Repositório ISCTE-IUL

: The main aim of this paper is to show that even when integrating the findings of behavioral economics, agency theory’s conception of interactions at work does not actually account for cooperative behavior. The paper draws on the distinction between the concepts of individual and person to critically examine this conception and show that, while work is mostly organized on the assumption that workers are self-interested individuals , management rhetoric addresses workers as persons in an attempt to prompt their cooperation and personal commitment. This managerial paradox may partly be due to the prevalent influence of agency theory’s prescriptions and has been contributing to a severe deterioration of the quality of working life. But it also indicates that agency theory has to confront serious theoretical and prescriptive dilemmas.


Introduction
The various paradoxes that always marked the world of work have sharpened dramatically in recent decades.Pressured by markets and shareholders to simultaneously innovate and reduce costs, management responses became ever more schizophrenic.On one hand, management rhetoric focuses on trust-building and cooperation, thereby addressing workers as persons, i.e. socially embedded beings endowed with moral capacity; on the other hand, actual management practices tighten control and individualize incentives, thereby addressing workers as individuals, i.e. non-cooperative beings predominantly driven by self-interest.This paper argues that the sharpening of this paradox might be partly explained and legitimized by the prevailing influence of agency theory.
Agency theory was launched in economics by Jensen and Meckling in 19761 .
Though the agency literature now encompasses different fields and strands, including a sociological and organizational integrative approach (Shapiro 2005, Mitnick 1992), our focus is on the agency theory of the firm 2 and more precisely on its behavioral assumptions, which have been adopted by all mainstream economic theories of the firm.In addition to its academic influence, agency theory has diffused into the business press and business schools and taught to millions of MBA students and executives around the world.Goshal (2005) and Duska (1992) contend that agency theory's assumptions have been incorporated in the worldviews of managers, thereby becoming a powerful normative model.
The building block of agency theory is a particular social interaction, namely a relationship viewed "as a contract under which one person (the principal(s)) engages another person (the agent) to perform some service on their behalf which involves 3 delegating some decision-making authority to the agent" (Jensen and Meckling 1976: 308).This scope of discretion allows rational agents, for whom effort generates disutility, to shirk and behave opportunistically at the expense of principals.High powered incentives were initially recommended to align the principals' and agents' interests.But the behavioral economics findings threaten the standard agency assumptions by showing that economic behavior is driven by pecuniary incentives but also by intrinsic motivations and social preferences, thus making the principal-agent relationship substantially more complex.Behavioral agency models have then been developed that integrate these non-standard motives.However, a close examination reveals that standard and behavioral agency models share a similar structure; the sole difference is that additional psychological factors are included in the utility function agents are supposed to maximize (Rebitzer and Taylor 2011;Berg and Gigerenzer 2010).
The main aim of this paper is to critically examine the agency theory's conception of social interactions at work and related conception of cooperation.We proceed by grounding our argument on the distinction between the individual and the person.Our purpose is to show that the interpersonal, i.e. relational and moral dimensions of the relationships engaged in at workplaces are overlooked in (behavioral) agency models.We also briefly characterize the recent evolution of management practices to point out how close they are to the individual-based agency prescriptions and how deleterious they prove to be for the quality of the person-based interactions at work.Our argument is structured as follows.We begin in section Two by distinguishing between the concepts of individual and person.This distinction serves a rhetoric as well as a substantive function throughout the paper aimed at clarifying our argument.Section Three examines the evolution of agency theory's conception of social interactions at work since the 1970s and section Four highlights the extent to which it departs from a person-based conception of behavior.Section Five presents the management rhetoric aimed at mobilizing persons and the actual management practices designed to motivate and control individuals.Section Six gives some evidence of the perverse effects of standard agency theory's prescriptions and makes alternative normative recommendations.Section Seven concludes.While the concept of individual refers to the internal attributes and uniqueness of humans primarily conceived as separate beings, the concept of person adds to these substantive characteristics the recognition that humans are constitutively social and relational (Roger 2012;Harris 1989).In the atomistic, Hobbesian view, humans are able to survive and grow outside the world of social interaction but other philosophical strands hold that humans are constitutively relational beings.To start with, the genesis of the human mind is not monological, something each person accomplishes on her/his own, but dialogical, something that results from the exchanges with others, namely through language (Taylor 1989).Persons are inextricably shaped by the context in which they live and the persons with whom they interact; their preferences and goals evolve throughout life under the influence of such interactions.By contrast, the concept of individual, defined by his "internal attributes" -stable preferences in economics' jargon -fits well the constitutive feature of homo economicus.Two intertwined traits distinguish the person and the individual: moral capacity and relational needs.
