2
Bureaucracy is that form of permanent organization of social action that is directed by a central authority.1 It is characteristic not only of state instrumentalities but also of large non-state institutions, business enterprises and universities, for example. Because the organization is permanent, most of its functioning elements are also permanent, and even the individuals who fill various offices in the organization are normally permanent and full-time members of it. Its various functions are regulated by fairly specific operating procedures and are subject to an hierarchical chain of command. This chain of command is supposed to ensure that the organization carries out the policy that is decided at the top, by the relevant political authority, board of directors or trustees. Democratic control of bureaucracies is supposed to be exercised mainly through their controlling bodies.
The attractions of bureaucracy are obvious enough. To the extent that it works according to rules, its operations are predictable and reliable in normal circumstances. A great deal of significant activity goes on without most of those who depend upon it having to worry about it. In so far as a bureaucracy is governed by a chain of command a policy formulated in fairly general terms is translated into ever more specific directives that result in a co-ordinated set of specific outputs giving effect to the policy. In the reverse direction the flow of information from the base operating agencies up through the system ensures that a large amount of material that is relevant to the assessment of policy is brought together as a basis for policy review. Bureaucracy, it might seem, is simply rational administration.
Actual bureaucracies are no doubt always imperfect, as are actual machines and organisms. But since organization is indispensable, so is bureaucracy, at least where there are permanent functions to be carried out and controlled. Permanent officers gain a great deal of experience that is indispensable to effective, predictable action, and are well motivated to make a success of their work in the hope of promotion and influence in the organization. Stable roles seem obviously important, and continuous monitoring of operations very much preferable to sporadic responses to breakdown in operations. Careful planning and co-ordinated implementation seems much more rational than ad hoc attempts to deal with problems as they arise. 41
Whatever the defects of actual bureaucracies the remedy lies in better organization, better control, better flow of information and better policy. The widespread illusion that the upper levels of the pyramid are simply parasitic on the lower levels that produce tangible outputs is dangerous. Abolishing co-ordination and control would not make the lower levels more responsive to their clients, much less to the needs of the people as a whole. It would leave them with arbitrary and autarchic power. Abolishing bureaucracy would destroy in both public and private organizations significant economies of scale, the concentration of resources needed for important projects and the accumulation of information necessary to understand the impact of policy over a variety of sectors and over significant periods of time.
The case against bureaucracy is equally well known. We have already mentioned some aspects of it. Simply because they work with a limited and entrenched repertoire of things that they can do bureaucracies are highly inflexible. The range of policy choices they can articulate is constricted by the sort of information they gather, the sort of failures or needs to which they normally respond and the sort of outputs that they are equipped to produce. They inhibit novel understanding, choice and action. The longer the chain of command the greater the distortion in the message that reaches the other end as at each level it is interpreted and translated into more specific terms and filtered by the bias, inertia and myopia prevalent at that level.
The more the chains of command fan out the less likely are the interpretations put on the original directive to result in a set of specific activities that are even consistent, let alone integrated with each other. While the bureaucracy may monitor its own output, it will record that information in quantitative terms that have little regard to quality. The picture that is transmitted back up the line is almost always extremely misleading and very much filtered. The bureaucracy is almost always both incompetent and poorly motivated to assess the impact of its output on the needs and problems the output was intended to deal with. At the same time it is highly resistant to the use of other means to assess its performance.
However, a bureaucracy is not for most of the time engaged primarily in responding to directives. Each office is normally engaged in its routine work, in relative isolation both from its superiors and from other offices at the same level in the same department, let alone in other branches of the organization. In order to ensure the smooth operation of its routine it will tend to make itself as operationally self-sufficient as possible, often reduplicating services available elsewhere in the organization in order to avoid having to rely on “outside” cooperation. Similarly, in order to be capable of responding to fluctuating demands on its services it will tend to try to hoard the resources 42to meet any eventuality, making “busy” work of a routine kind to justify the need for them in times of slack demand. Notoriously, this functional tendency is reinforced by the anxiety of officers to increase the size and apparent importance of their empires. The result is not only a very uneconomic use of resources but a very cumbersome organization, almost incapable of responding appropriately to any actual demand for action.
Resources become locked into established agencies that are reluctant to make them available to other agencies, for fear of disturbing their routine, weakening their case for needing those resources, or losing functions to other agencies. Cooperation must be paid for, and the price is as high as the agency can exact. Policies that demand close and continuing co-operation of diverse agencies are inherently very difficult to implement and even more difficult to assess when they fail. Reorganizing the chain of command is the usual response to such failures. New agencies are set up, old ones transferred to new organizational locations, responsibilities redefined in order to avoid a repetition of the same situation. But already the situation has changed and the new organization has not anticipated the change.
One could go on detailing the inherent problems of bureaucracy. Even most bureaucrats do not like it very much, They feel constrained by all those rules even when they are seen as having great power.
No doubt a certain amount of bureaucracy is necessary, since some permanent and centralized agencies of substantial size are necessary in a highly complex society. The question is whether it can be minimized. Two ways have been suggested, market control and political decentralization. I shall return to the question of the market in the next chapter. Let us assume for the moment that a large range of public goods cannot be supplied by the market but only by politically, preferably democratically, controlled organizations. Two questions then arise: (1) Is it practicable to break down agencies providing public goods into units that are small enough to enjoy a relatively short chain of command, granted the need for reliability, co-ordination and adequacy to meet very broad social problems? (2) Is there a feasible way of ensuring democratic control over such autonomous specialized agencies? This latter question is, I believe, the crucial one. It will take up a great deal of the rest of this book. Meanwhile, it is important to face the first question.
