62

3

Democracy and representation

I Voting

Most contemporary authors explicitly or implicitly treat free elections for the supreme offices in the state as a defining characteristic of democracy. By contrast, I shall argue that electoral systems are inimical to rule by the people for the people. In order to bring this out it is necessary to look closely at what voting can achieve, what alternatives night be envisaged in the abstract and what are the conditions under which they might produce acceptable results.1

Revealing preferences

Theorists of voting who have used the tools of economic analysis have been much troubled by the problem of how to get individuals to reveal their preferences. Let us grant, just for the sake of argument, that people have a consistent and complete set of definite preferences ranging over all the alternatives available in a given context of decision, and that each voter, (person or institution) knows what its preferences are. There is quite literally a problem about getting people to reveal those preferences in many voting situations where it pays to conceal them for strategic reasons. But the more fundamental problem is, How can votes express preferences?

What the economists work with are indifference curves that represent a set of trade-offs between costs and benefits. People’s preferences are expressed by the relative prices they are prepared to pay for a certain additional amount of each specific good. But voting rarely expresses a clear willingness to pay any definite price for any definite good. The act of voting is almost costless. The cost of getting what I vote for is often shifted on to somebody else. Again, my voting for rather than against does not say how strongly I am in favour of a proposal. It cannot express anything like a price I am willing to pay. A majority with slight preferences one way may outvote almost as many strong preferences the other way.

Moreover, voting is usually a matter of simply “buying” or “not buying” large packages, much of which one does not want. Mere voting tells us very little because it registers so little.

There might be a great deal to be said for people having to “put their money where their mouth is”, or, if not their money, since that is very unequally distributed, their time and effort. For example, one might acquire votes to be used as one sees fit by performing quanta of useful drudgery in approved public works. Voting would then measure the seriousness and strength of one’s preferences. The difficulty is, of course, that few people would bother 63to acquire votes at any substantial cost if their votes had as little effect on the outcome as they normally must have in any large organization. Where there is a very large number of voters any individual’s parcel of votes is unlikely to be significant. The larger the organization and the broader the issues presented to the voters the less likely is it that the outcome of voting will be translated into the sort of concrete output that any particular voter wants.

On the other hand, if the number of voters fell so dramatically that small numbers of votes were significant, such a system would encourage “monomaniacs”, those who were intensely devoted to a single issue. It is, of course, not too difficult to think up variants to guard against such dangers. For example, everybody might have a free vote on each question, and additional votes might be earned on a diminishing scale that made it costly to acquire a relatively large number of votes. Or there might be a limit on the number of votes that could be applied to a single issue. Again, various forms of trading in votes and proxies might be allowed under specific conditions, making it possible for the committed to acquire the votes of the indifferent by competing for them in some way.

The difficulty with all proposals that try to make voting a genuine expression of preferences by relating the act of voting to willingness to bear costs is that any given system of this sort will work only under a very narrow range of conditions. So as conditions change it will be necessary to make continuous adjustments to the system. Such changes inevitably favour some interests rather than others. An authority of sufficient integrity and ability to make these adjustments would have to be chosen on some other basis than the voting it is supposed to regulate. In any case there is no guarantee that there are solutions to many of these problems even in theory, let alone in practice.

Rationality and preferences

In practice people do not have definite preferences over the whole range of alternatives that affect them. They have neither the information nor the analytical skill nor the imagination to construct the sort of stable schedules that the economists’ calculations require. They simply do not have the sort of preferences that economists want them to “reveal”. What they usually have are a few likes and dislikes that are not systematically evaluated in relation to the feasible alternatives. To do that is, of course, a difficult thing, and often not worth our while. Of course, this applies to market transactions just as much as to voting. But there is a great difference. The market ultimately forces everybody into a set of precisely quantified bets about what will give them satisfaction. These bets are continuously readjusted in the light of experience. Many of them are related to very short-term outcomes. There is a tendency for the process of continual adjustments of expenditure to produce some sort of consistent allocation. 64

Even so, a great many “preferences” are stable simply because other alternatives have never been considered. One may say, with some justification, that in such cases people prefer stability to any other advantages. In fact they often do, quite explicitly. But it is a completely blind preference when it consists in a refusal to contemplate alternatives. It does not consist in setting a finite price on the inconvenience of change. It is not a quantified factor among other factors but a refusal to quantify and calculate. So it is very misleading to argue that because such people pay a definite price for their “choice”, it does constitute a preference that is “revealed” by the fact that the price is paid. The price that people in fact pay in such cases is not even evidence of a willingness to pay that price, since they may be quite unaware that they are paying what they are. Much less is it a guide to what they would be willing to pay if they knew. So people may stay on in a house as long as they can pay the rent, even though much more suitable accommodation could be found much more cheaply elsewhere, not because of any positive attraction to their present house, but simply to avoid having to contemplate change and the risks it involves. The avoidance of risk comes to have infinite weight.