All concepts of person explicitly emphasize the moral/judgmental capacities of humans.And moral attributes are inseparable from relations with others and language: "I become a person and remain one only as an interlocutor" (Taylor 1985: 276).Persons are capable of submitting their conduct to shared values and thus of honoring their commitments.The concept of person stresses that agents may act against their self-interest to honor their commitments or their duty.This has long been argued by Sen (2002) who claims that, since commitment breaks the link between individual welfare and the choice of action, the human capacity for commitment breaks with rational choice theory.
By contrast, individuals are seen in mainstream economics as likely to renege on their commitments.Hart (2002) maintains that although firms would have difficulty surviving in a norm-less society, it cannot be assumed that employers or employees are to be trusted to keep their word.He associates keeping one's promise to "irrational" behavior (p.1705) and explicitly states that self-interested parties are not trustworthy parties because there will always be an incentive to break a promise (p.1703).Thus, though a "justice motive" -or some other moral norm -is sometimes included as an argument in utility functions, this very inclusion implies that justice is a tool in the pursuit of (enlightened) self-interest: so long as behaving equitably is perceived as maximizing utility, individuals will do so.
As for relational needs, the person-based tradition highlights the human affective vulnerability and struggle for social recognition; persons enter into relations with others because of their need for relatedness and social esteem (Honneth 1995).
And it must be noted that morally-driven behavior necessarily involves both cognition and affect (termed "cold" and "hot" motives in psychology).That relational and moral motives are intertwined is empirically documented by social psychologists and management scholars.For example, managers adhere to justice rules the more they experience a positive affect (Scott et al, 2014).Positive affective states trigger prosocial behavior such as cooperation and reciprocal helping, which have ties to the rules of justice.By contrast, for rational choice theory, individuals enter into relationships with others when and if they need them to reach their self-goals, not for the relationship per se.
Due to its individual-based behavioral assumptions, mainstream economics primarily associates the agency relationship with agency problems, which stands in sharp contrast with the notion of agency as first introduced by law scholars.In law, the agency relationship presupposes that "the alleged agent and principal have met each other face to face, or have talked on the telephone, or have otherwise communicated in a specific, individualized way" (Clark, 1985: 85).Furthermore, in law, the notion of an agent is one who acts for another even at a cost to him/herself (it hence corresponds to our notion of person); the laws of agency impose a specific duty of loyalty on the agent (Duska, 1992).For Sen (2002), the agency relationship is an instance of commitment: the agent sees himself as charged with furthering the principal's interest; he acts as if the principal's goal is also his own.Mainstream economists adopted the notion of agency but removed the attribute of loyalty/commitment precisely because of their individual-based conception of human behavior.This conception is consistent with their conception of firms as "nexus for a set of contracting relationships among individuals" (Jensen and Meckling 1976: 310).In firms, relationships are not different in any substantial way from market relations (Alchian and Demsetz, 1972).
Assuming that humans are not loyal or bound to their word in the first place, the aim of agency economists is to identify the management practices that would lead the agents to fulfill their agency obligations.They do so by concentrating on the calculative abilities of individuals, required for utility maximization.But, according to law scholars, the functioning of firms does depend on the person-based abilities of persons, their communicative and normative capacities.Acknowledging these capacities supposes breaking with ontological isolation and recognizing the intersubjective ontology of persons (Favereau, 2008).

Agency theory's conception of social interactions at work: costly contractual exchanges
As referred, the agency theory of the firm is grounded on a particular kind of social interactions.Firms are a cascade of sequential principal-agent contracts in which principals delegate work to agents to act on their behalf.As agents are rational utilitymaximizers, principals must bear costs to mitigate agents' opportunism.Agency costs turned out to be the core concept of agency theory and its core aim is to design contractual arrangements that minimize agency problems.In the first agency models, only the vertical -principal-agent -relationships were deemed of interest.
