Agency theory and performance appraisal: How bad theory damages learning and contributes to bad management practice

Introduction

In an influential article published posthumously in 2005, Sumantra Ghoshal argued that what he termed ‘bad’ theories were wreaking havoc on management practice. His argument was focused with particular force on agency theory. In Ghoshal’s view, this had produced a preoccupation with self-interest among managers who then engaged in behaviours inimical to their organisations’ long-term interests. Immediate support came from leading management thinker Jeffrey Pfeffer (2005), who argued that the bad theories in question become self-fulfilling. That is, business schools teach them as though their empirical basis was much sounder than it is. Students emerge with an enriched theoretical lexicon but a diminished sense of social responsibility, and poor management practices become institutionalised to such an extent that although they fail to deliver on their original intentions they become naturalised and, therefore, are assumed to be beyond interrogation (see also Ferraro et al., 2005).

Consistent with this critical approach, we argue that many of the problems associated with performance appraisal can be explained by the wider influence that agency theory has had in organisations. To be clear, it is not our intention to suggest that all problems with appraisal can be explained in this manner. ¹ Nor do we deny that many employees and managers find appraisals to be a positive experience – for example, when there is a high degree of employee participation in the process (Roberts, 2003). However, our focus is on those instances where negative outcomes are the norm. In exploring why this is so often the case, we argue that agency theory has had a significant and often harmful impact on management learning and practice.

We discuss the nature of agency theory throughout this article. Here, we simply note its foundational assumption that ‘there is potential for mischief when the interests of owners and those of managers diverge. In those circumstances … managers may be able to extract higher rents than would otherwise be accorded them by owners of the firm’ (Dalton et al., 2007: 2). Agency theory, therefore, explores how firms can ensure an alignment of interests between the principal (owners) and agents (employees). It is assumed that people are motivated by self-interest, mainly or exclusively in the form of economic calculations that over-ride such issues as trust, loyalty and friendship networks (Besley and Ghatak, 2005).

We argue that the theory’s influence on the practice of appraisal deserves more serious attention than it has so far received. We focus on this since the manager–employee appraisal process remains a widely used mechanism for performance management in most organisations (Chartered Institute of Personnel and Development (CIPD), 2009, 2014). Researchers have identified many problems associated with how they are implemented. It is therefore not surprising that much research has focused on ‘improving’ the process of performance appraisal. However, this tends to stress the psychological rather than contextual/ideological aspects of dysfunctional appraisals (Ilgen and Favero, 1985). It depicts appraisers as suffering from biases in their perceptions of others, lacking in feedback skills, poor at listening, too wedded to appraisal forms, insufficiently wedded to appraisal forms, inclined to offer feedback too frequently or too infrequently or as otherwise deficient in some toolkit that it is assumed will improve their practice. This research lacks a deeper understanding of the root causes of the problems it addresses and avoids a critical engagement more generally with management theory. To address these deficits, we explore the problems of performance appraisal systems in the context of the wide-ranging influence of agency theory.

Thus, our argument, and contribution, is as follows. First, we briefly discuss both the positive intentions of performance appraisals, and some of the most pressing problems associated with it that have been widely identified in the literature. Second, we consider the wide-ranging influence of agency theory and argue that, as a result, an agency-induced mind-set often impacts on the practice of appraisal. We argue that this undermines the developmental intentions of performance appraisals that are regularly espoused by many researchers, consultants and managers.

Consequently, we suggest that the incremental approach of most research into appraisal, which focuses on its limited effectiveness, empirically explores incidental aspects of its operation and postulates minor adjustments that will purportedly ‘improve’ its utility are misplaced. ² These approaches encourage managers to keep on adopting it. The view seems to advance through the stages of it should work, it might work, it will work (eventually) – if only we correct this or that detail. This begs a question posed by Metz (1988) that could equally well be asked today: ‘why, in the constant process of appraisal systems revisions, can’t we seem to get it right?’ (p. 47). An obvious answer might be that no one can get it ‘right’ and that a more critical engagement with the practice of appraisal is required. We therefore argue that research needs to reject the neo-human relations tradition that has dominated much of the research on performance appraisal. In contrast, we advocate a critique that problematises issues of power, control and conflicts of interest beyond what such analyses have offered to date.

Performance appraisals – good intentions versus bad outcomes

It has been claimed that performance appraisal brings a multitude of benefits to organisations and their employees. These include the opportunity to ensure that staff pursue goals that are aligned with the wider organisational objectives set by senior managers; the provision of objective assessment and regular feedback, which it is assumed will improve learning; heightened commitment and motivation; improved career management through the identification of training and development needs; the creation of legal documentation for use in cases of discrimination, grievance and disciplinary processes or wrongful dismissal; an improved correlation between the wages bill and organisational performance, through linking appraisal to performance-related pay; and an overall increase in performance (Nikols, 2007). Many more such claims can be found in the literature.

