Merit and Power

Australian Joumal of Social Issues, Vol. 22, No. 2, May 1987, pp. 436-450
pdf of published article

The merit principle, used as a way of removing biases in occupational structures, is flawed. The evaluation of merit is tied to the interests of those in positions of power, so that application of the merit principle is compatible with continued structural inequalities. In practice the concept of merit is used as a resource in organisational power struggles. A deeper problem with the merit principle is its acceptance of occupational hierarchies. A more fundamental equal opportunity requires some form of self-management.

Paradoxically, the concept of merit has become a catchphrase of both supporters and opponents of legislated equal employment opportunity. On the one hand, advocates of women's and other minority rights have supported application of the merit principle, underpinned by affirmative action to remove unfair starting handicaps, to provide real equal employment opportunity. On the other hand, traditional bureaucratic elites have invoked the merit principle to oppose what they allege is reverse discrimination against white middle-class males. In the struggle over equal opportunity, the concept of merit is being used by both sides as a tool. To understand how this comes about, it is necessary to examine the power structures of organisations.

The merit principle is most commonly applied in bureaucratic or semi-bureaucratic organisations. These organisations are constructed most fundamentally on hierarchy and a division of labour. The classical picture sees bureaucracy as an instrument for rational administration, operating on the basis of universalistic principles (Weber 1964). Seniority is commonly used as a basis for advancement in a bureaucracy simply because it is easy to measure and because it supposedly shows commitment to the organisation. But seniority does not seem to be a suitable principle for maximising organisational performance. Merit, an assessment of a person's ability, experience and motivation relevant to performing a job, would seem to be a more suitable principle for measuring and rewarding performance in an ideal bureaucracy.

What is merit in the conventional theory of bureaucracy? 'Ability plus effort' is a typical formula. There are three standard ways of assessing merit: examinations, credentials and work experience and performance. Examinations are often used in recruitment into bureaucracies, but much less commonly in promotion. Credentials are especially important in professions: credentials are a prerequisite for entry to law, medicine, engineering and many other occupations. It is commonly argued that credentials demonstrate one's ability or suitability to undertake a job in bureaucracies as well. Finally, work experience and performance provide indicators of merit, especially in demonstrating whether abilities shown by examinations or credentials are actually utilised on the job.

In practice the measures of examinations, credentials and performance are mixed together in the assessment of merit. Credentials become important, for example, when professional and bureaucratic systems of social organisation are intertwined, as they often are. In hospitals, advancement through the organisational hierarchy usually depends on having a medical degree as well as experience in administration. In many high-technology companies, degrees in engineering or science provide a basis for rising into the management stream. To become head of a school a background as a teacher, implying possession of suitable teaching credentials, is usually a prerequisite. In these and other cases there is commonly a parallel 'administrative stream' for those without professional credentials. This administrative stream may or may not provide access to the highest levels of the organisation, depending on the relationship of the professional system compared to the bureaucratic system.

The standard picture of the role of merit then is as a means of assessing a person's suitability for a job in a bureaucratic and/or professional system. The question which needs to be faced by supporters of equal opportunity is whether merit can indeed be applied in an unbiased manner, or at least a relatively unbiased manner. If this cannot be done in practice, it may not be wise to rely too heavily on merit in pursuing the project of overcoming discrimination in employment. To approach this issue, I will look more closely and critically at the way in which power is exercised in large organisations.

I will begin by arguing that the struggle for advancement within complex organisations is part of a wider power struggle, in which resources provided by white middle-class male culture are crucial. Since work is embedded in the organisational context, evaluation of merit is shot through with the biases of those in positions of power as well as biases in the very definition of merit in a system which imposes unequal life opportunities. The usual definition of merit within organisations may seem fair, and be applied fairly within its own terms, while being entirely compatible with the continuance of structural inequalities such as male dominance. Merit can never be unambiguously assessed, but is rather one tool among others in organisational power struggles.

Going beyond this, a fundamental flaw in the merit principle lies in its acceptance of occupational hierarchies in which the decisions of the few control the work of the many. Rather than trying to equalise opportunity within a hierarchical system - which at most can change the individuals who occupy the slots - a more fundamental 'equal opportunity' would involve redesigned occupational structures in which workers defined the conditions and goals of their work. This would allow women, the disabled and others to be involved in an equal way by restructuring the hitherto distinct and unequal arenas of public and private life.