It is obvious that there is no difficulty in principle about devolving autonomy to smaller units. It is a principle of bureaucratic organization that the operating agencies in an organization should, as far as possible, each be autonomous within its special domain. The various agencies in a bureaucracy are usually not primarily instruments of a higher purpose but operating 43bodies with regular direct responsibilities to various clients to whom they respond. Some of those clients may be other agencies rather than the public. Even so, the realtionship will frequently not be one of detailed subordination in operation, even if the agency was set up by its client. In an administrative structure, unless function overrides status, chaos will ensue. The stability of modes of operation, the attention to detail and clear definition of responsibilities that are conditions of a successful operation, presuppose the principle of subordinate autonomy, even in organizations as tightly hierarchical as an army.
There is no inherent difficulty in specifying and giving autonomy to various functions even when these functions are highly interdependent. On the contrary, a function that depends on another usually needs to be able to rely on that function being carried out without its having to do anything to ensure that it is. For the most part diverse independent agencies rely on higher levels of the bureaucratic hierarchy not to ensure that there is some explicit co-ordination of the agencies on which they depend, but simply to ensure that those agencies do their proper jobs. Is bureaucratic oversight a better means of ensuring that organizations do their jobs properly than direct democratic control?
For there to be democratic control the members of the controlling body should be representative of those with a stake in the good performance of the functions they are superintending, and should understand the problems involved in carrying them out. The question then is how to ensure that people of this sort are appointed. But even granting that this problem can be solved, the defenders of bureaucracy will still raise objections. The most important is that a controlling body of this sort is much more likely than a bureaucratic superior to negate the proper autonomy of the agency it controls. The bureaucratic superior has usually been through the ranks and knows from experience and training the importance of autonomy. Moreover, officialdom can resist pressures to intervene in ad hoc ways much more easily than representatives that are beholden to particular interests that are affected by the agency’s operations. The higher official can be and is motivated to take a broader and longer-term point of view than a committee representing interested parties.
It is easy enough to dismiss the bureaucrat’s defence in the name of popular sovereignty and to raise counter-objections based on the tendency of higher-level bureaucrats to protect their empires from scrutiny of any sort. The bureaucrat is also subject to pressures that have little to do with the welfare of the clients that the office serves. Promoting ease of administration, the appearance of efficiency, loyalty to the department and 44the interests of his or her immediate superiors are some of these pressures. Even a bureaucrat’s genuine concern for the welfare of the agency’s clients is inevitably paternalistic, shaped by his or her conception of what is needed rather than by what the clients want, much less really need.
Nevertheless the bureaucrat’s objections to democratic control raise a crucial point. It is desirable that the controllers should have direct knowledge of and interest in the needs the agency is intended to serve. It is not desirable that they each conceive their task simply as one of maximizing the pay-off to his or her particular constituency in the short term. They need to be motivated to seek to build up the agency and direct its policies in such a way that the long-term interests of all concerned will be maximized, often at some cost to a number of quite powerful short-term interests. It is a crucial test of any proposal for democratic control that it can give a plausible solution to this problem. If not we may be forced to accept a fairly high level of bureaucracy. A local health clinic, for example, will prefer to be subject to predictable regulation by a bureaucracy rather than constant unpredictable interference by people who are more concerned to show off their power and concern than to understand the problems of running a clinic and contribute constructively to solving them.
The difficulty might seem to be compounded when there is a need for explicit co-ordination and co-operation between various agencies to carry out projects and broad social programmes that transcend the functions of any one of them. It is important to my case to argue that this difficulty is less troublesome than it might appear. It has two aspects, one that might be called a technical one, concerned with joint planning and execution of specific projects, and a social policy one, concerned with questions about directions of social change. They involve very different considerations. The technical question turns on a comparison between largely voluntary co-operation between distinct agencies and largely enforced co-operation in a bureaucratic structure. The policy question is a matter of the ways in which decisions are to be reached about social priorities and how those decisions are to be fed into the planning process.
The technical question is the more easily answered. If the control of each of the various bodies that are involved in some community project such as the development of a new housing estate is vested in a committee that represents community interests, co-operation is likely to be very much easier than between the distinct agencies under a variety of bureaucratic controls. Standing departmental rivalries, fears of “outside” interference, red tape of all kinds and the inertia of routine are much less likely to be in evidence. Particular projects can be originated from a variety of different sources, 45and the ones that will win out are those supported by the widest consortia of agencies concerned. The detailed planning and execution of particular projects can be handled by co-ordinating committees. There would, no doubt, be a need for recognized arbitrators with authority to determine issues that were deadlocked between agencies, but these would have no authority to intervene until negotiation between the parties broke down. How such machinery might work and what authority it would have I shall show in more detail in a later chapter.