Paradoxically, performance appraisal is also widely viewed as one of the most persistent problems faced by both managers and employees. Concern goes back almost 60 years, when McGregor (1957) published an article in Harvard Business Review entitled ‘An uneasy look at performance appraisal’. It is often viewed as an annual fiasco that some have suggested should be abolished altogether (e.g. Coens and Jenkins, 2000). One major review of the area has concluded that there is no evidence as yet to connect individual appraisals with firm-level improvements in performance (DeNisi and Smith, 2014). Given the huge volume of research on the topic, much of it practitioner oriented; this is remarkable. There can be few other examples of something that promises so much, delivers so little, but which is so universally applied – an observation made long ago by Grint (1993) and which retains its validity today. How can this be?

One of the central foci of human resource management (HRM) has been how effectively approaches to people management support the overall business strategy of organisations (Martin et al., 2016). When those strategies are dominated by agency perspectives and economic considerations it is hardly surprising that human resource (HR) practices, including appraisal, come to share those influences. Issues of context, power and control exert an enormous influence on employee-manager relationships. But these are banished to the side-lines, while managers are exhorted to improve their techniques for evaluating performance, offering feedback and setting goals. This reinforces the glaring disconnect between positive intentions and dysfunctional outcomes.

The influence of agency theory

Economists have long used agency theory to promote a particular understanding of the relationship between performance measurement systems and the provision of incentives (Fama and Jensen, 1983). It is not without significance that Jensen and Meckling’s (1976) seminal article was titled ‘Theory of the firm: Managerial behaviour, agency costs, and ownership structure’. ⁴ From the outset, it seems that its authors intended it as an all-encompassing theoretical explanation of organisations. The theory has been most often employed in research on the mechanisms used by owners to align the interests of chief executive officers (CEOs) with those of organisations (Gomez-Mejia and Balkin, 1992). The so-called principal–agent problem (Spencer, 2013) revolves around the extent to which a principal must devote effort to minimising shirking behaviour by an agent who is motivated by self-interest and cannot be trusted. In its own way, the theory thus recognises a core tenet of Critical Management Studies (CMS) – that is, that organisations have differentiated rather than unitarist interests. But the similarity ends there. Agency theory explores variegated interests from the perspective of ‘How can an organization, through its owners and stewards, minimise the posited tendency for managers to inappropriately leverage their advantage when managers’ interests are not consonant with those of owners?’ (Dalton et al., 2007: 2). The primacy of shareholder value and owner ‘rights’ is taken for granted. It is assumed that an owner’s expression of self-interest is tolerable since it somehow embodies a greater good, while that of other organisational actors does not. The function of management systems, including appraisal, is viewed as one of aligning everyone’s activities with the needs of owners, rather than ensuring that owner behaviour is aligned with the needs of other stakeholders.

If anything, the problem of conflicting interests within organisations has intensified. Davis (2009) argues that corporations are less concerned than ever with long-term relationships and building in-house capacity with self-interest increasingly at the fore of organisational behaviour. Although agency theorists sometimes acknowledge that this self-interest is ‘bounded by norms of reciprocity and fairness’ (Bosse and Phillips, 2016), it is also assumed that ‘the interests of the principal and agent diverge and the principal has imperfect information about the agent’s contribution’ (p. 276). It follows that incentives are required to narrow the gap in interests. Assembling more information (if necessary, through tight reporting mechanisms and close surveillance) is also helpful. Defenders of agency theory have argued that ‘conflicts of interest’ are not the same as the preponderance of self-interest and that agents and principals have a variety of motives for their actions (e.g. Buchanan, 1996). Be that as it may, in practice, its proponents lean overwhelmingly to a narrow view of self-interest around financial calculation (Heath, 2009).