Horizontal interactions3 among co-workers were nonetheless soon acknowledged, under the label of side-contracting.In the eyes of the then leading agency theorists, side-contracting takes the form of "bribes, personal relationships and promises of reciprocation" (Holstrom and Tirole 1989: 94).These "contracts" that agents enter into cannot be fully controlled by principals, which generates agency costs and adds "costly constraints to the owners' optimization problem" (Holstrom and Tirole 1989: 3).Side-contracting is hence considered undesirable and firms must take measures to prevent it.Suggested measures include the limitation of personal relationships through isolation and the restriction of reciprocity through the promotion of short-term relationships (Holstrom and Tirole 1989).Nevertheless, in passing, these authors comment that the measures they recommend can have organizational drawbacks since they may undermine the development of trust, which they view as crucial for cooperation.In subsequent papers, side-contracting is considered as taking two possible forms, collusion or cooperation, depending on its effect on the organization.The incentive structure should then regulate the degree of cooperation in order to limit collusion.In short, employers should consider social interactions among workers with suspicion.
These radical assumptions about both vertical and horizontal interactions at work softened once the findings of behavioral economics started being taken into account.Because it draws on laboratory and field experiments, whose findings differ to a great extent from the neoclassical predictions, behavioral economics claims to improved empirical realism.Experimental labor economists showed that the propensity to provide effort is largely dependent on the nature of social interactions at work, and specifically on whether the principal-agent relationship is perceived to be trustful and fair (Charness. and Kuhn, 2011).Interactions with co-workers also prove to significantly influence the subjects' productivity.New behavioral assumptions, such as inequity aversion and reciprocity -generically considered "social preferences" -, are introduced in extended agency models in the form of arguments into utility functions.
For example, Rob and Zemsky (2002)'s model assumes that workers derive utility from cooperation (defined as reciprocal helping) and that cooperative behavior among workers is a source of competitive advantage for organizations.It is recognized that "preferences for cooperating" are partly endogenous, which means that cooperative behavior may and should be fostered by appropriate incentive systems.
Likewise, while the first agency theorists assume that workers behave a-or immorally towards principals, some recent models elaborate on Akerlof (1982)'s view of labor contracts as partial gift exchange.Dur et al (2010) call attention to the many benefits good relationships between principals and agents may yield to the firm and emphasize the fact that such relationships may motivate workers more powerfully than pecuniary incentives.Employees' effort and the employer's benevolent treatment of workers are modeled as reciprocal gifts, and the signaling of good intentions on the part of principals, though costly, appears as a possibly efficient strategy (Non 2012;Sliwska 2007).
Although behavioral agency models include "person-based attributes", their structure is the same as that of standard models; both fit the axiomatic formulations of rationality (Rebitzer and Taylor 2011;Berg and Gigerenzer 2010).Indeed the models referred above, like all social preferences models, assume that individuals maximize a behavioral (other-regarding) utility function along with a traditional utility function that depends on own payoffs.Utility calculations now also involve psychic costs/benefits.The benefits of own payoffs and the psychic benefits/costs derived from own or others' behavior are weighted and summed to produce a utility score for each allocation, and choice is by definition the allocation with the highest score (Berg and Gigerenzer, 2010).But adding new parameters to the standard utility function further complicates the constrained optimization problem faced by agents and requires even more heroic assumptions about decision processes and calculation abilities.These assumptions are increasingly being denounced as psychologically implausible (Baron and Kreps, 2012;Berg and Gigerenzer, 2010).
The "relational contracts" literature, which emphasizes the role relational contracts play in enforcing incomplete contracts, also attempts to instill more realism in agency models.Relational contracts are considered efficient when outcomes are not contractible ex ante and are observable ex post only by the contracting partiesfeatures typical of employment relationships.Agency models define relational contracts as informal agreements and unwritten codes of conduct (Baker et al. 2002).
Because the relationships are valuable in the future, the contracting parties do not wish to renege on previous commitments.Thus, the honoring of contracts is based on self-interested, utility-maximizing, calculations, not relational traits.The aim of agency relational contracts models is to design the incentive payments that comprise the "necessary and sufficient condition for the relational-employment contract to be selfenforcing" (Baker et al. 2002: 74-75).
To sum up, even in the agency models that introduce relational and moral motives into utility functions, individuals are calculative rather than relational beings (Non 2012;Dur et al 2010;Sliwka 2007).The relational and moral abilities of persons are submerged by the self-interested calculative abilities of individuals: relating with others and following behavioral norms result from individualistic, enlightened, calculations, not from moral capacity: "We do not assume that people follow a social norm for its own sake, but we investigate how such a rule is sustained by selfinterested community members" (Kandori 1992: 63).
