested elsewhere that the best way to assess intrinsic motivation is to measure both free choice behavior and self-reported interest and to consider them as intrinsic motivation only when they are correlated within conditions and studies (Ryan, Koestner, & Deci, 1991). Another result of the DRK-99 study is of greatest interest. The analysis found considerable age effects showing that tangible rewards are more detrimental for children than for college students on both measures of intrinsic motivation (Deci et al., 1999a, p. 656). This set of findings has never

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

been predicted before. Possible explanations are that college students, having greater cognitive capacity and being more accustomed to the use of reward, have more expectations about whether rewards are appropriate, and are thus more likely to interpret rewards as indicators of their effective performance than as controllers of their behavior (Deci et al., 1999a, p. 656). This account provides a strong rationale for limitations of detrimental effects due to the subjects’ cognitive capacity and socialization while dealing with rewards, indicating that from an empirical point of view any generalization about undermining effects in corporations may have to be made with greatest caution. Summarizing, the empirical evidence on postreward performance lessening, even in the metaanalysis that provides the strongest support for CET, remains scant, when it comes to performance contingent rewards. Relying on only 10 studies concerning the measure of free choice behavior, it fails to be supported by the self-reported interest measure. Additional restrictions are given by the supposition that the subjects’ maturity may play an important role in determining whether postreward performance lessening will happen or not. 4.3. Newer experiments on pay for performance and intrinsic motivation Three new studies by Eisenberger, Rhoades and Cameron (1999), none of them included in the above meta-analyses, provide a strong rationale for the assumption that previous experiments may have been addressing the role of incentives inadequately. In everyday life, the question may not be whether rewards diminish intrinsic motivation, but rather whether their administration counters any motivational loss already produced by the heteronomous imposition of tasks and performance objectives (Eisenberger, Rhoades, & Cameron, 1999, p. 1027). People in everyday life learn that rewards are based on utilitarian considerations, being believed necessary to induce desired behavior. Precisely because this is so, reward may not convey social control, as suggested by CET, but rather freedom of action. In view of Eisenberger et al., (1999, p. 1027) the promise or repeated use of performance-contingent

289

reward not only conveys that (1) the person, group or organization giving the reward lacks control over the performance of the potential reward recipient, but additionally that (2) the potential recipient can, if he or she so wishes, decline the reward and not act as requested. Thus, the expectation of performance-contingent reward may increase experienced self-determination, thereby enhancing rather than reducing intrinsic motivation. Before, effects of performance-contingent reward on perceived autonomy have not been investigated in one published study (Eisenberger et al., 1999, p. 1026). Testing their hypothesis in three studies, the authors found strong empirical support for it, no matter whether subjects were students performing a novel task or employees engaged in their daily work activities. Results indicate that performance-contingent rewards increase students’ subsequent expression of task enjoyment and free time spent performing the task. Perceived self-determination reliably mediated the effect of reward on task enjoyment, but not on free task behavior (Eisenberger et al., 1999, pp. 1029–1031; study 1). High performancereward expectancy reliably contributed to employees’ perceived self-determination. Employees who perceived increased autonomy owing to performance reward expectancies, experienced more organizational support, a more positive mood at work, and also showed superior work performance (Eisenberger et al., 1999, p. 1031– 1034; study 2). The incremental relationship found between employees’ expected reward for superior performance and expressed work interest was reliably greater for employees with a high or medium desire for control. Expectation of reward for high performance reliably lead to greater perceived selfdetermination, attracting the interest of individuals who had a high desire for control (Eisenberger et al., 1999, p. 1034–1038; study 3). Since performance-contingent reward increased perceived autonomy, intrinsic motivation, and related outcomes in all three studies, their findings contradict the assumption that performance-contingent reward constitutes an invasive form of social control. Practical implications of these results are that pay for high performance may constitute an adequate means to increase the

290

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

intrinsic motivation of persons who have been assigned performance objectives. These results are in accord with previous findings demonstrating incremental effects due to a symbolic function of tangible rewards received for high performance (Bandura 1997; Harackiewicz, Abrahams, & Wageman, 1987; Sansone & Harackiewicz, 2000b). Consequently, the authors suggest replacing CET’s presumption of an inverse relationship between reward and perceived autonomy by the assumption that reward for high performance may increase perceived self-determination (Eisenberger et al., 1999, p. 1039).