The policy problem is much more difficult to answer, if only because it is much more elusive. The full answer to it can emerge only when we have thought through the whole matter of democracy in a radically new way. At this stage what needs to be challenged is the question itself. On most issues there is no problem of policy. The illusion that there is is an artefact of our present forms of government. It is true, of course, that even in the richest societies there are choices to be made between priorities among desirable objectives, discriminations to be made between desirable and undesirable forms of social development and between different long-term courses of action. But it is almost wholly misleading to think of social construction and change in terms of the progressively more detailed application of high-level principles and objectives to ever more concrete circumstances either by processes of choosing means to ends, or by ways of instantiating general ideas. The importance that this image of social policy-making has assumed is due almost completely to the centralized character of the state bureaucracy and the peculiarities of the electoral process.
Political parties attempt to persuade the electors that in voting they are deciding between rival social policies. That is a contention I shall examine later on. In fact there are issues of broad social policy that are vitally important, but our present electoral systems are powerless to deal with them. At the level of decision between different constructive projects, however, questions of policy are largely irrelevant. What matters are the likely effects of the various practical schemes and their incidence on the various groups of people affected by them. The problem is to make best use of existing resources to produce the whole gamut of results that are desirable in different ways and degrees. As I shall argue in more detail later, there can be no golden rule here that can be applied to assess which is the best proposal. The important point for our present purposes, however, is that it is very often much easier to get agreement about which solution to a given problem is best in the circumstances than it is to get soundly based agreement on a comprehensive policy that purports to supply a general criterion of what is best. 46
A viable policy presupposes that the policy-makers have a very sound idea of what things can be done, on what scale and in what areas, given the resources available. Policy that is not based on such information is merely a recipe for a mixture of failures and missed opportunities. Too many projects are promoted because they purport to be in line with slogans, or left undone simply because they are not.
In originating, elaborating and securing support for specific projects that make maximum use of the resources available to meet the real needs of those affected, consortia of agencies in direct contact with the specific resources and needs in question are likely to be more effective than relatively inflexible and remote bureaucracies. They may lack the permanently installed experts and planners that a bureaucracy comprises, but it is usually more efficient to hire teams of such experts for specific projects than to be saddled with the limited capacities of permanent officers. Obviously such a practice permits of much greater experimentation and innovation. It also cuts down standing overheads. Policy assessments in most cases cannot consist of sets of fixed priorities. Any practical decision between competing proposals has to be based on attempts to quantify the relative costs and benefits involved. A reasonable and responsive policy will emerge from such a consideration of competing proposals, if those making the assessments are guided by specific knowledge of and sensitivity to the real needs of those affected.
In so far as there are policy questions of a more general sort, questions about the overall balance of different kinds of projects and the general emphasis of development, it is undesirable that these be handled bureaucratically. Ideally speaking, in our present parliamentary systems these are matters for political parties, the cabinet, the legislature and specific political appointees. It is notorious, of course, that these political agencies are not very successful in imposing on the bureaucracies the emphases and orientations they favour. Nevertheless, one of the most difficult tasks that the advocate of decentralization has to face is just this deeply rooted conviction that it ought to be possible for the popular will to express itself by imposing policy on the “public service” through centralized organs of political decision-making. It is widely believed, especially by those who favour social change, that strong centralized political control is a necessary condition of coherent change. Even those who recognize that bureaucracies are very difficult to control and biased in favour of the status quo are inclined to believe that weakening the power of central government inevitably gives more power to existing interests.
The short answer to these contentions is that the result of weakening centralized authority depends entirely on what other authorities take up its 47functions. In this case it is a question of how political priorities can be articulated and expressed in the community as a whole. The positive solution of this problem will engage us in later chapters. In the present context, I want to emphasize only one crucial point. I agree that devolving the functions of central government down to smaller local units of generalized political decision-making, states, provinces, counties or municipalities, or leaving them to the “private sector”, does tend to transfer power to entrenched interests. The alternative I am elaborating is quite different. It is a matter of giving autonomy to highly specialized agencies, many of which would have a limited geographical scope, though a number would be national or international. The point is to break down comprehensive political power at each and every level.
How this is possible remains to be shown. The present point is that if it is possible centralized bureaucracies are not necessary for carrying out public projects. The need for centralized bureaucracy is wholly a result of the alleged need for centralized political control. It is a functional necessity of the state, not a functional necessity of any kind of constructive activity.
It is not only states that generate bureaucracies. The growth of state bureaucracies has been paralleled by a comparable growth of large privately owned corporations, and, in some countries, of publicly owned corporations that operate in the market in much the same way as the private ones. Such corporations are responsible in some degree to their shareholders, whether these be individual investors, other companies or the state. It is clear that this responsibility is mainly limited to their paying some reasonable dividend to shareholders and building up their assets. Normally shareholders can neither have nor aspire to have any significant say in the corporation’s activities, except in those cases where the corporation is effectively a subsidiary of some very powerful shareholder with other interests which it helps to promote.
As for the general public, it can put some constraints on what corporations can do by using state power, but any influence that it has on the positive direction of corporations is exercised through the market place, and such pressure as consumer groups and special interest groups can use to affect the corporation’s reputation. The large corporation is, by and large, very stable, highly autonomous and largely self-financing, growing mainly from retained profits rather than new capital. At first sight, then, it seems to be a very efficient form of organization, well adapted to its environment. It is also very powerful. In the market place it can and often does exercise monopolistic 48or oligopolistic power, setting and maintaining prices, effectively excluding outsiders from entering the market and exercising a very strong influence on the pattern of life of both consumers and its own workers. Through its economic power it can also exercise great influence on government decisions.