Mention must be made of a separate line of behavioral agency theory developed by organizational scholars. 4This strand of literature proposes a version of agency theory that breaks with the standard assumptions by incorporating, for example, bounded rationality and the results of research on choice and uncertaintynamely Kahneman and Tversky's prospect theory (Martin et al 2013).It is also acknowledged that other new behavioral assumptions about time discounting, inequity aversion and intrinsic motivation influence agents' and principals' behavior (Pepper and Gore, 2015).Unlike most behavioral economics, this literature does not systematically retain the standard utility maximization models.means that compliance with commitments and expectations about others' behavior acquires crucial importance.Yet, individuals who do not abide by social norms for their own sake would not expect co-workers to cooperate.
In fact, it is the relational needs and moral abilities of persons that both establish and enforce mutual expectations.More precisely, workers cooperate because this produces mutually beneficial outcomes but also because it provides the opportunity to enter into interactions that satisfy the human need for relatedness.
The notion of "relational goods", defined as the communicative, cognitive and affective components generated by the interpersonal relations one enters into (Bruni 2008;Gui 2000;Uhlaner 1989), has been advanced to designate the relationships that are intrinsically motivated.But it is to be noted that cooperating gives rise to relational goods only if and when the interacting workers jointly value it.Bruni et al (2014) show that cooperation depends on the value that one's interacting partners give to cooperating.Relational goods (friendship, camaraderie, reciprocal helping) are more than just a combination of private goods; they go beyond the individualistic derivation of utility to the extent that they require a common valuation (Lopes 2011).Relational goods explain why workers do not systematically exploit cooperative partners (requisite (a)).
Cooperating also involves a series of mutual commitments that, depending on whether or not they are honored, result in moral goods/bads (e.g.feeling fairly treated and respected versus deceived and humiliated).Moral goods may be defined as the outputs of a moral nature generated by the social interactions one enters into (Lopes 2011).Moral goods involve and require the common valuation of the norms that guide the actions of the interacting workers.The normative appeal of moral norms derives precisely from their being commonly shared.The fact that moral norms prescribe actions that may lead to acting contrary to self-interest is of special importance.
Indeed, only persons abiding by moral norms can expect others, whom they know share the same norms, to also abide by them even when it runs counter to their interest.The existence of moral goods explains how requisite (b) is met.
To sum up, cooperative behavior at work cannot be accounted for without considering the role played by relational and moral goods in motivating and sustaining cooperation.As such, these goods may be seen as a kind of productive factor.While cooperation undoubtedly entails a calculative facet, it also inevitably calls for the relational and moral capacity of persons.Only morally able persons can effectively enter into commitments with others, be them colleagues or principals.Figure 1 synthesizes the differences between the person and the individual-based conceptions of cooperation.Although the behavioral developments of agency theory recognize the cooperative dispositions of workers, the way in which the latter are modeled does not actually acknowledge their relational and moral nature.To begin with, agency theory does not consider cooperation as an outcome of continued social interactions; that is, it is not viewed as a collective endeavor rendered possible by constitutive traits of humans.Rather, cooperating is considered a residual kind of behavior: "voluntary cooperation is defined as the difference between actual and privately optimal effort" (Non 2012: 322).The decision to cooperate is taken separately by each worker depending on her/his preferences rather than emerging from the fact that workers are engaged in a collective venture.Cooperative dispositions, or any other ad hoc social preference, are introduced into the individualistic theoretical apparatus by adding arguments of a different nature into a single utility function.But this operation actually overlooks the interpersonal dimension of social interactions, i.e. their specific affective and moral nature.
Firstly, in interactions at work, the identity of the person with whom one interacts has affective significance.Sugden (2005), inspired by Smith's notion of sympathy, argues that what occurs in interpersonal interactions is a correspondence of sentiments, i.e., a two-way interaction at the level of sentiments and emotions.It is the mutual awareness of the correspondence of sentiments that makes it a truly interpersonal phenomenon and that confers subjective (affective) value to interpersonal relations (hence the term relational goods).What behavioral economics calls "social preferences" are often actually affective states.Yet, for Sugden, affective states cannot be considered preferences, "they are not, as preferences are, comparative relations among objectively described options; rather, they are our subjective experiences of the world" (Sugden 2005:54).These experiences influence the process of preference formation and hence cannot be assimilated to preferences.
Being neither preferences nor beliefs, affective states do not fit into the ontological framework of rational choice theory (Sugden 2005); their inclusion in utility functions may well be illegitimate.