This presents an imbalanced view of human behaviour since although we may be hard wired to prioritise our own interests and to compete, we are also hard wired to cooperate, reciprocate favours, behave altruistically and value fairness (Bowles and Gintis, 2011; Highfield and Nowak, 2011). The ultimatum game offers a good illustration (Güth et al., 1982). One person with a sum of money (the proposer) offers it to another (the receiver) who knows how much the ‘proposer’ possesses. Self-interest would suggest that if the sum is US$10, the proposer should offer only US$1 and that this would invariably be accepted since the rules of the game dictate that refusal by the responder means both sides get nothing. Repeated studies show that in actual fact the proposer generally offers 50 percent of the sum available and that if their offer falls below 25 percent, it will most likely be rejected. Perceptions of fairness appear to trump naked financial self-interest. Powerful group identities, forged in fire, can also find people subordinating their own fate and interests to the welfare of the group. ⁵

Regardless, agency theory assumes that in most organisations the principal’s goal is the promotion of shareholder value – generally held to be a good thing – while agents are assumed to be primarily interested in self-aggrandisement (Angwin, 2015). In our view, this is a highly idealised view of the principal’s role in principal–agent relationships. Applebaum and Batt (2014) document predatory rent seeking behaviours by private equity firms that are more concerned with the short-term enrichment of their funds than adding long-term value, often destructively. In such instances, the architects of corporate misfortune turn out to be owners rather than managers. Despite such shortcomings, the theory has continued to acquire traction and, as we argue in this article, become an ideological template for management–staff relationships within organisations. Consistent with this, there have been calls for agency theory to be extended beyond the economics perspective and to encompass both organisational behaviour and non-traditional settings (Eisenhardt, 1989; Kallifatides, 2011; Mitnick, 1992; Wiseman et al., 2012). This view has extended to the public sector, where new public management (NPM) applies the logic of agency theory in a culture of auditing, monitoring and appraisal (Ballantine et al., 2008; Craig et al., 2014). For example, appraisals are now commonplace in many environments where they would once have been disdained, including universities, and where they reproduce some of the negative effects that have been well documented in the private sector (Simmons, 2002). This is not without its ironies. Precisely, at a time when, in the aftermath of the Great Recession, the shareholder value model is most suspect (e.g. Mansell, 2013; Segrestin and Hatchuel, 2011; Stout, 2012), HRM seems to be embracing it with renewed devotion. In castigating what he terms HR’s desire to ‘look up the organization’, Marchington (2015) proposes that it has ‘become a servant to short-term performance goals and the mantra of shareholder value rather than the development of longer-term sustainable contributions based on shared values and fairness at work’ (pp. 176–77).

Thus, we suggest that agency theory has itself become an ‘agent’ within the theory and practice of management and has greatly influenced the concepts that managers learn during their formal and informal training. Pfeffer (1995) reported that the year before Jensen and Meckling’s article appeared, only 2.5 percent of the articles published in Administrative Science Quarterly and 0 percent of the articles in Academy of Management Journal cited economics or economists. By 1985, these proportions had become 30 percent and 10 percent. By 1993, they had risen still further, to 40 percent and 24 percent. It is, therefore, no surprise that agency theory has now ‘diffused into business schools, the management literature, specialised academic and applied practitioner journals, the business press, even corporate proxy statements’ (Shapiro, 2005: 269).

Levy and Williams (2004) argue,

that agency theory models have widespread implications for companies at both the individual and organizational level as the links between basic level constructs such as goals and participation could be examined and tied to employee attitudes, employer-employee relationships, employee performance, organizational effectiveness and employee withdraw behaviors. (p. 889)

These implications surely constitute one reason for its success since the theory has at least the merit of offering seemingly simple prescriptions for managerial actions. Another is that the theory chimes with the hierarchical and power-saturated nature of organisations and management work. It has an intuitive appeal for many since it seems to merely describe ‘what is’ and which thus surely must lie beyond interrogation. As Davis (2013) puts it, ‘Thanks to two decades of restructuring driven by a quest for shareholder value the global supply chains of contemporary corporations increasing resemble the “nexus of contracts” described by the finance-based theory of the firm’ (p. 35). Thus, the theory becomes naturalised by dint of its association with practice, and practice becomes further naturalised because of its association with the theory.

Other ironies abound. Agency theory is founded on distrustful and pessimistic notions of human motivation and behaviour. Conceived in the first instance in terms of managerial behaviour and the need to align it with the interests of shareholders, it is scarcely surprising that managers then come to view (other) employees in the same distrustful light in which they are viewed themselves (Roehling et al., 2005; Thompson, 2011). Managers thus become trapped in an Orwellian state of Doublethink. On one hand, they subscribe to the supportive, co-operative and developmental purposes of appraisal interviews (Mind-set 1). On the other hand, they are also influenced by the more sceptical notions of agency theory (Mind-set 2). These contradict and undermine the positive aspirations of Mind-set 1. Such role conflict undermines the good intentions offered in defence of HR practices, such as performance appraisals. Yet line managers primarily adopt Mind-set 2 and implement HR policies with the primary intention of meeting cost-focused performance targets (Evans, 2015).