5. Implications for agency theory After elaborating both a theoretical and empirical analysis, we now want to draw some conclusions. Restricting attention to performance pay in business corporations, the starting point of this paper was to clarify whether the agency perspective faces an urgent need to incorporate the construct of intrinsic motivation. Incorporation may be worthwhile if undermining effects prove to be a reliable empirical phenomenon, whose theoretical explanation by the construct of intrinsic motivation helps understanding and modelling these effects. The empirical analysis reveals good and bad news for agency theory. Obviously, since many experiments succeeded in documenting detrimental effects, it is bad news for agency theory that ‘hidden costs of reward’ do indeed exist. Nevertheless, it is good news that the empirical evidence on undermining effects cannot be interpreted as being contradictory to agency theory. First, antecedents for postreward performance lessening are highly stylized and easily avoidable in real life situations. Secondly, recent research has found conditions under which reward will have positive effects on subsequent behavior, even if the deleterious antecedents were prevailing. Thirdly, the way agency theory recommends designing incentive schemes appears to be one of these conditions. All three arguments are briefly discussed below. Inspection of the circumstances under which detrimental effects of reward became demon-

strable in the laboratory reveals that these conditions are highly specialized. Undermining effects could only be demonstrated when all of the following conditions were met: 1. High level of initial task interest: The activity performed by the subjects needed to be highly attractive in itself, so much so that individuals possessed sufficient preexisting interest in performing it without any expectation or promise of reward. 2. Lack of control during the undermining phase: The situation in which undermining effects were measured was free of any kind of social surveillance, demand or expectation of continued reward. 3. Exclusion of performance improvement: Performing the activity provided no possibility for performance improvement, presumably instigating feelings of satisfaction, thereby enhancing rather than reducing the subjects’ interest in the activity (Bandura, 1986; Calder & Staw, 1975). 4. Rewards were situationally inappropriate: Those experiments that succeeded in documenting detrimental effects of reward did so by rewarding subjects for activities for which one normally does not expect to be compensated for. Hence, a crucial antecedent seems to be that the provision of incentives is situationally inappropriate, indicating that undermining effects appear to be produced by the violation of a situational norm of no payment (Dickinson, 1989; Snelders & Lea, 1996; Staw, 1989). It must be emphasized that these conditions rarely prevail in real life work settings. Within business corporations, people expect to be paid in exchange for selling their labor. Thus rewards are the norm rather than being situationally inappropriate. Moreover, most jobs involve supervision and social surveillance of some kind, either by superiors, subordinates, shareholders or the public. Likewise, most positions involve possibilities for performance improvement which may lead to increased task interest. In addition, a large number of jobs are not inherently interesting enough to foster high intrinsic motivation. Some

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

jobs are even performed solely because of their extrinsic inducements. In such cases, detrimental effects due to increased extrinsic rewards must be precluded because of the person’s initial lack of a sufficiently high level of intrinsic interest. Nevertheless, even if all jobs were interesting enough to foster high intrinsic motivation, a powerful norm exists in the business world for extrinsic payment. An instrumental relationship between professional behavior and extrinsic rewards is supported by both social and legal standards; moreover, pay even seems to have become of great symbolic importance as a surrogate of worth and social status in modern business society (Pearce, 1987). Since extrinsic reinforcement is the norm, tasks are likely to be experienced as more valuable or more interesting if they lead to higher compensation, thereby tending to increase rather than diminish postreward performance. Thus, in practice the danger may not be that rewards are perceived as inappropriate, but rather that not rewarding good performance might be perceived as unfair, leading to frustration and diminution of interest and commitment towards the activity. Recent findings report on some contextual and personal patterns successfully countering negative after-effects of rewards, indicating that such effects are easily avoidable. The style and the language in which rewards are administered are such factors (Ryan et al., 1983). If rewards are administered in an autonomy supportive context, they enhance rather than diminish intrinsic motivation (Deci, Conell, & Ryan, 1989; Grolnick & Ryan, 1989; Williams, 1991). The subjects’ cognitive capacity and socialization in dealing with rewards appears to be another limiting factor. Adults, being accustomed to the use of rewards, are likely to interpret performance pay as a sign of freedom, conveying the reward-giver’s lack of control over the performance of the potential recipient. Thus, previous studies mostly involving pre-school children may have found detrimental effects mainly because children were less accustomed to the use of rewards and thus were less likely to interpret rewards as indicators of their autonomy and competence. This conjecture appears to be substantiated by the latest meta-analysis, since the