The Political Perspective on Organisations

Traditional views on bureaucracies and professions have been challenged in recent years by scholars who see these as political systems. First consider bureaucracy. The traditional Weberian picture of bureaucracy as a rational system of organisation has strongly influenced much organisational theory, which concentrates on how management can overcome obstacles to efficiency. Nevertheless, it is widely recognised that actual bureaucracies do not behave precisely according to Weber's ideal-typical picture. There is a constant struggle for power within bureaucracies, a process which includes power relations embodied in interpersonal discourse, competition between internal factions, conflicts between bureaucrats and clients, and struggles between bureaucracies.

Rejecting the Weberian assumptions is the political perspective on organisations presented by Deena Weinstein (1977,1979), according to which bureaucracies are best understood as political systems. Individuals and groups within bureaucracies exercise power using the particular social and material resources available, including control over labour power, positions, clients, funding and knowledge. Overt political activity by workers is commonly inhibited by fear of reprisals and by habits of acquiescence, often based on beliefs or fear.

The key role of power is most obvious when oppositions form within bureaucracies, either individual dissidents or whole movements. What usually happens is not a rational assessment of the claims of the dissidents, but rather suppression of the dissidents themselves: corporation executives who speak out critically about their employer's behaviour typically will be sacked; soldiers who rebel will be court-martialed; rank-and-file revolts by trade unionists may be smashed by expulsions (Weinstein 1977, 1979; see also Anderson et al. 1980; Perrucci et al. 1980; Zald & Berger 1978).

Weinstein likens bureaucracies to authoritarian states in their political structure in that control is exercised from the top with few opportunities for popular participation. Democratic structures are token or non-existent, and the elites control the main political and economic resources. The main difference between bureaucracies and authoritarian states is that normally bureaucratic elites cannot use violence to enforce their control. One of the consequences of this power struggle is that the goal of the bureaucracy may be displaced from outer service to inner control (Hummel 1977). This picture does not assume that bureaucracies are monolithic structures, nor that the only political power exercised is strictly according to formal structures. Indeed, it is the existence of oppositions which shows that power is based on a set of practices which do not accord with formal rationality.

A similar analysis of professions has been developed in recent years. Contrary to the traditional picture of professionals using their knowledge and skills purely in the service of society, this analysis sees occupational groups using knowledge and skills as resources to increase their power, wealth and status. Doctors for example use credentialing and licensing systems to keep down numbers in the medical profession and then use their control over the resulting scarce resource of 'legitimate' expertise to push for higher salaries and control over medical decision-making (Freidson 1970; Johnson 1972; Larson 1977; Parkin 1979; Willis 1983; for a critique see Saks 1983). Opposition movements also arise in professions. The medical profession has routinely taken action to deny status and resources to challengers such as practitioners of 'fringe therapies', for example by lobbying to deny their inclusion under medical benefits schedules.

The political picture of bureaucracy and professions provides a useful way to begin to analyse their interactions with systems based on class, gender and ethnicity, which can be treated as power systems which cannot be reduced to each other nor to some other essence (Brittan & Maynard 1984). Class, gender and ethnicity in this picture are resources which can be deployed in bureaucratic and professional struggles, and conversely bureaucratic and professional power are resources which can be deployed in struggles involving class, gender and ethnicity.

For example, there are various ways in which male domination is maintained in bureaucracies and professions. Overt discrimination against women is the most obvious way. This is made possible by control of the top decision-making positions by men. Secondly, external support systems for employees are biased against women. Since women continue to carry most of the burden of housework and child care, most women employees are doing 'double duty' compared to men, especially those who have the support of wives for these tasks. Thirdly, organisational styles usually favour those with personal characteristics which are typically masculine, such as aggressiveness and competitiveness. Women who do not behave this way are not deemed as suitable material for promotion, while women who do behave this way are denigrated as not behaving appropriately for their sex. Finally, the gender division of labour excludes most women from opportunities to rise in organisational or professional hierarchies. Similar mechanisms are used to limit the advancement of members of ethnic minorities, those who maintain links with working class culture, and those who take threatening political stands.

This example of the meshing of bureaucracy and patriarchy is silent about the precise mechanisms of domination and group cohesion within bureaucratic and patriarchal systems. One possible analysis, avoiding the resort to a dominant ideology thesis (Abercrombie et al. 1980), could be based on the material interests of most men in patriarchal systems and of most bureaucratic elites in their positions. These elites, for example, mobilise the support of other men, lower in the job hierarchy, by supporting them against the challenge to jobs or status by women. This can explain management's promotion or at least acceptance of a constantly renegotiated gender division of labour (Game & Pringle 1983) which keeps women in the inferior jobs at the cost of overall efficiency.