This confusion represents a step backward for the HR profession and management more generally. Lubatkin (2005) argues that agency theory makes assumptions ‘about the nature of individuals, organizations and markets that take the model out of the realm of organizational reality’ (p. 213). In looking much more critically at appraisal from this perspective, it becomes apparent that the further adoption of an agency perspective will intensify the difficulties with how it is usually practised rather than ameliorate them. We highlight these issues in Table 1, where we identify some of the major claims made for appraisal in the literature, show the parallel problems that have arisen in their practice and relate these to accompanying assumptions from agency theory that sheds light on the widespread nature of the problems concerned. We then focus in-depth on each of them in the text that follows.

OPEN IN VIEWER

Table 1.

Appraisal systems and the influence of agency theory.

Intentions with appraisal	Problems with implementation	Influence of agency theory
• Alignment of employee performance with organisational goals	1. Reliance on short-term measures to assess individual employee performance	• A predominantly economic view of the employment relationship ensures that short-term financial measures dominate assessment of organisational performance
• Provision of useful developmental feedback for individuals and teams • Strengthening of employee motivation through clear goal setting and better linking of rewards to goal attainment	2. Developmental feedback undermined by relating appraisal to pay and/or close scrutiny of performance   Leads to negative or insignificant effect on affective commitment, job satisfaction, organisational citizenship behaviour   Promotion of self-interested behaviours   ‘Gaming’ behaviour by employees	• Privileging of management voice in determining goals, rewards and training needs • Assumption that employees will shirk without close monitoring of their behaviour • Individual rather than collective interests dominate behaviour
• Open discussion of performance	3. Prioritisation of individual rather than team performance	• Individual rather than collective interests dominate human behaviour
• Joint commitment between managers and staff to goal attainment	Leads to increased surveillance by managers and greater shirking by employees
• Objective and accurate record of performance • Creation of legal documents for organisational reference, of use where discrimination occurs and also where disciplinary action is required	4. Difficulty in accurately and objectively measuring performance   Leads to increased mutual distrust	• The notion of ‘moral hazard’ suggests that people will take risks or avoid effort if they imagine the costs of doing so will be borne by others • Employees likely to shirk
• Provision of honest, accurate feedback to guide improved performance	5. Self-efficacy biases cause employees to have more favourable view of performance than their managers do   Leads to employees growing more fearful and distrustful of managers	• Predominantly instrumentalist view of employment relationship • Principal must monitor and control the efforts of wayward agents, who will otherwise deviate from goal pursuit

Problem 1: reliance on short-term measures to assess individual employee performance

Surveys of appraisal practice (e.g. CIPD, 2009) consistently show that ‘a results orientation has come to be the dominant approach for expressing performance requirements’ (Ward, 2005: 5). Monks et al. (2012) also reported a tendency in their sample for the close monitoring of behaviour:

through metric-based performance management systems that focused on the achievement of targets … individuals were castigated for poor performance … if they failed to meet the performance standards required in the initial training. Reward systems were very closely tied to performance metrics that related to output. (p. 389)

Indeed, Edwards and Wajcman (2005) drew the conclusion that the increased use of appraisal systems contradicts the notion that we have entered an era where formal bureaucracy has given way to systems built on trust and autonomy. Instead, they argue that it is indicative of a growth of bureaucratic control mechanisms, of the kind that we suggest are largely induced by an agency mentality.

The predominance of measurable targets and the close monitoring of employee achievement of such targets are, of course, aligned with agency theory assumptions and, in particular, with the notion that economic interests are the key driver of human behaviour. It assumes that agents are shirkers, with a self-interested incentive to avoid work and viewed as ‘resourceful, evaluative maximizers’ (Jensen, 1994: 1), pursuing money, respect, honour, love and whatever else is in their interests, while being willing to sacrifice the common good to do so. Agency theory’s assumption of individualistic interest is explicit. It is also aligned with transaction cost economics (TCE), a first cousin of agency theory (see Williamson, 1979). In Ghoshal’s (2006) words, TCE ‘assumes that individuals are self-interested and opportunistic in nature, and they will cheat the system if they can’ (p. 14). In this world view, departures from self-interest are irrational, aberrational and, ultimately, inexplicable. Shirking is therefore inevitable (Rocha and Ghoshal, 2006). On the other hand, the principal is motivated to ensure that no shirking occurs. But it is often the case that the principal cannot be sure if agents have applied maximum effort in pursuit of the goals and tasks to which they have been directed (Holmstrom, 1979). It follows that surveillance and tighter supervision is required. The tension here is between allowing agents an element of discretion – often the reason that they are chosen as agents in the first place, particularly when the principal is unsure of what their own precise interests or objectives will be (Hendry, 2002), or counter-productively eliminating the scope for such discretion by tight specification and close monitoring. Either variant is liable to incur agency costs, creating an irresolvable paradox (Shapiro, 2016). Regardless, organisations often attempt to overcome this paradox via a Sisyphean default to complex incentive and performance management/appraisal systems that make extensive use of hierarchical authority (Monks et al., 2012).