291

authors found considerable age effects showing that undermining effects were far more detrimental for pre-school children than for students (Deci et al., 1999a). Three new studies are also supportive of this view. They all found that performance pay reliably increased perceived autonomy, work interest, work mood, and work performance with adults who had been assigned performance standards (Eisenberger et al., 1999). These results are in accord with previous findings, showing that rewards can enhance intrinsic motivation due to their symbolic value (Harackiewicz et al., 1987; Sansone & Harackiewicz, 2000b). Reviewing the whole empirical evidence on undermining effects revealed that the experimental findings cannot be interpreted as being contradictory to agency theory. This is particularly so when referring to economically meaningful incentive systems. Two recent meta-analyses, condensing the whole empirical research in this area found deleterious effects only for reward categories that are generally not among the ones advocated by the agency perspective, e.g. engagement-contingent, completion-contingent and tasknoncontingent rewards. Both meta-analyses failed to prove undermining effects of performance-contingent rewards. One even found increasing effects. Thus, when considering the condensed empirical material of more than a quarter of a century, the current criticism against agency theory is not supported by empirical data. It even seems as if providing incentives according to economic theory violates the necessary conditions for postreward performance lessening, and therefore helps to supersede its detrimental effects. Summarizing, the empirical evidence on postreward performance lessening indicates that such effects constitute stylized exceptions rather than a rule. Two meta-analyses fail to prove detrimental effects when referring to performance-contingent pay. In addition, the experimental evidence showing such effects primary relied on pre-school children and college students, and not on adults in business corporations. Moreover, the scant empirical evidence we have on incentive schemes in business corporations suggests that detrimental effects of rewards are avoidable. Not only are their antecedents too artificial to prevail in real-life

292

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

work places, in addition, by making rewards contingent upon performance and administering them in an autonomy supportive context, will violate these antecedents. Thus, when considering the entire empirical material on undermining effects, at least when in comes to business corporations, we do not find much evidence substantiating any claims in favor of incorporating postreward performance lessening into economic theory. Nevertheless, the experimental evidence showed that undermining effects exist under special circumstances. One might think that incorporating intrinsic motivation into agency models may help economic theory to understand and deal with such exceptional behavioral patterns. Unfortunately, the literature review revealed that the construct of intrinsic motivation is not of much help in so doing. Overall, the construct fails to provide objective criteria for classifying rewards as being controlling, informative or expressive of some other aspect. Consequently, its main contribution lies in intuitive post-hoc classifications of reward systems, without being capable of predicting behavioral outcomes beforehand. Since major mechanisms through which postreward performance lessening occurs still remain widely unclear, the construct’s actual state of the art obviously cannot serve as a model for incorporating postreward performance lessening into the agency perspective. Agency theory would inevitably lose its predictive quality at the gain of being able to model a stylized but still largely unclear phenomenon. Thus, since the dichotomy between intrinsic and extrinsic motivation does not help in explaining and overall predicting the problem at hand, we plead for skipping this terminology and for the pursuit of more promising paths in research. What are these new paths? Obviously, answers have to vary across disciplines. As shown, most psychologists have already recognized the disadvantages of polarizing between intrinsic and extrinsic motivation and are currently focusing on the complex interaction between both personal and situational variables. As far as economics is concerned, the need to develop the theory in order to incorporate additional behavioral aspects has long been understood. Consequently, the theory has noticeably matured in the past years. Overall,

incorporation of behavioral effects similar to performance lessening are increasingly being taken into account by the theory of implicit contracts (MacLeod, 1995), research on self-perception and social interactions (Benabou & Tirole, 2000), fairness and reciprocity (Falk & Fischbacher, 2000; Fehr & Schmidt, 1999; Rabin, 1993), social norms (Fudenberg & Levine, 1997; Kandori, 1992; Knack, 1992) and the analyses of fuzzy incentives in organizations (Bernheim, 1994; Kreps, 1997). Thus, although the standard agency model does not account for non-monetary behavioral aspects per se, game theoretic variations thereof increasingly do. Consequently, as fas as economics is concerned, we argue in favor of skipping any reasoning about incorporating intrinsic motivation into agency models, and think it more worthwhile to continue along the promising paths where economic theory is presently heading. Nonetheless, we by no means question the importance of psychological insight for economic theory and real life situations, nor do we think that the agency model is perfectly elaborated. Owing to cognitive evaluation theory, we learned about possibly deleterious consequences of autonomy restrictions and negative feedback on human motivation. We merely argue that the current empirical material on undermining effects does not indicate that performance pay in business corporations will always or even usually be interpreted that way. Consequently, neither the phenomenon of postreward performance lessening nor its theoretical explanation by the construct of intrinsic motivation challenge economic theory to the extent many contributions have lead us to believe.