The apoligists for large corporations sometimes make much the same claims on their behalf as are made for government bureaucracies. They are agencies of rational, long-term, and relatively enlightened planning. They can muster the resources to bring about major productive undertakings, exploit economies of scale and introduce stability and predictability to key sectors of economic life. To some limited extent these contentions are correct. As in the case of government bureaucracies the question is whether these advantages could not be secured in other ways and whether they are not overwhelmingly outweighed by other considerations.
The only additional features of the private corporations that are alleged in their favour by comparison with government bureaucracies relate to their dependence on the market. The market forces on the private corporation a concern for efficiency, as well as a responsiveness to the needs of consumers and a readiness to innovate that has no parallel in the public sector. It is forced into these concerns in its own interest, but ultimately they are also in the public interest. The validity of these claims will be examined when we come to examine the market mechanism. In any case they are not arguments for size. On the contrary, the larger the corporation the less tightly is it governed by market constraints.
In the case of the so-called conglomerates, such as ITT, which control a great variety of companies with no functional relationship between their operations, one of their main purposes is to insulate the component companies from certain effects of the market. In particular, they can raise capital more cheaply, pool reserves and assist each other in ways that would not be possible without central co-ordination. Their viability has nothing to do with any increase in productive efficiency or even marketing, but only with the business of monetary management. Even in the case of single-industry giants such as the car manufacturers, the structure of the corporation has more to do with making money than making cars. All such manufacturers buy in substantial quantities of components because it is more efficient to do so. That they seek to expand their capacity to produce most of the components themselves is more a matter of securing their independence from outside supplies and their ability to keep potential competitors out of the industry than of technical desirability. It is also, of course, a matter of building up the capital assets of the corporation and the power of those who control it. 49
The corporation that acts for profit in the market place is in a very much simpler position for monitoring its impact and efficiency than a public corporation. Its impact is measured by its share of the market and its efficiency by its profitability. These measures of success can be applied directly to most operating divisions by comparing the costs of one division with another or with the costs of buying in various products from outside. Because they are subject to such clear and inexorable judgement divisional managers can be given a much freer hand than in a bureaucracy producing public goods. Empire building, lack of responsiveness to needs and opportunities, and sheer laziness show up sooner and much more clearly and quickly in falling sales and profits than in public dissatisfaction. Precisely because there is always a great deal of public dissatisfaction it is extremely difficult to distinguish clear messages from the general noise.
It is because of these advantages that private bureaucracies have a much better reputation than public ones. Their power is often feared, but their efficiency is less frequently questioned. Nevertheless, they are inefficient when it comes to innovation. New products, and technical and productive revolutions tend to come from small firms. There is an inherent tendency for large firms to entrench existing technical practice in an organizational structure that is not easily changed, to concentrate on marginal improvements in current practice rather than radical rethinking, to attempt to persuade consumers to buy what the corporation is geared to produce rather than explore new needs and new ways of meeting them, to economize by squeezing suppliers, workers and retailers rather than by radical re-organization.
Much more will have to be said about efficiency in allocation of resources and in responding to needs when we come to criticize the market mechanism. Meanwhile, I want to examine a number of questions about industrial democracy, assuming for the moment that the market mechanism is an inadequate form of popular control over large productive enterprises. I shall also assume that the point of industrial democracy is not primarily to enhance the interests of people as consumers, or even to protect their common interest in their material heritage of natural and produced resources, but to give greater scope for all-round development of people’s capacities and social relations by giving them more control over and responsibility for their work.
Obviously, advocates of industrial democracy often want to argue that there is no systematic conflict between these objectives, and even that each reinforces the other. But their central case rests on the inherent desirability of enhancing the satisfaction people get out of their work, simply because it is so substantial a part of their lives. Sacrificing satisfaction in one’s work for the sake of leisure activities is to be avoided if possible. The desirable thing 50is to break down the work-leisure dichotomy, but that seems to be much more a matter of agreeable working conditions than of control of the firm.
It is by no means clear exactly what the members of a work team need to control or participate in to get satisfaction from their work. Discharging the tasks of management is itself difficult work, and not everybody finds it congenial. By and large we enjoy doing those things we are good at doing, and aptitudes for negotiation, persuasion, policy decision and administration are as unevenly distributed as aptitudes for scientific experimentation, theorizing, playing the piano or any craft skill. Doctrines that claim that the proper development of human personality is possible only if all these rational capacities are developed to the full in every person are essentialist, dogmatic and ultimately repressive. They attempt to force people into a single mould. They arouse little enthusiasm in most workers because they implicitly devalue other talents. Indeed, the main attraction of proposals for participatory democracy is their opposition to the prevailing practices that tend to assign workers narrowly circumscribed roles and very little autonomy within those roles. Such interest as most workers have in having a say in management is related to defence of their jobs and conditions of work. By and large they prefer to have these interests defended by a special independent organization, the trade union, and fear that co-option of their representatives into the tasks of management will blunt their effectiveness in advancing the specific interests of workers. Yugoslav theory and experience of management elected by workers attests to these points.2 There have been strikes in worker-controlled enterprises.
A fairly sharp distinction and tension between management and labour is inevitable in a market economy where the success of a firm is crucially dependent on risk-taking decisions about the mix and quantity of production, pricing, contracts with suppliers and retailers, and the constant monitoring of costs. Nor can analogous risks and problems be avoided in a non-market economy. Ultimately the fate of the enterprise will be judged by its usefulness in the plan, and management decisions will be taken in the light of the needs of the plan rather than the preferences of the workers. Management would like to put the brunt of the risks on to workers, workers to insist that the firm must accept them.