This reinforces short-term measures of financial performance since these are viewed as capturing the primary purpose of organisational activity. Organisations become viewed as ‘simply legal fictions which serve as a nexus for a set of contracting relationships among individuals’ (Jensen and Meckling, 1976: 310). This contributes to a focus on the most immediately visible and quantifiable aspects of performance, thus undermining the opportunity to identify long-term employee development and performance needs (Antonsen, 2014).

Problem 2: developmental feedback undermined by relating appraisal to pay and a close scrutiny of performance

Many performance appraisal systems attempt to serve both administrative purposes and the developmental and learning needs of employees (Kondrasuk, 2012). But driven by administrative imperatives and agency assumptions that immediate self-interest and tangible rewards are what most motivate people, organisations also use appraisal systems to determine pay increases. In the United Kingdom, the proportion of organisations linking pay to appraisal outcomes in this way rose from 15 percent in 2004 to 24 percent in 2011 (Van Wanrooy et al., 2013). The practice has also become more prevalent in the public sector (Ballantine et al., 2008). However, appraisals are simultaneously expected to address developmental goals associated with helping employees to improve their individual and collective job performance. As McGregor (1957) complained, this requires managers to play God. Employees grow reluctant to openly discuss performance problems since it may damage their pay and career prospects. This reluctance undermines the developmental, learning and supportive intentions purportedly associated with performance appraisal and which depends on trust and open two-way communication for their realisation.

Linking appraisal to pay and attendant tight supervision is consistent with the predominant agency theory assumption that there is no such thing as non-pecuniary agent motivation or that if there is, it is insignificant (Besley and Ghatak, 2005). The assumption is that people could work more effectively, efficiently and smartly if they wanted to but will usually choose not to. Consequently, close monitoring, regular feedback and complex systems of rewards and punishments are required to compensate for their deficit of motivation. This approach tends to undermine the developmental intentions that are generally held to be appraisal’s primary purpose. Thus, the evidence suggests that performance-related pay linked to appraisals does little to actually improve employee performance (Kennedy and Dresser, 2001), that it demotivates staff (Smith and Rupp, 2003), does not help retain high performers or encourage poor performers to leave and that it creates perceptions of unfairness (Varma and Stroh, 2001).

None of this should be a surprise. Studies of motivation have long suggested that close monitoring of behaviour and performance-related pay reduces the intrinsic motivation so crucial to much modern work (Heyman and Ariely, 2004). Once a premium on rewards is installed, people become less willing to engage in any form of work-related activity without them, however poor the resultant quality or their own lack of interest in the outcome. ⁶ The use of appraisal is in danger of producing further negative self-fulfilling prophecies, in that managers are encouraged to believe employees will only put in an effort if pay is closely linked to the effort required, employees become more distant and disengaged from any intrinsic investment in the effort in question and managers become even more convinced that it is only through such rewards that employees will do anything since they are ever-more inclined to demand more money and neglect those tasks not directly related to pay progression. As so often, McGregor (1960: 41–42) anticipated this argument. He observed that when managers focus on money, people respond by demanding more of it. They will also

behave exactly as we might predict – with indolence, passivity, unwillingness to accept responsibility, resistance to change, willingness to follow the demagogue, unreasonable demands for economic benefits. It seems that we may be caught in a web of our own weaving.

Among its other effects, it appears that agency theory has erased the memories of those who research, teach and practice management, condemning them to forever reinvent the wheel but immediately forget that they have just done so.

Problem 3: prioritisation of individual rather than team performance

There has been growing acknowledgement of the multiplicity of relationships that form complex organisational structures and the value of a stewardship approach in such a context (Roberson et al., 2007). Despite this, the majority of performance appraisals are conducted on an individual basis, while individual merit pay raises and bonuses are the most common form of performance-related pay (CIPD, 2009). These approaches mirror the agency notion that individual self-interest invariably trumps concern for the collective (Sen, 1994).