Acknowledgements We are grateful to Armin Falk, Ernst Fehr, Bruno S. Frey, Simon Gaechter, Anthony Hopwood (the editor), Margit Osterloh, and two anonymous referees for helpful comments and discussions. A. H. Kunz gratefully acknowledges financial aid by research grant FHA-99-U-049-4040.

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

References Alchian, A. A., & Demesetz, H. (1972). Production, information costs, and economic organization. American Economic Review, 62, 777–795. Baiman, S. (1990). Agency research in managerial accounting:a second look. Accounting, Organizations and Society, 15, 341– 371. Baker, G. (1992). Incentive contracts and performance measurement. Journal of Political Economy, 100, 598–614. Bandura, A. (1986). Social foundations of thought and action. A social cognitive theory. New Jersey: Prentice Hall. Bandura, A. (1997). Self-efficacy: the exercise of control. New York: Freeman and Co. Bates, J. A. (1979). Extrinsic reward and intrinsic motivation: a review with implications for the classroom. Review of Educational Research, 49, 557–576. Beatty, R. P., & Zajac, E. (1994). Managerial incentives, montoring, and risk bearing: a study of executive compensation, ownership, and board structure in initial public offerings. Administrative Science Quarterly, 39, 313–335. Benabou, R., & Tirole, J. (2000). Self-confidence and social interactions. NBER-Working Paper No. 7585. Bem, D. J. (1967a). Self-perception: an alternative interpretation of cognitive dissonance phenomena. Psychological Review, 74, 183–200. Bem, D. J. (1867b). Self-perception: the dependent variable of human performance. Organizational Behavior and Human Performance, 2, 105–121. Bem, D. J. (1972). Self perception theory. In Berkowitz, L. (Ed.). Advances in experimental social psychology (pp. 1–62). New York: Academic Press. Benjamin, A. C. (1937), An introduction to the philosophy of science, New York: Thatas. Berlyne, D. E. (1960). Conflict, arousal, and curiosity, New York: McGraw-Hill. Bernheim, B. D. (1994). A theory of conformity. Journal of Political Economy, 102, 841–877. Bernstein, D. J. (1990). Of carrots and sticks: a review of Deci and Ryan’s intrinsic motivation and self-determination in human behavior. Journal of Experimental Analysis of Behavior, 54, 323–332. Boal, K. B., & Cummings, L. L. (1981). Cognitive evaluation theory: an experimental test of processes and outcomes. Organizational Behavior and Human Performance, 28, 289–310. Brittan, S. (1997). Work as its own reward. Financial Times, 16 Monday April 7th. Calder, B. J., & Staw, B. M. (1975). Self-perception of intrinsic and extrinsic motivation. Journal of Personality and Social Psychology, 31, 599–605. Camerer, C., Babcock, L., Loewenstein, G., & Thaler, R. (1997). Labor supply of New York City cabdrivers: one day at a time. Quarterly Journal of Economics, 112, 407–441. Conlisk, J. (1989). Three variants on the allais example. American Economic Review, 79, 392–407. Csikszentmihalyi, M. (2000). Beyond boredom and anxiety, San Francisco: Jossey-Bass.