Even if there were some way in which the workers could have an effective and relatively direct participation in these decisions, another problem emerges. If workers can change jobs easily there is little incentive for them to place the long-term viability of the enterprise before their own short-term interests. If they cannot easily change jobs they do not have one of the most effective ways of both securing work that suits them and ensuring that 51work that is generally regarded as unpleasant is compensated adequately by other rewards. From this point of view it is probably more important for the welfare of workers both individually and collectively that they have a positive freedom to choose what jobs they will accept than that they enjoy either security of employment in a particular job or responsibility for the way that enterprise is managed.
Such freedom depends on a host of factors such as support while they are looking for a job, opportunities for retraining, portability of entitlements such as superannuation and leave, and in some cases assistance with the costs of moving themselves and their families. Ideally, perhaps, nobody would be compelled legally or economically to work, at least for most of the time. In such a situation people would only work because of the intrinsic rewards and the goods and services they could purchase with additional income. It sounds utopian. However, it is a striking fact that most people who are rich enough to live at least moderately comfortably on their unearned incomes do undertake full-time jobs. Many women who are under no social or economic pressure to work look for outside employment when there is not enough interesting work at home. If the opportunities for interesting work and the rewards for uninteresting work were greater there is little reason to doubt that most people would want to work. In such circumstances, given modern productivity, advanced societies (and very soon not so advanced ones) could afford to carry those who, whatever their reasons, did not want to work, and were content with a minimal income. The cost would be the price of ensuring that work would be freely undertaken.
In such a situation the problems of participation might be expected to be solved in a very large variety pf ways, depending on specific contracts of employment between individuals and teams of workers and management, however management is chosen and whatever the criteria of success for the enterprise. Such a system should be efficient, flexible and dynamic, if only it could be financed and administered satisfactorily. A host of questions arise. Who is going to decide on the level of payments to those who do not work? Is the money to come from taxation on those who do work, or is there some better way of raising it? Does it not in any case require a very powerful state apparatus of the kind we have already rejected? The answers to these questions are complex and will, I hope, emerge in the course of subsequent chapters. At the same time I shall attempt to produce reasons for thinking that a social structure based on negotiated arrangements between genuinely autonomous groups can better satisfy the conflicting requirements of libertarian and socialist aspirations than any form of capitalism or of planned economy. Finally, as this is not just a conceptual exercise, I shall try 52to show that the solutions I am advocating could work under social conditions that either exist or could be brought about by means that are already available. All that is lacking is the political will.
[I confess that the last two paragraphs now seem very dated.]
There is great danger in talking of bureaucracy that the enormous variety of kinds of organization is lost from view. So far we have concentrated on productive agencies that deliver positive services to consumers such as public utilities, educational institutions, hospitals and the like. At the other end of the spectrum purely regulatory agencies are concerned only with keeping the activities they regulate within certain bounds, for example, police, safety and pollution control authorities. But many agencies have mixed roles, or play one kind of role as a means to another. So a consumer protection agency may, in order to carry out its regulatory role, produce a great deal of information that is useful to other agencies and to individuals for a variety of purposes.
One of the things that people most frequently complain about when they object to bureaucracy is the rigidity and arbitrariness of its control functions, the impenetrable thicket of rules and regulations, the ways in which rules are applied, the counter-productive character of the rules, the amount of work that goes into getting permissions and making reports. All of this makes it unnecessarily difficult and costly to do many things that are otherwise needed or useful, and it is open to both a host of petty abuses and some significant major abuses. Hence the very widespread belief that it would be much better if productive organizations could be so designed that they were as nearly self-regulating as possible.
In order to understand the possibilities of self-regulation it is important to look more closely at the differing kinds of control that are involved in organizations of a relatively permanent sort. It has become fashionable to use the word “control” in an increasingly abstract way, which in turn relates to very vague conceptions of power, authority and agency. Take, for example, the dining room of a large hotel. From one point of view it might be said to be under the control of the customers. Individually they give the “orders” that determine what is produced when and in what quantities. Collectively they establish the pattern of demand and constrain the operators to meet that pattern. At the other end of the operation the hotel management controls what is done. It not only sets policy and decides what resources are to be put into these operations but monitors them through its accounting procedures and quality control. In between, the restaurant manager controls 53the actual production, assigning responsibilities to cooks, waiters and other staff, determining the menu and supervising every aspect of the operation.
It is obvious enough that each of those controls serves a different specific set of interests in the operation, each operates by very different mechanisms and each requires different knowledge and skills. A similar structure of different kinds of control is evident in most government departments. Policy is set at the top, resources assigned and monitoring procedures set up. Precise regulations are made, procedures of implementation and assignments of responsibilities instituted by permanent officers within the organization. The activities of clerks at their desks and officers in the field are controlled by clients with claims or by delinquents to be brought into line. From this point of view the police are “controlled” by criminals and the army by the enemy.