Secondly, as referred above, Sen (2002) contends that commitment or moral motives in general cannot be reduced to arguments in utility functions because actions driven by duty are "counterpreferential".Acting in committed ways, i.e. to further the goals of another party or of a group, is not representable within the belief-desire schema.Additionally, the specific features of moral behavior, namely discontinuity, "lumpiness" and non-marginality, raise serious technical issues when it comes to modeling.But even the authors who propose models that acknowledge these specificities and make them consistent with the economic view of rational behavior point out that straightforwardly assuming tradeoffs between moral and non-moral considerations may well be illegitimate (Dowell et al, 1998).That is, they feel uncomfortable with the assumption of universal commensurability between arguments of a different nature.Can a price be assigned to the "psychic cost" of not having honored a commitment and can it then be lumped together with ordinary costs and benefits?
In sum, the desire to propose models that fit the axiomatic formulations of rationality leads behavioral agency theorists to downplay the abilities of the person.
This downplaying is consistent with agency theory's conception of the firm which, in defining firms as "privately owned markets" (Alchian and Demsetz 1972: 795), is in fact claiming that the functioning of firms does not require morally-driven behavior.The argument developed in this paper is just the opposite.One of the basic distinguishing features of firms when compared to markets is that firms provide the opportunity for intensive interpersonal relationships.It is precisely the unspecified obligations and the relational goods deriving from social interactions that, combined with the economic advantages of cooperation, contribute both to generating the desire to cooperate and to rendering cooperation sustainable.In short, while mainstream economics assumes that relationships established within firms are like those established in markets5 -that is, contracts between individuals -we consider that within firms workers establish relationships as persons because they are engaged together in productive endeavors.
Relational and moral goods may obviously also be "bads" (animosity, disrespect, resentment, etc), in which case they can be destructive for the workers' well-being and organizational performance.The fact that workers are persons as defined above does not obviate that they do often behave as self-interested, indeed immoral individuals.Moreover, because of their high emotional potential, personalized interactions can disrupt cooperation more than impersonal contexts.It is therefore of prime importance that work environments allow the nurturing of relational/moral goods to avoid the losses and inefficiencies caused by relational/moral bads.

From mobilizing persons to managing individuals -towards the depersonalization and individualization of work
The individual-based and the person-based conception of social interactions and cooperation lead to very different normative prescriptions.The aim of this section is to show that while the person-based conception underlies management rhetoric, it is the individual-based rationale as conveyed by agency models and prescriptions that seems to inspire management practices.
Managers know that cooperation is indispensable to achieve profitability and foster innovation.They also know that workers are morally-endowed persons who, for the most part, are willing to behave cooperatively rather than opportunistically.They therefore try to draw workers into a cooperative endeavor by asking them to invest themselves in work cognitively and affectively.The prevailing management rhetoric solicits the mobilization of the "whole person", i.e. requires that workers internalize the organizational goals and commit themselves to achieving them.Indeed, the workers' loyalty is more efficiently fostered through internalized commitment than through obedience, technical prescriptions and bureaucratic control. 6  Management rhetoric puts strong emphasis on trust and the cooperative spirit.
To be consistent, management practices should give great leeway to workers and 6 This raises a crucial issue: the fact that workers as persons need to nurture good relationships and are prone to behave cooperatively is often used by firms to manipulate them and perpetuate oppression in the name of overcoming it (Alvesson and Willmont 1992).Our intent is obviously to argue that firms must acknowledge the social and moral abilities of workers without using them to further exploit workers.
signal that they trust them, which means enhanced autonomy at work, decentralized decision-making and workers' participation, above-market pay, collective rewards and job security.This would foster a kind of relationship between workers and the organization based on "partial-gift exchanges" (Akerlof 1982), in which workers reciprocate in effort and dedication the trust and good working conditions offered to them.If firms fail to honor their commitments, workers disengage from the moral obligations that such practices are supposed to generate.
In fact, evidence shows that firms' practices habitually break the "psychological contracts" presumed in the cooperative rhetoric (Thompson 2013).The trends that characterize the world of work to-day7 sharply depart from the rhetoric of trustbuilding and empowerment and have instead converged with the prescriptions of early agency models.To minimize the agency problems generated by the workers' pursuit of their self-interest, Jensen and Meckling (1976) prescribed two kinds of arrangements: As for monitoring, direct types of control, such as surveillance and giving orders, are replaced by "unobstructive" controls like standardization, involvement of workers in the setting of their performance targets, individualized and quantified appraisal systems, extensive reporting procedures and other sophisticated devices that, thanks to the new management technologies, often conceal their true nature.