293

Davis, D. D., & Holt, C. A. (1993). Experimental economics. Princeton: University Press. Deci, E. L. (1971). Effects of externally mediated rewards on intrinsic motivation. Journal of Personality and Social Psychology, 18, 105–115. Deci, E. L. (1975). Intrinsic motivation, New York and London: Plenum. Deci, E. L., Conell, J. P., & Ryan, R. M. (1989). Self-determination in a work organization. Journal of Applied Psychology, 74, 580–590. Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and selfdetermination in human behavior, New York: Plenum. Deci, E. L., Ryan, R. M., & Koestner, R. (1999a). A metaanalytic review of experiments examining the effect of extrinsic rewards on intrinsic motivation. Psychological Bulletin, 125, 627–668. Deci, E. L., Ryan, R. M., & Koestner, R. (1999b). The undermining effect is a reality after all - extrinsic rewards, task interest, and self-determination: reply to Eisenberger, Pierce, and Cameron (1999) and Lepper, Henderlong, and Gingras (1999). Psychological Bulletin, 125, 692–700. Dickinson, A. M. (1989). The detrimental effects of extrinsic reinforcement on intrinsic motivation. Behavior Analyst, 12, 1–15. Dyer, L., & Parker, D. F. (1975). Classifying outcomes in work motivation research: an examination of the intrinsic-extrinsic dichotomy. Journal of Applied Psychology, 60, 455–458. Eisenberger, R., & Cameron, J. (1996). Detrimental effects of reward. Reality or myth. American Psychologist, 51, 1153–1166. Eisenberger, R., Rhoades, L., & Cameron, J. (1999). Does pay for performance increase or decrease perceived self-determination and intrinsic motivation. Journal of Personality and Social Psychology, 77, 1026–1040. Eisenhardt, K. M. (1989). Agency theory: an assessment and a review. Academy of Management Review, 14, 57–74. Falk, A., & Fischbacher, U. (2000). A theory of reciprocy. Working Paper No. 6, Institute for Empirical Economic Research, University of Zurich. Fehr, E., & Gaechter, S. (2000a). Fairness and retaliation: the economics of reciprocity. Journal of Economic Perspectives, 13, 159–181. Fehr, E., & Gaechter, S. (2000b). Cooperation and punishment in public good experiments. American Economic Review, 90, 980–994. Fehr, E., & Schmidt, K. M. (1999). A theory of fairness, competition, and cooperation. Quarterly Journal of Economics, 114, 817–868. Feingold, B. D., & Mahoney, M. J. (1975). Reinforcement effects in intrinsic interest: undermining the overjustification hypothesis. Behavior Therapy, 6, 367–377. Flora, S. R. (1990). Undermining intrinsic interest from the standpoint of a behavior analyst. Psychological Record, 40, 323–346. Frey, B. S. (1997). Not just for the money. An economic theory of personal motivation, Cheltenham and Brookfield: Elgar. Fudenberg, D., & Levine, D. K. (1997). Learning and evolution in games. Boston: MIT Press.

294

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295

Gibbons, R. (1998). Incentives in organizations. Journal of Economic Perspectives, 12, 115–132. Gneezy, U., & Rustichini, A. (2000). A fine is a price. Journal of Legal Studies, 29, 1–18. Grolnick, W. S., & Ryan, R. M. (1989). Parent styles associated with children’s self-regulation and competence in school. Journal of Educational Psychology, 81, 143–154. Harackiewicz, J. M., Abrahams, S., & Wageman, R. (1987). Performance evaluation and intrinsic motivation: the effects of evaluative focus, rewards, and achievement orientation. Journal of Personality and Social Psychology, 53, 1015–1023. Harter, S., & Jackson, B. J. (1992). Trait vs. nontrait conceptualizations of intrinsic/extrinsic motivational orientation. Motivation and Emotion, 16, 209–230. Haubrich, J. (1994). Risk aversion, performance pay, and the principal-agent problem. Journal of Political Economy, 102, 258–276. Hebb, D. O. (1955). Drives and the C.N.S. Psychological Review, 62, 243–254. Heckhausen, H. (1977). Achievement motivation and its constructs: a cognitive model. Motivation and Emotion, 1, 283– 329. Heckhausen, H. (1989). Motivation und handeln, Heidelberg: Springer. Helson, H. (1964). Adaption-level theory. New York: Harper & Co. Holmstrom, B. (1979). Moral hazard and observability. The Bell Journal of Economics, 13, 324–340. Jensen, M. C., & Smith, C. W. (2000). Stockholder, manager, and creditor interests: application of agency theory. In: M.C. Jensen (Ed.). A theory of the firm: governance, residual claims and organizational forms. Boston: Harvard University Press. Kagel, J. H., & Roth, A. E. (Eds.) (1995). Handbook of experimental economics. Princeton: University Press. Kandori, M. (1992). Social norms and community enforcement. Review of Economic Studies, 59, 63–80. Kelley, H. (1967). The process of causal attribution. American Psychologist, 28, 107–128. Kelley, H. (1973). Attribution theory in social psychology. In Levine D. (Ed.) Nebraska symposium on motivation, 192–238, Vol. 15, Lincoln. Klinger, E. (1971). Structure and functions of fantasy. New York: Wiley. Knack, S. (1992). Civic norms, social sanctions, and vote turnout. Rationality and Society, 4, 133–156. Koch, S. (1956). Behavior as ‘intrinsically’ regulated: work notes towards a pretheory of phenomena called ‘motivational’. In Jones M.R. (Ed.), Nebraska symposium on motivation 1956, Lincoln. Kreps, D. M. (1997). The interaction between norms and economic incentives. intrinsic motivation and extrinsic incentives. American Economic Review (Papers and Proceedings), 359–364. Kruglanski, A. W. (1975). The endogeneous-exogenous partition in attribution theory. Psychological Review, 8, 274–287. Lambert, R. A. (2001). Contracting theory and accounting. Journal of Accounting and Economics, Forthcoming.