How can the conventional application of the merit principle be compatible with the sort of bias consequent on the routine exercise of power to serve group interests? Let me look in turn at examinations, credentials and job performance.

The strengths and limitations of examinations in assessing merit have been studied in some depth. Their advantages include uniformity and anonymity (of those taking an examination to those marking it). But this does not remove the possibility of cultural bias. Examinations usually select out those schooled in the 'cultural mainstream' and also those skilled in taking examinations, which is again an indication of success in the dominant culture within the schooling system.

Studies have shown the class bias in IQ tests (Kamin 1974). The IQ, which is widely believed to signify innate ability, is not able to explain educational and economic inequalities (Bowles & Gintis 1976). A few decades ago, only a relatively small fraction of children completed high school and therefore those who competed for professional and managerial jobs were a select group, mostly from better-off families. As progress through high school became more universal, it seemed that everyone would be able to share in the 'common culture' sufficiently to obtain equal opportunity through examinations. But as schooling has become universalised, its focus has changed from preparation for elite careers to inculcating middle-class culture (Collins 1979) - something which works better for those with the appropriate class background and family support. Working class boys, for example, typically resist this process (Willis 1977). In any case, the outcome is culturally biased: those with the appropriate cultural background will learn most readily and be able to show their skill on ostensibly neutral examinations.

Another problem with examinations is that most of them have limited relevance to the job, since jobs are quite different from answering questions. The only really relevant examination is an 'on-the-job' test, which displaces the assessment of merit to job performance.

Even if these problems with examinations could be overcome, there is the problem of their limited use. Typically, examinations are used as a screening mechanism in recruitment, but once inside a bureaucracy, tenure is reasonably secure. A true merit system would have regular open competitions, in which any person could vie for any post. In only a few occupations, such as competitive sports and the performing arts, is there a continuous and public examination of talent and performance which affects a person's standing and career. Significantly, these occupations are notable for their absence of formal examinations, and for the relative ease with which individuals can establish themselves without credentials.

Credentials are a second major way of measuring merit. Credentials, like examinations, commonly have little relevance to the job. Most of the skills for professional and managerial positions are learned on the job (Collins 1979). The role of formal education in reproducing and legitimating the social class structure has been widely documented (Bowles & Gintis 1976). Credentials are the symbols and currency of competence, but numerous studies have shown that there is little connection between educational performance and job performance. I have talked to many students who comment, 'this course has nothing to do with my job, but I need the degree to get a promotion.'

The use of credentials to judge merit is not just to rely on irrelevant measures. Rather, the cultural and gender biases in educational systems serve as a filter which legitimates unequal outcomes. The social biases inherent in credentialing are most pronounced in poor countries such as Kenya or Sri Lanka in which qualifications are the means by which individuals can join the privileged western sector of government employment. Students desperately cram down irrelevant information in an attempt to snare one of the few high-paying jobs at the end. The result is a massively wasteful educational system, irrelevant to the needs of the bulk of the population, which serves to justify and promote the imposition of an inappropriate western model of development. The problems with credentialing are not so dramatic in richer countries, but many of the same problems exist (Dore 1976).

There are problems with using examinations or credentials to assess merit. What then about job performance? The first point here is that it is very hard to separate out the performance of an individual worker within a bureaucracy (Collins 1979; Emery & Emery 1974). Work is embedded in the organisational context. It is done within overall policies and procedures set from above. It requires cooperation with other people and resources provided by other groups. Above all, performance depends on working as part of an overall system, on making the whole operation go along more smoothly. Who is to say what was a particular individual's contribution?

This question is a key one. The answer of course is that it is those people in the top positions who evaluate performance. This gives them the opportunity to judge on the basis of conformity to the cultural norms of the organisation. The concept of 'political labour' is a valuable one here (Collins 1979). Much of the activity in bureaucracies and professions, especially in the higher echelons, consists of interacting with others through conversations, committee meetings, lunches and so forth. What is happening is a continual process of 'social negotiation' over preferences, policies and alignments. The status of any individual is at stake in this process. Those who play the game poorly, or opt out entirely, are liable to lose out in some reshuffle or power play. Others are able to forge alliances and mobilise resources to promote the interests of themselves and those similar to them (Kanter 1977). The problem here is that political labour, or in other words playing the game within organisations, is something which depends sensitively on possessing the right cultural attributes and resources. Verbal skill, personal attributes and social compatibility are important.