In addition to these “vertically” differentiated controls, there are usually a variety of “horizontal” controls. The firm is controlled by its competitors, the government department by its rivals within the bureaucracy that want to take over its functions or resources. Most organizations are sensitive to fashion, the approval or disapproval of opinion-makers. Their professional standing is not only a means to whatever other ends the organization or its members may have, but puts the seal on the intrinsic worth of what they are and what they do. What these horizontal controls are worth as controls depends largely on the appropriateness and openness of the competition through which they exert their effects. Those who articulate the verdicts must be governed by solid, adequate information and appropriate criteria of assessment. They must not be manipulated by public relations exercises, bemused by their own capacity to orchestrate public opinion, or in the service of extrinsic purposes.
For many purposes avoiding the undesirable aspects of bureaucracy is mainly a matter of substituting horizontal for vertical controls both within the organization and over the organization. Horizontal controls, especially peer assessments, can be much more flexible and much more apt to produce excellent performances and readiness to take proper risks than any controls that can be exercised from above or below. Assessment from below is necessarily largely assessment on results, especially unpalatable results. The consumer, the lay person, cannot know whether an organization is making the most of the opportunities available or is technically efficient. The public may not know what it is missing. Even where it does, it can voice its discontents only in negative terms.
Control from above is even more unsatisfactory. The further it is away from the ultimate product the less it can assess accurately its quality and appropriateness. The further away it is from the ultimate production process the less it is in a position to be aware of or appreciate alternative possibilities. The more tightly it attempts to control the organization the more it becomes 54a prisoner of its subordinate command structure. The more it has to rely on negative sanctions, the greater the emphasis it puts on mere quantity, regularity and avoidance of error rather than quality, adaptability and experiment.
Increasing horizontal control, however, presupposes not only dismantling control from above but ensuring that the legitimate need for control can be met. In order to understand this problem better it is useful to distinguish two aspects of organizations that are not usually contrasted strongly enough. Take an extreme case, an army. One purpose that control and central planning must serve is to ensure that the army is well prepared to meet a variety of contingencies. It must be kept functioning, and occupied in a host of day-to-day activities when it is not fighting. It must be equipped, organized, trained and exercised on a continuing basis. At the other extreme when it is actually engaged in warfare it must respond to a variety of different demands in a flexible, prompt and effective way. As a standing peacetime army it is organized by regulation, budgeting, long-term programmes and established routines. As a combat unit it is deployed by command, leadership, co-operative action and detailed response to the contingencies of battle.
Every organization has these two aspects to some degree, often involving different systems of control. The organic analogies are striking. One system (or complex of systems) builds up our muscles, keeps them in dynamic equilibrium and repairs them. Another puts them into action. The normal administrative apparatus of an organization ensures its supply and general preparedness to produce a certain sort of performance by regular standardized routines. The specific controls that bring about particular performances may come through quite distinct channels. Even when they come through the “normal” channels they take a different form, orders of all sorts, decision-making conferences between various units, and specific programmes rather than rules, routine checking or standing instructions.
It is relatively easy to monitor and assess particular performances. The results are often unequivocal and the relations between the decisions and those results easy to trace. Horizontal controls are usually prominent and powerful in such cases. But what an organization can do in response to specific and episodic demands on it depends on the soundness of its preparation and the range of its repertoire of performances. These things are very much harder to assess prior to any actual performance. The preparations may be irrelevant to the challenge that in fact emerges. It is often said that armies are usually prepared to fight the last war rather than the next. The same tends to apply to large organizations generally. Their very success in building up a capacity for production and action of a certain kind makes it difficult for them to do 55anything else. They build in so many constraints on variety and initiative precisely because they are well designed and integrated to produce certain results and not others.
Formalized horizontal controls are not likely to be very much better than vertical ones in this respect. The standards of sound practice are inevitably conservative and conventional. There is, in fact, as the advocates of “free enterprise” have so insistently reminded us, no solution to these problems in terms of planning and control. If we are going to have productive organizations that are innovative, risk-taking and easily changed, they must be controlled mainly by inducing those who direct them to face the risk of failure in the hope of the rewards of success. Individuals and organizations must have access to the resources necessary to experiment. To deter recklessness they must pay a penalty for failure. To motivate them to try they must receive rewards for success. The criteria of success must be appropriate and they must be applied appropriately. Assessment must be directed mainly at performance rather than at preparations for performance.
If that is the case, then there is not much hope for the army. The proponents of “free enterprise” are equally certain that there is no hope for public enterprise either. The crucial problem, assuming for the moment that success and failure can be assessed and rewarded appropriately, is the allocation of resources to individuals and groups that want to experiment. It seems fairly obvious that these will have to be distributed in fairly small parcels. The larger the stake in proportion to the total resources available the less sense it makes to encourage risk-taking. The larger the scale of the experiment the more likely it is to exclude undertaking other experiments with which it might be compared. The larger the organization the more it is likely to become inert or inflexible.
On the other hand, the larger the number of organizations competing to supply a similar service the more difficult it is to decide which are to get the chance to do so, the more likely it is that there will be duplication, overlapping and pointless competition between them. Co-ordination; negotiation and stable joint enterprises are the more difficult the more independent sources there are of variation and aim between the agencies whose co-operation is needed for a given project. Where one is not relying on the market, which distributes resources mainly on the basis of past success in the market, it is difficult to see how one escapes the need for a centralized bureaucracy to make the distribution.