The outcome is an overall decline in perceived autonomy at work in the last two/three decades (Lopes et al. 2014).These trends in the evolution of work are reinforced by those in the evolution of employment, marked by depressed labor markets and a decline in job security.
There is nothing new in the observation of a gap between managerial words and deeds; the novelty lies in its marked increase.Management rhetoric advocates personal commitment and trust-building, but in practice firms invest in increasingly sophisticated surveillance and control devices.The person-individual distinction captures this paradox: firms want workers to involve themselves at work as persons but they organize work as if workers were opportunistic individuals in need of strict direction and monitoring.By individualizing incentives, controls and rewards, firms rely on the self-interest of workers and their competitive dispositions rather than on their trustworthiness and cooperative dispositions.In the end, is organizational efficiency best served by competition, rivalry and envy, or by mutual trust and reciprocity as advocated in management rhetoric 8 ?Since work is not only a means to consumption but also an opportunity for workers to realize their potential and satisfy their social needs, workers tend to subjectively engage in their work and actually devote large amounts of cognitive and emotional resources to it.Thus, to the growing gap between management rhetoric and practices corresponds an increasing divergence between the workers' expectations of meaningful work and the working conditions they actually face.The individualization and depersonalization of work hence creates a tension between the need for meaningful work, which necessarily includes the feeling of usefully contributing to and being part of a community, and the pressure to enter into a competitive, zero-sum game.On the one hand, workers involve themselves as persons in work9 but, on the other, they are compelled to behave as self-centered individuals to meet the quantitative objectives set by management, hence contributing to the erosion of the cooperative spirit needed for their psychological well-being.If one could observe the relational and moral goods at work, one would certainly see them regressing.
These management practices have resulted in a considerable intensification of work, increasing levels of stress and emotional exhaustion, substitution of extrinsic for intrinsic motivations and the dissolution of collective solidarities.Intensification of work is reported as being accompanied by a decrease in the time available for socializing -fewer "non-productive" moments; less time to learn, teach or help; fewer opportunities to meet and communicate -as well as by a general depersonalization and deterioration of work relations (Le Gall 2011).The increase in work-related social and psychological problems is a good indicator of this unprecedented deterioration in the quality of working life.In the last decades, the workers' vulnerability and psychosocial disorders have intensified considerably (Netterstrom et al. 2008;Siegrist 2006).
Burn-out and stress, feelings of isolation, and feelings of culpability when performance targets are not met have become prevalent phenomena of contemporary workplaces (Le Gall 2011).

The perverse effects of standard agency theory's prescriptions and beyond
Although as yet the existence of a direct link between the evolution of management practices and the influence of agency theory is insufficiently empirically documented10 , a number of papers convincingly argue that agency theory's basic assumptions have become a dominant ideology, and consequently a powerful normative model (Roger et al 2012;Goshal 2005;Roberts 2005;Duska 1992).As mentioned, the standard prescriptions of agency theory emphasize pecuniary incentives and control devices.experimental economists and organizational scholars, shows that pecuniary rewards and tight control may have deleterious and counterproductive effects on intrinsic motivation (Larkin et al 2012, Charness and Kuhn 2011, Deci and Ryan 2000).Indeed, strict monitoring is shown to threaten the sense of personal autonomy, thereby damaging self-esteem and decreasing intrinsic motivation (Gagné and Deci 2005).
When workers are excessively controlled, intrinsic motivation is supplanted by defensive or self-protective processes such as the tendency to withdraw concern for one's job and focus on oneself (Deci & Ryan 2000).Workers then tend to adhere to more materialistic values and behave less cooperatively.Sheldon et al. (2004) show that workers involved in controlling environments declare to be less satisfied at work and more focused on pay and benefits.
In the same vein, the use of monitoring tools by managers leads them to distrust workers and triggers a pathological spiral.These processes are well-known by psychologists: "Surveillants come to distrust their targets as a result of their own surveillance and targets in fact become unmotivated and untrustworthy.The target is now demonstrably untrustworthy and requires more intensive surveillance, and the increased surveillance further damages the target.Trust and trustworthiness both deteriorate" (cited in Goshal 2005: 85).This counterproductive effect has been shown to also affect top executives, for whom it is recommended that compensation schemes are aligned with shareholders' interests.Many of the reforms set up to curb top executives' opportunism in the last decades ended up augmenting rather than diminishing the governance problem (Roberts 2005).