The importance of political labour means that it is difficult for cultural outsiders to get ahead. This mean women, ethnic minorities, nonconformists or indeed anyone falling outside the standard mould in a particular organisation. Of course, organisations vary in their cultural style. In some cases a particular group, such as graduates of a particular university, constitute the in-group. In other cases personal attributes, such as a hard-driving entrepreneurship - which can be closely linked to gender and class background - may be most important. The point is that the cultural milieu of the organisation is not neutral, but rather favours some groups over others.

A further problem is that most assessments of merit are confounded with simultaneous assessments of 'loyalty' and 'fitting in': whether a person gets along with workmates and superiors, is unlikely to rock the boat, associates with the right people, etc. This type of conformity is especially prized by organisational elites, since an uncommitted employee is a potential leader of or trigger for wider opposition. In evaluations of workers, loyalty in this sense is typically a prime consideration, often completely overwhelming the formal merits of a person who is perceived as difficult. Convincing people of one's suitability is of course a crucial part of political labour, and is inextricably linked with gender, class, ethnicity and other ascribed characteristics.

This point about political labour can be made clearer by looking at the performance of academics, who work in a more professionalised and less bureaucratised system in which achievement of individuals can be judged according to teaching and research. Teaching can be ignored here, since it is rarely judged and is not treated as important in career advancement. Research performance on the other hand is supposedly quite important, and can be judged by the public record of publications.

In practice, research productivity is not nearly so important in career advancement as commonly believed. For example, Lionel S. Lewis has shown that a much more significant factor is years of service (Lewis 1975; Lewis & Doyno 1983; Lewis & Gregorio 1984; Lewis et al. 1979). How can the apparently objective record of research be compatible with biases arising from systems of power? There are several factors.

Firstly, in research teams, the supervisor or team leader often takes undue credit for the work of subordinates. This is even more common in bureaucracies. Among public servants I have talked to, it is commonly taken as a fact of life that some superiors will claim or attempt to claim credit for the work of subordinates. Secondly, credit for research work depends on being in the right place at the right time. The same contribution is likely to be given much more attention and credit if the author is at the right university and known by the key people in the field. Those scholars who are too young, too old, women, in minor institutions or in the wrong field will be ignored. Work is judged according to who did it, not solely on the basis of the work itself (Caplow & McGee, 1958).

Thirdly, people look almost solely at research outputs, and not at the inputs to achieve them. Many male academics receive major career support from their wives (Papanek 1973). Young men who have had full support from wives are deemed more successful than somewhat older women with the same record of performance who have devoted significant time to housework and child rearing. Fourthly, appointments and promotions depend only partly on research performance. Also important are fitting in socially, working in a suitable speciality and having credentials from the right places. Finally, dissidents - those who do research or teaching or make public statements that challenge the interests of powerful groups, especially within academia itself - often encounter severe difficulties in their careers (Martin et al. 1986).

What is in operation is a highly biased system in which particularistic practices are justified by the merit principle. The widespread belief that publication is the key to success in academia testifies to the importance that merit has in legitimating academic position and power. This is compatible with system bias through the means noted above: performance is attributed to individuals even when much credit is due to subordinates or spouses, and perceptions of performance are shot through with distortions due to the current distribution of status and power.

There is really little hope of overcoming these biases by a 'true' application of the merit principle, because there are no neutral judges who sit outside the academic system. For example, judgements of academic merit partly depend on how central a contribution is to the recognised core of the discipline. This would not be a problem, except that judgements of what the 'core' is depend on the outcome of continual power struggle between individuals and groups who have their own status and power at stake in the struggle. Performing successfully means adapting to the conception of the discipline which is promulgated by the most powerful figures and groups in the discipline. Of course, these individuals and groups promote a vision which puts them in the prime positions. In other words, the assessment of academic merit within a discipline depends on the outcome of a power struggle in which the criteria of merit itself are a key stake.

All this is quite compatible with continued bias against women, ethnic minorities, members of the working class and so forth. For example, scholars who choose to develop forms of feminist analysis are almost always seen as marginal to an academic discipline. The elites of the discipline define the essence of its study in male terms or simply in terms which they, who happen to be men, also happen to control.

The importance of power struggles in academia, in which merit is a tool rather than an objective assessment, throws light on struggles in more bureaucratised organisations. Performance within bureaucracies is much more embedded in the organisational context than it is in academia. Unlike academia, it is not easy to appeal to experts in other organisations to judge performance by assessing publications. The good graces of supervisors, the cooperation of co-workers and the availability of support systems for non-work life are all vital.