There are formidable problems. The solution that I shall advocate is in principle very simple. We shall come to it in due course. For the moment I shall try to clear the ground a little more. 56
One of the paradoxes of bureaucracy is that it is generally perceived as powerful and oppressive by its clients while those who work in it at all levels see themselves powerless and burdened with responsibilities. The paradox is, of course, easily resolved. The bureaucrat is perceived as having great power because he or she is in a position to invoke a complex of regulations, provisions, procedures and requirements that impose burdens on clients, structure very tightly the choices open to clients and bring great pressure on them to act or refrain from acting contrary to the bureaucrat’s desires. The bureaucrat is harassed by the pressures from above, below and each side to make rules and practices work in spite of their being ill-designed to cope with circumstances that arise, and in spite of the failure of other parts of the machinery. The penalties for error and unauthorized behaviour are high, the rewards for success are low. That problems get solved is seen merely as fulfilling one’s duties, even when the solving of them is in fact a result of great effort and ingenuity. In these circumstances there is a very strong temptation to fall back on a literal interpretation of instructions and regulations, to attempt to shift the locus of responsibility to somebody else and refuse to respond to substantive need whenever they threaten to create problems. It is very dispiriting.
There is no doubt that overcoming this paradox involves giving much greater freedom to those who are called on to solve the base level operating problems and much greater rewards for producing creative and beneficial solutions to them. This clearly involves more flexible controls, more attuned to substantive rather than procedural criteria. In general this is much easier to achieve when the context is one of meeting a substantial present challenge than in the regular context of maintaining an organization. An army in battle encourages initiative, even at the cost of bending the rules, calls for actions that are beyond the call of duty rather than routine compliance, and rewards the successful taking of risks. The contrast with regular discipline and management practices is as striking as the contrast in attitudes, behaviour and relationships among the participants in the two situations.
It is impossible, of course, to eliminate the routine work that is necessary to keep an organization functioning. Records must be kept, resources marshalled, training exercises repeated and structures maintained. By and large this “housekeeping” work is much more tolerable if it is not an individual’s or agency’s unique occupation and if it can be seen as integral with and tailored to the requirements of substantive action by the agents themselves. So what is needed are small organizations that can be put together as occasion demands into more complex organizations, rather than large standing organizations from which units are dispatched to carry out 57particular tasks. The survival of any unit would be made to depend on its actual usefulness to its ostensible task and to other agencies, not just on the tendency of a large organization to hoard resources that might prove useful, and to multiply functions whose sole purpose is to keep the organization busy. In other words there needs to be a form of decentralization that puts resources in the hands of agents, gives them a flexible area of substantive responsibilities and rewards or penalizes them according to their success in discharging those responsibilities. Co-operation and coordination with other agencies must be included in those responsibilities and the manner and precise objectives of those joint efforts left to their initiative.
The terms of office of those who staff these agencies must be long enough for them to get to know the problems, evolve practicable solutions and carry through their projects to completion, but not long enough for them to sink into routine or prevent others from trying a fresh approach to the problems. The reward for success must be explicit recognition by their successors, clients and peers of the quality of their work. This would be expressed in part by a recognition of their suitability for tackling more difficult tasks, such as that of adjudicating conflicts between agencies, reviewing their structures and functions and allocating resources between them. Obviously, the key problem in making such a system work is, Who is to articulate the host of particular judgements of success and failure on which it depends and make and give effect to the consequent decisions?
It is obvious that such flexible organization, assuming it can work, could not be an effective instrument of any single narrowly defined policy, whether it be the military policy of a particular state, a definite centralized policy of income redistribution or the market strategy of a large corporation. It is designed to decentralize power and initiative and encourage experiment not just with means but with diverse specific goals. If there is some precise set of objectives that constitutes the will of the people it will not be an appropriate way of achieving them. On the other hand, if the will of the people is simply the very abstract objective that the best use be made of a great variety of community resources to do whatever can be done to meet the specific needs of many diverse, overlapping and dispersed groups of people, then it might be expected to be far more responsive, creative and effective in doing so than any centrally controlled bureaucracy.
It would not, however, be blind in the sense in which the market or an ecosystem is blind. Through such a system, I shall argue, much more dynamic and well-directed forms of social change would be possible than can be achieved by any form of centralized direction, however democratically controlled. If there were a widely shared view about the changes that need to 58be made in a society, it could articulate them in practice much more quickly, appropriately and effectively than any centralized process. As each agency in its own area of responsibility, and in co-operation with other agencies, attempted to advance those specific interests that were seen as disadvantaged, to give effect to those values that are most widely regarded as important and relative to their concerns, and to monitor and adapt their procedures in the light of their specific effects, the result should be a massive shift of policy.
[The successes of feminist and ant-racist campaigns over the last generation illustrate this point.]
Because the detailed implementation of policy would be so closely tied to specific problems and possibilities the danger of pure dogmatism and of systematically counter productive policies would be minimized. Entrenched interests that exercise great conservative force on a central government because of their ability to organize and concentrate their efforts on a large scale would usually be much less effective in dealing with a variety of autonomous specialized agencies. Even very powerful local interests that can dominate a centralized local authority would find it harder to dominate a variety of specialized authorities many of which had no narrowly local base or little organic connection with the activities in which the local interest was based.