Lanen, W. N., & Larcker, D. F. (1992). Executive compensation contracts adoption in the electric utility industry. Journal of Accounting Research, 30, 70–93. Lazear, E. (2000). Performance pay and productivity. American Economic Review, 90, 1346–1361. Lepper, M. R., & Greene, D. (1978). Overjustification research and beyond: toward a means-ends analysis of intrinsic and extrinsic motivation. In M. R. Lepper, & D. Greene (Eds.), The hidden costs of reward: new perspectives on the psychology of human motivation, (pp. 109–148), Hillsdale: Erlbaum. Lepper, M. R., & Greene, D. (1976). On understanding ‘overjustification’: a reply to reiss and sushinsky. Journal of Personality and Social Psychology, 33, 25–35. Lepper, M. R., Greene, D., & Nisbett, R. E. (1973). Undermining children’s intrinsic interest with extrinsic rewards: a test of the overjustification hypothesis. Journal of Personality and Social Psychology, 28, 129–137. Lepper, M. R., Henderlong, J., & Gingras, I. (1999). Understanding the effects of extrinsic rewards on intrinsic motivation—uses and abuses of meta-analysis: comment on Deci, Koestner, and Ryan. Psychological Bulletin, 125, 669–676. Locke, E. A., & Henne, D. (1986). Work motivation theories. In C. L. Cooper, & I. Robertson (Eds.), International review of industrial and organizational psychology 1986, (pp. 1-35), New York. MacCorquadale, K., & Meehl, P. E. (1948). On a distinction between hypothetical constructs and intervening variables. Psychological Review, 55, 95–107. MacLeod, W. B. (1995). Incentives in organizations: an overview of some of the evidence and theory. In H. Siebert (Ed.), Trends in business organization:do participation and cooperation increase competitiveness? (pp. 3-42), Chicago: Kiehl Institute World Press. Mas-Colell, A.,Whinston, M. D., & Green, J. R. (1995). Microeconomic theory, New York: Oxford University Press. Milgrom, P., & Roberts, J. (1992). Economics, organization and management, London: Prentice Hall. Morgan, M. (1983). Decrements on intrinsic motivation among rewarded and observed subjects. Child Development, 54, 636– 644. Morgan, M. (1984). Reward-induced decrements and increments in intrinsic motivation. Review of Educational Research, 5–30. Murphy, K. J. (1999). Executive compensation. In O. Ashfelter, & D. Card (Eds), Handbook of labor economics, Vol. 3, North-Holland. Notz, W. W. (1975). Work motivation and the negative effects of extrinsic rewards. American Psychologist, 30, 884–891. Osterloh, M., & Frey, B. S. (2000). Motivation, knowledge transfer, and organizational forms. Organizational Science, 11, 538–550. Oyer, P. (2000). Why do firms use incentives that have no oincentive effects? Working Paper, Stanford University. Pearce, J. L. (1983). Job attitude and motivation differences between volunteers and employees from comparable organizations. Journal of Applied Psychology, 68, 646–652. Pearce, J. L. (1987). Making sense of volunteer motivation: the sufficiency of justification hypothesis. In Steers, R. M. &