All this means that performance must be judged by a 'merit system' which includes an enlightened evaluation of these factors. But good intentions and enlightened assessments are hardly sufficient to establish a merit system. As long as the organisational resources are available to be used in power struggles, then they will be used to maintain and extend the power of those who are best able to deploy them.

Within bureaucracies, the Weberian model of rational administration is itself a powerful rhetorical and ideological resource for maintaining commitment to the organisation. The belief in bureaucracy as administration is favoured and promoted especially by those who are successful in the system, since it attributes that success to individual performance rather than to being on the winning side of a power struggle reliant on cultural resources.

In summary, individual productivity has a relatively low impact on career performance in large-scale organisations. Advancement depends greatly on jockeying for position, for example by forming alliances, obtaining personal credit for collective work and increasing the evaluation of particular types of contributions. In this struggle for position, resources provided by culture - namely the dominant white middle-class male culture - are vital.

While the reality is that the struggle for advancement is basically part of a wider power struggle, it is no longer acceptable to admit the importance of system biases in an era in which egalitarian sentiments are strong. Hence it is very convenient to justify decisions on the basis of merit - for example academic credentials - because this legitimates inequality. This is the basis of the defence of 'merit' by traditionalists who in practice oppose changes in the gender and ethnic composition of hierarchies. The supporters of legislated equal opportunity on the other hand recognise the biases built into normal selection processes and therefore push for 'true' adherence to the merit principle, in other words merit stripped of its links to gender stereotypes, middle-class culture and so forth. This is the explanation of the paradox of merit: that both opponents and proponents of equal opportunity make appeals to it.

What then should the harried equal opportunity officer or personnel manager do with the merit principle? Is it simply a tool of those already in power? Or is it, as usually thought, a tool of those who are actually achieving valuable things in their work? The answer is that merit is a tool for both groups at the same time.

To recapitulate: under the ideal-typical conception of bureaucracy as a system of rational administration, merit can be used unambiguously as a measure of performance, and affirmative action then can be used to assure true equal opportunity. I have argued that the picture of bureaucracy and professions as political systems is more realistic. In this picture, 'merit' is not something judged in the abstract, but rather is something continually created, debated and negotiated through the social processes of the organisation. Those individuals and groups with the greatest power are in the best position to shape the conception of merit to serve their own ends. But, at the same time, others with less power can argue for a different interpretation of merit which promotes their interests.

In this view, what is called merit is a tool in the power struggles within bureaucracies and professions. Those promoting the interests of disadvantaged groups should not assume that an argument based on merit will be taken on face value. Rather, an argument promoting the interests of disadvantaged individuals or groups, based on claims about merit, is part of a political move to change the distribution of power, in either a minor or major way. It should be expected that some of the merit arguments will be contested, less by direct rejection than by a different interpretation of merit. The change in general definitions of merit over the years (Stanton 1978) - for example from legislation against discrimination to affirmative action, and then to comparable work - is a consequence of the challenges mounted by blacks and women. Interpretations of merit are intrinsically political, just as are bureaucracies and professions, in the sense that power is always involved rather than just rationality and formal procedures.

The Division of Labour

If merit is a tool or resource used by different groups in struggles within bureaucracies and professions, the question arises, who can most easily use this tool? Tools, whether butter knives, nuclear weapons or ideas, are not equally useful for performing tasks. Butter knives are more suited to cutting butter than are chain saws. The idea of merit in many circumstances is of great value to disadvantaged groups: the most blatant forms of discrimination are harder to sustain. But merit is not a perfect tool for creating equal opportunity: it has a fundamental flaw.

The most massive barrier to application of the merit principle in bureaucracies is the division of labour. It is virtually impossible for a cleaner or a steno-secretary to move into top management - or even to move from her position - no matter how well she does her job. In any situation where there are large differences between jobs, the problem arises that people can't be judged as unsuitable until they have been given a chance in the other job. The incongruities involved in actually putting people into jobs of completely different status forms the basis of the humour in the story of a millionaire and a pauper who change life situations, for example in the film Trading Places.

in some bureaucracies a limited degree of job rotation is practised. This certainly has the advantage of allowing individuals to show their ability in a variety of positions. But usually job rotation is only carried out with tasks on a similar level. Managing directors are not rotated to be gardeners, nor vice versa. In practice, many workers are refused any real chance to move out of dead-end positions.

The main way this is justified is through appeal to the merit principle! For example, typists are refused opportunities to undertake research positions because they have neither the right credentials nor appropriate experience.