Yet systems thinkers argue that emphasising individual performance and self-interest to the exclusion of team performance results in an ineffective system (Seddon, 2008). It is difficult to reconcile team responsibility and commitments with an emphasis on individual responsibility. The encouragement of self-interest, through a stress on individual effort, also incentivises employees to cover up errors and inflate claims for their own performance, potentially at the expense of team performance. This negates efforts to promote organisational learning capability, despite its claimed potential to make a positive contribution to organisation performance (Camps and Luna-Arocas, 2012).

We offer an example that may resonate with many readers. Targets for research income have been introduced at one in six UK universities, at either individual or departmental level (Jump, 2015). However, the UK-based annual report of Economic and Social Research Council (ESRC) for 2014–2015 disclosed that only 1 of 21 open call business and management applications secured funding. In short, success or failure depends to some significant extent on factors beyond the control of the applicant. Paradoxically, the more academics who apply for funding – driven by targets – the lower their chances of success will be. This is a zero sum game. But they are still liable to be ‘performance managed’ if they fail to achieve their targets. Perversely, the process of appraisal encourages managers to see problems created by systems as evidence of individual weaknesses that must be ‘managed’. They therefore promote an ever more individualistic approach to how academics are managed and how they work. In a further instance of managerial amnesia, Deming (1982) warned against appraisal systems decades ago on the basis that they had precisely these detrimental effects.

Even where appraisals have been designed around team-working principles, it has been found that issues of individual gaming are still prevalent while team-working relationships are not necessarily enhanced (Toegel and Conger, 2003). Appraisal processes founded on principles of maximised performance and measurable objectives tend to ‘be self-serving, irrespective of any inclusive, team-working rhetoric’ (Tourish et al., 2010: 53). Once more, employee–manager relationships are shunted onto a track consistent with agency theory and away from building trust, loyalty and reciprocity. The benefits of this are not immediately obvious.

Problem 4: difficulty of accurately and objectively measuring performance

Ultimately, the developmental and supportive intentions behind appraisal systems that are routinely articulated in the literature rest on the assumption that assessment methods can measure employee performance with reasonable accuracy. Even scholars who advocate appraisals and the use of financial incentives within them admit that performance measures of appraisals must be both complete and accurate. Otherwise, they ‘lead to undesired behaviours’ (Shaw and Gupta, 2015: 288). Agency notions that people should be closely monitored rest on the same assumption since it is imagined that such supervision yields a more or less accurate impression of people’s work effort. This is questionable on two main grounds. First, some research into the accuracy of what are known as frequency-based assessments of behaviour in garment manufacturing plants compared estimated and actual frequencies of behaviours on the part of sewing machine operators (Deadrick and Gardner, 1997). The authors report a correlation of .59. Given a sample size of 397, this was statistically significant. However, the effect size is small. These data would appear to suggest that even when tangible and repetitious tasks are at the heart of a job, it is difficult to estimate people’s behaviour, no matter how closely they are observed. The more intangible the effort in question, the harder this will be. Moreover, the effects of close monitoring on job satisfaction, commitment and intrinsic motivation are unlikely to be positive.

Second, a problem arises from the effects of the moral hazard assumption of agency theory on the accuracy of our perceptions. The idea of moral hazard suggests that people will take risks or avoid effort if they imagine that the costs of doing so will be borne by others, such as employers rather than employees (Dembe and Boden, 2000). This is obviously often true, as in the banking crisis: in this instance, as in others, there clearly are principal–agent issues to be addressed. But moral hazard does not invariably prevail, particularly in a context of deeply embedded and long-term relationships between people. Regardless of this qualification, the notion of moral hazard is generally taken to imply that employees should be closely scrutinised. Typical of many, Milgrom and Roberts (1992) devote a chapter of a book tellingly titled ‘Economics, Organization and Management’ to employee retention strategies. As Heath (2009) critically notes, they do not ‘once mention the fact that employees sometimes feel a sense of loyalty to the firm (and that managers have it within their power to cultivate such loyalties)’ (p. 501). A predominantly economic perspective informs project scrutiny, born of the fear of deviance, which is in turn facilitated by traditional appraisals. However, behaviours that are not easily observed may offer more important indicators of effectiveness, particularly in knowledge-oriented and creative workplaces such as universities. Context-oriented behaviours such as organisational citizenship behaviour, pro-social organisational behaviour or extra-role behaviour are often intangible and not easily captured within a performance appraisal process (Organ et al., 2006). Appraisers seeking to form an overall judgement of performance must, therefore, default to other criteria or resort to the perceptual biases with which all of us are fully equipped (Hoffman et al., 2010).