That is, instead of mitigating opportunistic behavior, the standard agency prescriptions may actually enhance opportunism (Goshal 2005;Roberts 2005).The trend towards the quantification of work, for example, may result in workers engaging in individual gain calculations rather than exercising their person-based abilities.As Hannah Arendt (1958) brilliantly contended, the danger with theories is not so much that they are false, it is that they may become true.In assuming that people are opportunistic and primarily motivated by pecuniary motives, agency theory may contribute to making people be just like that.
The behavioral agency models take stock of these perverse effects but stick to designing more appropriate incentive structures rather than devising adequate organizational solutions.For Rebitzer and Taylor (2011), it is pay structures that have to perform the multiple duties of motivating workers, signaling the principal's trustworthiness, making workers identify with the organization and taking account of agents' intrinsic motivations and social preferences.Rob and Zemsky (2002), who assume that cooperating increases the workers' utility, recommend that pecuniary incentives be set so as not to crowd out the workers' cooperative dispositions; but they say nothing about how these may be fostered though they recognize that cooperative dispositions are endogenous to the work context.Similarly, since it is now acknowledged that monetary incentives crowd out intrinsic motivations, high-powered incentives are considered "inefficient".Holmstrom (1999: 89) states that firms must often suppress "excessively strong incentives on individually measured performance for the benefit of enhancing the effectiveness of more delicate and subtle instruments aimed at encouraging cooperation".But nothing is said about what the content of these "subtle instruments".Besides, low-powered incentives are now assumed to potentially attract workers with high intrinsic motives, which may subsequently promote high effort work norms (Sliwka 2007).Notwithstanding, according to agency theory, this kind of incentive structure always entails a second-best situation.
In fact, the inclusion of social and moral preferences in principal-agent models results in contradictory claims and prescriptions.In Non (2012), for example, the good treatment of workers by managers is modeled as a cost for firms.So are cooperation and satisfying interactions a cost or an efficiency-enhancing factor?
Though behavioral economists declare their models to be purely descriptive, their findings may have serious effects on prescription.If "social preferences" were taken seriously, prescriptions should include not only adjustments in incentives but also recommendations about how to design social interactions at work -including hierarchical interactions, for it must be borne in mind that workplaces are structured in terms of the authority of managers.Bruni et al (2014) provide a fruitful indication: their evolutionary model shows that cooperation is enhanced both by social rewards for cooperation -i.e., the use of material rewards to signal social recognition -and by high frequency of social interactions.A mix of objective and subjective incentives is better at driving cooperation than sanctions and tight control.Hargreaves Heap (2013) insightfully suggests that, rather than assuming opportunism, managers should consider which preferences they would like to develop and adjust organizational structures accordingly.
In fact, contrary to what agency theory assumes, and as management rhetoric shows, real-world firms do expect workers to be loyal, conduct themselves in a trustworthy and cooperative manner and honor the non-contracted obligations inherent in incomplete contracts.Because the very existence of firms is justified by the possibility of cooperation, prescriptions should focus on fostering cooperative dispositions rather than trying to economize on and indeed discard the person-based moral and social capacities.This implies recognizing that firms are not nexus for contracts but rather organizations, and that management not only involves conceiving incentive systems but also designing organizational structures.
Managerial perspicacity cannot be relied upon to prevent firms and workers from being increasingly caught in the paradoxes described above.There is no robust evidence showing that cooperative or participative firms are more profitable than exploitive ones.It may well be the contrary since the externalization of social costs may compensate for motivational losses.The effective implementation of management's rhetoric, if it is not to be manipulative, demands giving much voice and power to workers, something most employers are certainly not prone to do; institutional changes that further the workers' rights are required.But to be effective changes in labor law that foster the workers' voice11 must be necessarily accompanied by changes in corporate law that institutionalize the workers' participation in the corporate governance structures.Questioning the behavioral foundations of agency theory hence leads us to acknowledging a need for juridical and economic regulation of firm governance -which is precisely what the launching of agency theory wanted (and succeeded) to avoid in the 1970s (Gindis 2013).