The usual pattern then is one in which credentials and experience are used to justify appointments and promotions. The problem is that the credentials usually have little relevance to the job, and previous experience says little about a person's ability to handle a different type of job. One example is teaching. The theoretical, classroom part of teacher training is notorious for being irrelevant to actual teaching. (Apprentice teaching, essentially learning on the job, is literally the exception which proves the rule.) Once on the job, some teachers are promoted to administrative positions. There is no logical reason why excellence in teaching would be a basis for promotion to administration, even if excellence in teaching were taken as the criterion of merit (male ambition seems a more common reason for promotion among teachers).

This brings me to what I believe is a fundamental flaw in the usual conception of the merit principle. The application of 'merit' assumes not only that individual achievement can be measured, but also that the existing occupational slots are suitable for the allocation of meritorious workers to them. In other words, the merit principle assumes that the slots are fixed while people are chosen for them - not vice versa.

The difficulties in this can be illustrated by a not-so-hypothetical example. Assume that there are 10 people working in a section. They are all literate and have an average potential for learning. In a bureaucratic system, the 10 people are ranked 1 to 10: one person administers from the top while those at the bottom do routine work requiring little or no initiative. The person at the top develops a broad view of the operations, hobnobs with corresponding figures in other sections, obtains inside information about the organisation and about the other workers, and gains increasing confidence and self-assurance through the exercise of power and from the deference of subordinates. The point of this is that the initial distribution of workers into slots could almost have been done at random, and the result would have been the same. The power system creates the very differences in information, confidence and performance which are used to justify it.

In practice, the workers are not allocated at random to different slots. White middle-class men hold most of the positions of power, and they use that power to maintain themselves and those like them in the top positions. The merit principle helps to sustain this system by justifying it; the merit principle may be the best justification yet for a highly stratified society (Young 1958).

But to be fair, it is not the merit principle which lies behind stratification. The most important support for system bias is the hierarchy itself. Resources are given to small groups which they can use to maintain their privilege. It is inevitable that groups with more cultural resources - whether this is maleness or ethnic dominance - will use those resources to gain organisational power, and use organisational power to promote group interests. While insisting on the application of a 'true' merit principle has the potential to make inroads into this system, a more far-reaching approach is to restructure the job hierarchy and division of labour.

The standard solution is worker self-management, in which workers collectively decide on how they will organise their work and carry it out. There can still be a division of labour, but it is one that is freely chosen (and it is likely to be a division that changes more or less frequently). Worker self-management needs to be accompanied by job redesign, in which the technology and the organisational decision-making process are set up in a way which fosters equality and participation by all workers (Bolweg 1976; Emery & Thorsrud 1976; Herbst 1976; Hunnius et al. 1973; Roberts 1973; Wilson 1974).

I have claimed that the power systems in bureaucracies and professions generate differentials in information, confidence and performance. What is the evidence for this claim? Actually, the best evidence comes from experiences in worker self-management, in which workers are given the chance to demonstrate their capacity in a supportive environment. Most people are able to learn the complex skills of driving a car or managing a household. Similarly, most people should be able to learn carpentry and learn to manage a carpentry business.

What would merit mean in a workplace in which the workers collectively controlled their work? In such a situation, merit could be recognised even though performance was embedded in the work organisation. For example, some workers might be able to do certain jobs quicker or more elegantly. Indeed, it would be easier to assess merit in many cases since different workers would be regularly doing the same tasks, for example on a rotating basis. It would also be easier to assess merit because many of those doing the assessing would have recently done the same tasks themselves.

Let me now compare the definitions of merit under hierarchy and self-management. The usual definition is built around picking the individual who will perform 'best' in a given job situation. If meritocracy is justified on the basis of its greater productivity, this usual definition is a micro principle: merit considerations apply to each job assignment in isolation from other jobs. A slightly wider picture arises from the aim of maximising productivity from a whole ensemble of jobs. People are assigned to jobs in a way which maximises productivity for the full set of jobs. This macro principle of merit on occasion may violate the micro principle. The macro principle again assumes a fixed distribution of jobs (Daniels 1978).

Under self-management, jobs are not fixed. Rather, the design principles for work are themselves something to be democratically decided. Under the project of socio-technical design (Herbst 1974), merit is just as usefully attached to the organisation of work as to individuals. The most 'meritorious work organisation' is the one which maximises the performance, satisfaction and personal growth of all individuals, within the constraints of current knowledge and resources. (Some might prefer a maximin principle in which the aim is to maximise, relative to feasible targets, the performance, satisfaction and personal growth of the person least advanced in these areas. The maximin principle is designed to avoid a utilitarian maximisation which perpetuates inequalities.) For example, socio-technical design typically implies elimination or automation of jobs which are satisfying to no one, rather than automation on the basis of profits and material productivity.