Thus, the focus on what is ‘measurable’ and hence observable supports agency theory’s assumption of widespread and inevitable employee shirking, its inclination to disregard the multidimensional nature of effort and the distinctive impact different types of effort can have on performance. The ability of managers to make sound and fair judgements of all facets of employee performance becomes progressively more difficult (Wilson, 2010). Such systems shift managers’ emphasis away from creating meaning and purpose and towards a micro-management of efforts that, despite being highly visible, may be much less important for long-term success than their visibility assumes. Managers often forget that what happens backstage is as vital as what happens onstage in the production of a compelling performance.

A further difficulty from the standpoint of moral hazard is determining precisely what employees have done and hence the extent of their contribution to organisational success or failure. Complex organisational structures create multiple priorities, conflicting instructions and a proliferation of targets. Spans of control and long-distance appraisals in multinational corporations make the principal–agent relationship and close monitoring of performance subject to more errors (Holmstrom, 1982). As a result, appraisal interviews permit managers, who perhaps know less and less about an individual’s work, to determine which aspects of their performance are to be evaluated, as well as to decide the outcomes of the measurement, including pay and career progression. Consequently, the context of performance may be lost, despite its importance for the ability of any appraisal scheme to even partially achieve its objectives (Farr and Levy, 2004). However, the need for decisiveness (e.g. when appraisal schemes involve a rating scale) encourages an attitude of certainty towards evaluations when they are objectively uncertain. This makes it difficult to deliver cogent, well-informed assessments of the performance of others – the very thing on which the whole system depends. The inherently doomed attempt to deliver it turns managers into the suspicious monitors of what is obvious rather than what is important and creates employees who are increasingly distant, disengaged and defiant. Once more, the advantages of this are not immediately clear.

Problem 5: self-efficacy biases cause employees to have more favourable view of performance than their managers do

Most people do not rate their own performance as either average or below average. Rather, they exaggerate their contribution to organisational success (Rollinson and Broadfield, 2002). This is important since it suggests that an employee’s evaluation of their performance may differ from that of the manager charged with conducting an appraisal interview.

Here, agency theory once more comes into play. We have already noted that it suggests employees must be closely monitored since they will otherwise deviate from organisational goals. This in turn effects how their performance is perceived. Some experimental work suggests that the more managers monitor performance, the more likely they are to value the end product highly since they have a strong belief in their own efficacy and hence on whatever they attribute to be the outcome of their actions (Pfeffer and Cialdini, 1998). However, this does not necessarily translate into an appreciation of the contribution that employees have made to such outcomes. While the evidence on this point is mixed, and beyond the scope of our article to fully evaluate, there is some to suggest that the greater a person’s power over others, the less likely they are to interact with them and the less favourable the evaluations of their performance will be (Kipnis, 1972). After all, it can be reasoned, if they were really good at their job, they would not require such close supervision in the first place.

These problems are compounded by performance-related pay. An experimental simulation study led those in the role of supervisors to falsely believe that those whose work they were overseeing were either enjoying what they did (Intrinsic Motivation) or, alternatively, doing it only for money (Extrinsic Motivation). When they believed that money was the driver of performance, the supervisors responded by becoming more controlling. In turn, ‘employees’ in the study became more disinterested in the task. The flipside was also observed, in that those in the ‘intrinsic motivation’ condition chose to spend significantly more of their free time on the task (Pelletier and Vallerand, 1996). As Ghoshal (2006) summarised, ‘Because all behaviour (especially that which is consistent with management’s objectives) is seen by management as motivated by the controls in place, managers develop a jaundiced view of their subordinates’ (p. 24). This does not displace their confidence that these judgements are accurate, even as they diverge from those of the employees in question.

In addition, if managers imagine that the work produced under close supervision is of higher quality than that which is less closely monitored (what Pfeffer and Cialdini (1998) call ‘the illusion of influence’), and that employees have, therefore, performed well, it follows that ever-tighter monitoring would confer even more benefits. Thus, good performance may be seen as occurring in spite of the attributes of the people involved. Their successes can instead be credited to the system of surveillance which, it is imagined, has reigned in their tendency to behave deviantly. In addition to the harmful effects of this on intrinsic motivation and the quality of work, there is the risk that ever tighter monitoring becomes a form of ‘petty tyranny’ and so triggers low self-esteem; damages performance; weakens work unit cohesiveness; and produces higher levels of frustration, stress, reactance, helplessness and work alienation (Ashforth, 1994), all of which undermine people’s capacity to learn.