Concluding remarks
The paper begins by distinguishing the notions of individual and person and then draws on this distinction to i) critically examine agency theory's conception of interactions at work and ii) show that management rhetoric addresses workers as persons but actually organizes work as if workers were opportunistic individuals.While the concept of individual in mainstream economics defines humans as unique and separate beings engaged in the pursuit of self-centered goals, the concept of person defines them as constitutively social, morally-endowed beings prone to cooperative behavior.
Being built on the specific interaction between a principal and an agent, the early agency models focused on the agency costs deriving from the misalignment of these interacting partners' interests.But presently most agency models integrate the findings of behavioral economics (in the form of psychological arguments in utility functions) and provide a more complex picture of social interactions at work, of their efficiency character and of their impact on incentive systems.Notwithstanding this substantive enlargement of the domain of economics, the way in which the -personbased -"social preferences" and "moral dispositions" have been introduced in agency theory leaves the individualistic utility-maximization framework unquestioned.Even though the mainstream begins to explicitly consider the employment relationship a "social relationship" (Baron and Kreps, 2012), the fact that work in firms necessarily entails personally relating with others continue to be actually disregarded.This is due to the incapacity of the atomistic ontology of the rational choice paradigm to effectively cope with the cooperative facet of work.
By contrast, the distinctive methodological feature of our approach, i.e. the focus on the interpersonal dimension, allows us to uncover the powerful role played by social interactions in work organizations.We argue that cooperation, on which the functioning of modern productive processes depends, as signaled by management rhetoric, requires the social and moral abilities of persons to be effectively established and sustained.A meaningful theory of the firm should hence conceive it as a "system of cooperative services of persons" rather than just the "sum of services of individuals" (Barnard 1938: 110, our italics).That is, the behavioral assumption placed at the core of the theory of the firm should consider both the cooperative dispositions and the self-interested dimension of human behavior.The study of social interactions at work, in particular, deserves greater attention than it has received so far.
The processes of individualization and depersonalization that mark present-day work are very much in line with agency theory's standard prescriptions.Beyond the deeply deleterious effects they have on the workers' well-being, these processes may be fostering the opportunism they were supposed to restrain.Recognizing that the good functioning of firms and their creativity necessitates treating workers as persons rather than individuals leads to questioning firms' governance structure.In order to compel firms to be more consistent with their own rhetoric, major institutional changes are required.

2.
The concepts of the individual and the person All heterodox economists emphasize the need to replace the mainstream atomistic conception of the individual with one that is socially embedded, but a change in terminology had not yet been envisaged.Conversely, in philosophy and anthropology the terms individual and person are employed to single out different views of human beings.(It must be mentioned that both concepts are historical creations.)

10 4 .
Social interactions at work as drivers of cooperation and major distinguishing feature of firms The insight that economic agents have other-regarding preferences represents a welcome expansion of standard agency models.It helps in understanding cooperation but still provides an insufficient account of it.Cooperation is critically required by modern production processes, characterized by high levels of uncertainty and strong interdependence.Most workers have to constantly engage in binding social interactions and mutual commitments to establish the common understandings and routines necessary for each to contribute her/his part to the productive endeavor.This section proposes a person-based conception of cooperation and highlights its differences with the individual-based conception.Cooperating implies (a) giving up on one's desire to cheat or exploit cooperative partners and (b) expecting that others cooperate too.Requisite (a) means that workers must commit themselves to the pursuit of the previously set goal rather than the onesided pursuit of self-interest -a behavior that unrelated and self-interested individuals would not exhibit if it happens to diverge from their private interest.Requisite (b)

Figure 1 :
Figure 1: The person-based and the individual-based conception of human beings and cooperation (a) incentives and compensation schemes aimed at aligning the principals' and agents' interests (particularly recommended for top corporate executives) and (b) control devices aimed at keeping self-serving behavior in check and providing information about what agents are actually doing.As for incentives, schemes such as performance-related pay, individualized and quantified performance targets, performance rankings and rising wage differentials are being widely implemented to boost productivity.Most performance measurement practices assign objectives and accountability for outcomes to individuals despite task and functional interdependencies.Larkin et al (2012) reveal that over 83% of papers on compensation in the leading strategy journals in the 2000s rest on standard agency theory rather than social psychology or behavioral decision research.The human activity of work is being translated in quantitative terms, in all economic sectors, to fit the constraints of financial analysis.It is the outcomes of work that are focused on, not the workers as persons.The subjective and collective experience of work is largely overlooked.