Although the introduction of self-management is often associated with lowering productivity, actually the converse is often the case (Melman 1958). Self-management can increase conventional productivity - not to mention worker satisfaction - because the capacities and enthusiasm of workers are more effectively used. Another way to look at this is that hierarchical structures carry a burden of inefficiency, namely the managerial and technological expense and organisational rigidity required to keep workers in their place (Braverman 1974). As often as not, the productivity loss required to maintain hierarchies is greater than the loss entailed in self-managerial decision-making processes.

One other point is worth making: there is no reason why merit and a more responsible position should automatically lead to greater rewards for a person (Daniels 1978; Luard 1979, pp. 89-75). Under a meritocracy justified by productivity, the only differentials in rewards would be those necessary to motivate workers, and these would be far less than those in present bureaucracies. According to the political perspective on bureaucracy and professions, the reason why merit and reward are conventionally linked is that those with power use that power to increase their income. Where the medical profession is powerful, as in the United States and Australia, average incomes of doctors are high. Where medical workers are more subordinate to the state, as in the Soviet Union, their average incomes are closer to the average wage.

Under self-management, there would be no reason for meritorious performance to lead to greater formal power to tell other workers what to do, nor to gain disproportionate economic rewards. That would not be compatible with the principle of self-management, and in any case it would not be necessary. Those workers who worked harder or came up with better ideas would be naturally recognised by the others for the value of their contributions. The reward for good performance would be the performance itself, more influence (since others would respect the person's views) and more recognition.

Self-management is not an alternative sitting on the shelf waiting to be applied should the decision be made. Rather, the operation of self-managing systems would depend on the political processes used to establish and maintain them, just as in the case of the operation of bureaucracies and professions. Still to be further studied and developed through practice are the limits of consensus decision-making (Mansbridge 1980), the role of community control (Morris & Hess 1975) and the nature of wider political decision-making (Benello & Roussopoulos 1971). Self-management is part of a particular strategy for social change. For example, whereas many liberal feminists use the merit principle to help the cause of women within bureaucracies, some radical feminists favour self-management as an alternative to bureaucracy itself (Ferguson 1984).

The alternative of worker self-management throws light on the assumptions underlying the usual use of the merit principle. Applied to advancement within a bureaucratic system, merit is inevitably linked to the exercise of power. The evaluation of merit is shaped by the current elites and is also used to justify the hierarchy itself. Those pushing for equal opportunity through application of 'true' merit have an uphill battle because they must accept the hierarchy and division of labour which give power to those groups with the greatest cultural resources. Merit in this situation can never be a neutral principle to judge contenders in the occupational power struggle. Rather, merit is a tool in the power struggle itself.

For those who are pushing for rights for women, ethnic minorities, the handicapped and so forth, it is important to realise that using merit as a lever for the advancement of individuals has important limits. It does nothing for the bulk of workers who are relegated to dead-end jobs and occupations, and it may help to legitimate the whole system of inequality and privilege. Instead of applying the merit principle only to individuals, it should also be applied to work structures. Instead of workers being judged according to how they fulfil fixed job specifications, work structures should be evaluated according to how they allow the greatest use and development of each person's potential contribution. The challenge is to turn this vision into a practical program for organisational change (Williams 1982).

Acknowledgements

Hester Eisenstein and two anonymous referees provided valuable comments on an earlier version of this paper.

References

Abercrombie, N., Hill, S. and Turner, B.S. (1980), The Dominant Ideology Thesis, London: Allen and Unwin.

Anderson, R.M., Perrucci, R., Schendel, D.E. and Trachtman, L.E. (1980), Divided Loyalties: Whistle-blowing at BART, West Lafayette, Indiana: Purdue University.

Benello, C.G. and Roussopoulos, D. (eds.) (1971), The Case for Participatory Democracy, New York: Grossman.

Bolweg, J. (1976), Job Design and Industrial Democracy, Leiden: Martinus Nijhoff.

Bowles, S. and Gintis, H. (1976), Schooling in Capitalist America, New York: Basic Books.

Braverman, H. (1974), Labor and Monopoly Capital, New York: Monthly Review Press.

Brittan, A. and Maynard, M. (1984), Sexism, Racism and Oppression, Oxford: Blackwell.

Caplow, T. and McGee, R.J. (1958), The Academic Marketplace, New York: Basic Books.

Collins, R. (1979), The Credential Society, New York: Academic Press.