A.H. Kunz, D. Pfaff / Accounting, Organizations and Society 27 (2002) 275–295 Porter L.W. (Eds.), Motivation and work behavior (pp. 545– 554), New York. Pinrich, P. R., & Garcia, T. (1991). Student goal orientation and self-regulation in the college classroom. In M. L. Maehr, & P. R. Pinrich (Eds.), Advances in motivation and achievement, (pp. 371–402), Vol. 7, Greenwich: Jai Press. Prendergast, C. (1999). The provision of incentives in firms. Journal of Economic Literature, 17, 7–63. Rabin, M. (1993). Incorporating fairness into game theory and economics. American Economic Review, 83, 1281–1302. Rigby, C. S., Deci, E. L., Patrick, B. C., & Ryan, R. M. (1992). Beyond the intrinsic-extrinsic dichotomy: self-determination in motivation and learning. Motivation and Emotion, 16, 165–185. Ryan, R. M. (1982). Control and information in the intrapersonal sphere: an extension of cognitive evaluation theory. Journal of Personality and Social Psychology, 43, 450–461. Ryan, R. M., & Deci, E. L. (2000). When rewards compete with nature: the undermining of intrinsic motivation and self-regulation. In C. Sansone, & J. Harackiewicz (Eds.). Intrinsic and extrinsic motivation. The search for optimal motivation and performance, (pp. 14–54), San Diego: Academic Press. Ryan, R. M., Mims, V., & Koestner, R. (1983). Relation of reward contingency and interpersonal context to intrinsic motivation: a review and test using cognitive evaluation theory. Journal of Personality and Social Psychology, 45, 736– 750. Ryan, R. M., Koestner, R., & Deci, E. L. (1991). Ego-involved persistence: when free-choice behavior is not intrinsically motivated. Motivation and Emotion, 15, 185–205. Sagotsky, G., & Lewis, A. (1978). Extrinsic reward, positive verbalizations, and subsequent intrinsic interest. Paper Presented at the Meeting of the American Psychological Association in Toronto. Sandelands, L. E., Ashford, S. J., & Dutton, J. E. (1983). Reconceptualizating the overjustification effects: a templatematching approach. Motivation and Emotion, 7, 229–255. Sansone, C., & Harackiewicz, J. (2000a). Intrinsic and extrinsic motivation. The search for optimal motivation and performance. San Diego: Academic Press.

295

Sansone, C., & Harackiewicz, J. (2000b). Rewarding competence: the importance of goals in the study of intrinsic motivation. In C. Sansone, & J. Harackiewicz (Eds.). Intrinsic and extrinsic motivation. The search for optimal motivation and performance. (pp. 82–103). San Diego: Academic Press. Sansone, C., & Morgan, C. (1992). Intrinsic motivation and education: competence in context. Motivation and Emotion, 16, 249–270. Sansone, C., Weir, C., Harpster, L., & Morgan, C. (1992). Once a boring task always a boring task? Interest as a selfregulating mechanism. Journal of Personality and Social Psychology, 63, 379–390. Snelders, H. M. J. J., & Lea, S. E. G. (1996). Different kinds of work, different kinds of pay: an examination of the overjustification effect. Journal of Socio-Economics, 25, 517–535. Staw, B. M. (1989). Intrinsic and extrinsic motivation. In H. Leavitt, L. Pony, & D. Boje (Eds.), Readings in managerial psychology (pp. 36–71), Chicago: University Press. Staw, B. M., Calder, B. J., & Hess, R. (1975). Intrinsic motivation and norms about payment, Working Paper, Northwestern University. Thierry, H. (1990). Intrinsic motivation reconsidered. In U. Kleineck, H.-H. Thierry, & H. Haecker (Eds.), Work motivation (pp. 67–82), Hillsdale: Erlbaum. Tosi, H. L., Katz, J. P., & Gomez-Mejia, L. R. (1997). Disaggregating the agency contract: the effects of monitoring, incentive alignement, and term in office on agent decision making. Academy of Management Journal, 40, 584–602. Vasta, R., & Stirpe, L. A. (1979). Reinforcement effects on three measures of childrens interest in math. Behavior Modification, 8, 98–108. Wentzel, K. R. (1991). Social and academic goals at school: motivation and achievement in context. In M. L. Maehr, & P. R. Pinrich (Eds.), Advances in motivation and achievement (pp. 185–212), Vol. 7, Greenwich: Jai Press. Williams, G. C. (1991). Self-determination and attrition from a weight loss program. Unpublished Manuscript, University of Rochester. Woodworth, R. S. (1918). Dynamic psychology, New York: Ayer Company Publishers.