The sort of social organization I am advocating would itself constitute a major change, one of the most significant in history. Nevertheless, although the problems of introducing it and making it work in practice may seem insurmountable, many of the social conditions of its working are already present in most advanced polyarchies and in the major socialist countries. People are very well aware of the enormous complexity of such societies. There is a very strong desire to dismantle centralized controls and disperse power to functionally appropriate units. In spite of all that has been said about the “privatization” of life as people have become aware of their powerlessness, there is a great interest in public affairs and, I believe, great willingness to participate in them if only it were more possible to do so in some limited but effective way.
The interests of most people are sufficiently various for them to be well aware that their overall interests in a complex society can be advanced only by constructive cooperation. They resort to confrontation reluctantly, as a last resort. They are well aware that a great deal of the avoidable evil in the world is the result not of malevolence but of agents being forced (or perhaps seduced) into inappropriate policies and actions by (what they see as) the requirements of their institutional roles. Even where malevolence is to blame, 59that it can have such disastrous effects is the result of institutions that make it possible for such people to acquire inordinate power.
Indeed it is remarkable that public and private bureaucracies in many countries function as well as they do in spite of their rigidity, their tendency to produce misinformation, their costliness and the lack of satisfaction of both their staff and their clients. It is, I believe, very good evidence that there are many people in such organizations who are prepared to work very hard to get the appropriate things done, to respond imaginatively to crises and to work for the public welfare as they understand it. That so many people are prepared to put up with the degree of bureaucratic control that they do, and with the very high taxation necessary to fund it, is evidence of a realism that is prepared to accept considerable costs to achieve social goods that are not necessarily in their own interests narrowly conceived. Many people vote for parties that will tax them more severely than their competitors because they are more concerned about the direction of social policy than about the cost or benefits to them individually in the short run.
Of course, people who would derive satisfaction from working for the public good for modest but solid rewards may be a relatively small minority. Nevertheless, they can be found in pretty well any section of the community, whether one divides it up by occupation, status or region. They crop up all over the place. It is not necessary to preach some narrow morality that exalts such people above those whose interests are more private or more special. All that is required is that most people recognize that public service is a worthy thing, along with devotion to one’s artistic talent or one’s family or science or dahlias. Much less is it necessary or desirable to attempt to mould everybody into a public person. Why should not those who are not interested in public office simply turn over the work to those who are, provided they are so organized and chosen as to be sensitive to the interests that need to be taken into account and adequately motivated to do a good job?
The result would be democratic if all interests were appropriately represented, if everybody who wanted to had a reasonable and equal chance of participating actively and if no individual or group could acquire entrenched power in the system. Most of our lives depend on other people doing their jobs properly without our having any substantial control over what they do. We have to rely on others to do what needs to be done and absorb the effects as best we can. Many of those effects are in any case unpredictable and uncontrollable. The culture we share is the product of artists and scientists pursuing their own work. The opportunities for consumption we enjoy are the product of many independent producers attempting to produce things that they think people will want. The richness of our lives is in part at least the result of producers disconcerting our expectations, creating new 60challenges and opportunities for us, and inevitably certain disappointments. The morality to which we must adjust our personal relationships and the valuations people put on us are matters over which we can have only partial control. Social norms, our friends, and people’s standards of judgement change in a variety of ways to which we have to respond.
For most people most public affairs are matters that they can no more control than they can control the weather. They adjust to one as the other as best they can. They do not understand the complexities of public policy, and they know very well that their influence in such matters is in any case infinitesimal. They are prepared to accept any reasonable system for getting things done. Indeed they accept that there are plenty of people who know a lot better than what can be done and what results it is likely to have in most cases. However, even such people often have a strong interest in some area of public decision-making that affects them in a particular way, and they may well have or be prepared to acquire the skills and knowledge needed to take an active part in attempting to deal with that area. Given the opportunity they may make a significant contribution to it.
In the practical context of attempting to make decisions about complex matters such people are likely to learn that these questions are not zero-sum games where a gain for one interest is necessarily a loss for another. On the contrary most of them involve a variety of dimensions in each of which various interests are advantaged or disadvantaged in different ways. The optimal solution results from a process of constructive trading in which the significance of losses and gains changes with the context set by other aspects of a proposal taken as a whole. In such a situation the ones who are likely to win are those who best understand what others have at stake and are able to articulate proposals that meet a wide variety of interests in various ways to a degree that is satisfying to each.
Such bargaining is impossible if everybody affected is supposed to participate. It is necessarily a matter for small numbers of people who are prepared to put in a great deal of work on it. They must not be too narrowly constrained by “non-negotiable” demands. They must rely to a large extent on their knowledge not just of what those they represent want, but of what they need and what they might reasonably be expected to settle for. The question then is how can we get this sort of negotiation into the hands of people who can be trusted to do their best for the interests they are supposed to represent? The almost universally accepted answer is by free openly contested elections. Our next task is to show that that answer is radically unsatisfactory.
[Readers who are already convinced that mass voting is a very inadequate decision procedure in large scale politics may safely skip the next section.] 61
1. This chapter is very indebted to Andrew Dunsire’s The Execution Process, vol. I Implementation in a Bureaucracy, vol. II Control in a Bureaucracy (Oxford, Martin Robertson: 1978). See also E. Kamenka and Alice E. S. Tay, Bureaucracy (Melbourne, Edward Arnold: 1979).
2. I am sorry that I have not the space to discuss Branko Horvat’s book The Political Economy of Socialism: A Marxist Social Theory (Oxford, Martin Robertson: 1982).