Of course, these are not the only problems here. But we are stressing those that arise from close supervision, driven by agency theory reasoning, and its ensuing negative impact on managers’ perceptions of the work of others. In aggregate, it means that during formal appraisal interviews, managers are often reduced to informing employees that their performance is weaker than what they themselves imagine it to be. The detrimental effects can be readily imagined. Even if attempts are made to ‘compensate’ for this by also providing positive feedback, a great deal of research has shown that ‘bad’ events, such as critical feedback, are much more powerful and memorable than those events regarded as good (Baumeister et al., 2001). Consistent with this, in the context of appraisal, DeNisi (1996) found that 75 percent of employees saw the evaluations they received as less favourable than their own self-estimates and, therefore, regarded appraisal interviews as a deflating experience. The numerous biases that managers themselves bring to the process of appraisal intensify this problem.

In line with the anxieties voiced by Ghoshal (2005) and Pfeffer (2005), these outcomes may well activate a large number of destructive self-fulfilling prophecies, whereby negative feedback creates resentment, places obstacles in the path of personal development and diminishes rather than enhances effectiveness. It equates to an instrumentalist view of the employment relationship whereby employees are viewed as a ‘resource’ to be employed or discarded on the basis of their short-term performance (Tourish et al., 2010). Moreover, close monitoring exacerbates the power relationships existent in traditional organisational hierarchies and endorsed by agency theory. Such a context only highlights the pervasive influence of agency relationships in the appraisal process and the role they play in maintaining structural power inequalities between managers and employees. Overall, it supports the conclusion that the (mal)practice of performance appraisal is intimately informed by the assumptions of agency theory and constitutes a further example of how this theory leads to what we view as bad management practice.

Discussion

While advice for practitioners on how to ‘improve’ appraisal is plentiful, it is clear that the ‘creation of a successful performance appraisal system remains largely an unrealised goal’ (Gordon and Stewart, 2009: 274). In problematising conventional perspectives on this issue, we argue that agency theory has influenced the implementation of performance appraisals, by virtue of the deeper traction it exerts within management theory and hence on the ideological context in which management is practised. We have argued that dominant assumptions within agency theory of economic rationality, self-interest and moral hazard have a negative impact on how performance appraisal systems are misused in many organisations. In doing so, we suggest that appraisals constitute a prime example of how a theory can damage learning and contribute to bad management practice. Their continued popularity is a classic instance of hope triumphing over experience. It owes little to any inherent utility. Therefore, we extend Foucauldian approaches by showing how agency theory perpetuates particular forms of power relationships in organisations, enacted through often questioned but astonishingly resilient HR practices, such as appraisals. The very ubiquity of appraisals has given them a ‘naturalised’ status in the minds, lexicon and practices of managers and researchers, even as their actual effects are deplored.

Managers are often only too well aware of these issues but face two key problems in addressing them. First, as Mintzberg (2009) has reminded us, management work is unrelenting, orientated to action, fragmented and full of interruptions. The time for reflection, including reading, is minimal. No wonder that folklore, tradition and the casual imitation of what others do frequently triumphs over a careful study of the evidence (Pfeffer and Sutton, 2006). Second, managers are encouraged to keep hoping that things will improve. Key texts promoted by business schools recognise that ‘good intentions in the PA area have often been associated with disappointing outcomes’ (Boxall and Purcell, 2003: 145) but still go on to assume that adjustments will enable managers to ensure that ‘formal PA systems reach more of their potential’ (p. 146). The problem is that each such fix has unintended consequences. Keen to ease their pain, managers look for yet more fixes. And they have an eager supplier – in the form of the HR industry. In contrast, our article denaturalises appraisals. We seek to situate debate on the issue in a deeper appreciation of the power-saturated and ideological contexts in which they are implemented. This suggests that the more formalised, ritualised and bureaucratised the process of appraisal becomes, the less helpful and more damaging it is likely to be. Perhaps, it is useful to recall the injunction of the Hippocratic School: ‘First, do no harm’. If appraisals risk doing more harm than good, perhaps, we should suggest abolishing them. Surely, by now, there have been enough attempts to fix the unfixable?

Conclusion

In the end, we return to Ghoshal (2006: 42). He advocated an alternative perspective which recognises that

… the advantage of organizations over markets may lie not in overcoming human pathologies through hierarchy, but in leveraging the human ability to take initiative, to cooperate, and to learn; it may also rely on exploiting the organization’s internalised purpose and diversity to enhance both learning and its use in creating purposeful and innovative adaptation.

We do not question the value of regular, informal communication and two-way feedback. But to be effective, these need to be informed by different values, based on trust and a diminution of power differentials within the workplace. By contrast, conventional appraisals prioritise hierarchy over intrinsic motivation, distrust over trust, and the importance of individual effort over that of building sustainable, co-operative systems. Without a major rethink along the lines advocated in this article, they will continue to blight relationships in the workplace far into the future.