Daniels, N. (1978), 'Merit and Meritocracy', Philosophy and Public Affairs, 7, 206-223.

Dore, R. (1976), The Diploma Disease: Education, Qualification and Development, London: George Allen and Unwin.

Emery, F.E. and Emery, M. (1974), Participative Design, Canberra: Centre for Continuing Education, Australian National University.

Emery, F. and Thorsrud, E. (1976), Democracy at Work, Leiden: Martinus Nijhoff.

Ferguson, K.E. (1984), The Feminist Case Against Bureaucracy, Philadelphia: Temple University Press.

Freidson, E. (1970), Professional Dominance: The Social Structure of Medical Care, New York: Atherton.

Game, A. and Pringle, R. (1983), Gender at Work, Sydney: Allen and Unwin.

Herbst, P.G. (1974), Socio-technical Design, London: Tavistock.

Herbst, P.G. (1976), Alternatives to Hierarchies, Leiden: Martinus Nijhoff.

Hummel, R.P. (1977), The Bureaucratic Experience, New York: St Martin's Press.

Hunnius, G., Garson, G.D. and Case, J. (eds.) (1973), Workers' Control, New York: Vintage.

Johnson, T.J. (1972), Professions and Power, London: Macmillan.

Kamin, L.J. (1974), The Science and Politics of IQ, Potomac, Maryland: Lawrence Erlbaum.

Kanter, R.M. (1977), Men and Women of the Corporation, New York: Basic Books.

Larson, M.S. (1977), The Rise of Professionalism: A Sociological Analysis, Berkeley: University of California Press.

Lewis, L.S. (1975), Scaling the Ivory Tower: Merit and its Limits in Academic Careers, Baltimore: Johns Hopkins University Press.

Lewis, L.S. and Doyno, V. (1983), 'The Definition of Academic Merit', Higher Education, 12, 707-719.

Lewis, L.S. and Gregorio, D.I. (1984), 'University Teaching and the Assessment of Merit', Teaching Sociology, 12, 32-46.

Lewis, L.S., Wanner, R.A. and Gregorio, D.I. (1979), 'Performance and Salary Attainment in Academia', American Sociologist, 14, 157-169.

Luard, E. (1979), Socialism Without the State, London: Macmillan.

Mansbridge, J.J. (1980), Beyond Adversary Democracy, New York: Basic Books.

Martin, B., Baker, C.M.A., Manwell, C. and Pugh, C. (eds.) (1986), Intellectual Suppression, Sydney: Angus and Robertson.

Melman, S. (1958), Decision-making and Productivity, Oxford: Basil Blackwell.

Morris, D. and Hess, K. (1975), Neighborhood Power, Boston: Beacon.

Papanek, H. (1973), 'Men, Women, and Work: Reflections on the Two-person Career', American Journal of Sociology, 78, 852-872.

Parkin, F. (1979), Marxism and Class Theory: A Bourgeois Critique, London: Tavistock.

Perrucci, R., Anderson, R.M., Schendel, D.E. and Trachtman, L.E. (1980), 'Whistle-blowing: Professionals' Resistance to Organizational Authority', Social Problems, 28, 149-164.

Roberts, E. (1973), Workers' Control, London: Allen and Unwin.

Saks, M. (1983), 'Removing the Blinkers? A Critique of Recent Contributions to the Sociology of Professions, Sociological Review, 31, 1-21.

Stanton, M.A. (1978), The Merit Principle: Its History and Future, Canberra: Australian Government Publishing Service.

Weber, M. (1964), The Theory of Social and Economic Organization, New York: Free Press.

Weinstein, D. (1977), 'Bureaucratic Opposition: the Challenge to Authoritarian Abuses at the Workplace', Canadian journal of Political and Social Theory, 1, 31-46.

Weinstein, D. (1979), Bureaucratic Opposition: Challenging Abuses at the Workplace, New York: Pergamon Press.

Williams, T.A. (1982), Learning to Manage Our Futures, New York: Wiley.

Willis, E. (1983), Medical Dominance: The Division of Labour in Australian Health Care, Sydney: Allen and Unwin.

Willis, P. (1977), Learning to Labour, Westmead: Saxon House.

Wilson, H.B. (1974), Democracy and the Work Place, Montreal: Black Rose Books.

Young, M. (1958), The Rise of the Meritocracy 1870-2033: An Essay on Education and Equality, London: Thames and Hudson.

Zald, M.N. and Berger, M.A. (1978), 'Social Movements in Organizations: Coup d'etat, Insurgency, and Mass Movements', American Journal of Sociology, 83, 823-861.