In a changing world attempting to avoid risk by refusing to change oneself is not a rational strategy for minimizing risk. To forego all possibility of adapting oneself appropriately to changes one cannot control is to run a very substantial risk of being needlessly disadvantaged by those changes. Nevertheless many people do adopt such a strategy in practice for a host of bad reasons. They may be bewildered or paralysed by fear. They may trust in their luck or their patron saints. Often they are deprived of the time, information and assistance they would need even to contemplate the problem. Sometimes they are just stupid or lazy or perverse.

Where many of the preferences expressed in voting are of this nature there seems to be little point in trying to construct systems of voting that are based on taking them as rational preferences in the sense required by the model of rationality on which these constructions rest. It is like building a skyscraper out of any old materials that happen to turn up even though the design specifies materials of the highest quality. It is important to avoid misunderstanding here. The argument is not that the wills of people who are irrational by some ideal measure should not be taken into account. It is not that it may not under some circumstances be expressed appropriately as a vote. The point is simply that there is no way of getting a set of precisely weighted preferences over a range of alternatives from a set of votes. They are not expressions of preferences in the required sense, since the value put on most of the alternatives is completely indeterminate or on others is fixed without any reference to their relative advantages and disadvantages. 65

It may be true that the market does in various ways put pressure on people to rationalize their preferences. In some sense most of us have to budget our money. But there is no way in which normal political processes force people to make realistic policy decisions between the range of alternatives open to them. There are just too many externalities, too many complications, and the time-scale is too long. If voting is to be a significant procedure in political decision-making it must be assigned a more modest role.

The limits of decision procedures

Even where there are genuine preferences there is no procedure, even in principle, by which all sets of expressed preferences can be aggregated into a social decision that guarantees an optimal solution to the task of reconciling them. There is a very substantial amount of precise knowledge about the theoretical impossibilities inherent in this kind of enterprise, especially since Kenneth Arrow discovered the “Arrow barrier”.2 It is very much less clear how relevant this knowledge is in a practical context. On the one hand, most social decision-making is at a level so far from the theoretical limits that those limits may well be as irrelevant to our purposes as absolute limits on the velocity of matter are to designers of motor cars. On the other hand, the most important practical difficulties are precisely those that are brushed aside in the theory, which tends to assume such conditions as perfect knowledge and rationality.

It is the practical problems that are most important in the present context. The work that has been done on theoretical models of the logic of preferences, voting and the conciliation of preferences makes several assumptions that are very unrealistic. One of these is that preferences between A and B are not affected by the outcome between, say, G and D. This assumption is necessary to ensure the transitivity of preferences, that is to say, that if A is preferred to B and B to C then A is preferred to C. The procedures of calculation require that we be able to construct alternatives that form constant and mutually independent objects of preference despite changes of context. However, in real life, states of affairs do not divide up so neatly. A conservative may prefer the status quo A to a change B and B to a greater change C. Nevertheless, it may well be the case that in certain circumstances he or she would prefer the combination of C and D to A, believing, for example, that a radical, consistent and thoroughgoing change is preferable to an unstable, incoherent compromise.

Again, preferences often depend on a context of repetition. Apple pie may be my favourite dish, but I do not want it every night. There are, of course, ways of getting around such difficulties. In the limit one can exhibit the alternatives as a set of “possible worlds” that conjointly exhaust all possible combinations of their components in every dimension in which they make a 66difference. The trouble is that the further one goes in that direction the further one departs from the original object of the exercise, namely to find a way of deriving an assessment of a complex from the assessments of its individual components or of comparing complexes on the basis of comparisons between their components taken discretely.

None of this is to say that there is no practical value at all in attempting to compare complex alternatives on the basis of preferences between their elements, but merely that such procedures are inconclusive. They may, for example, enable us to clarify why, although we prefer A to B and C to D we prefer B-with-D to A-with-C. The result may be a radical restructuring of our perceptions of the situation that is of the highest practical importance. Qn the other hand, it may well be that in a given context one gets a very good approximation to the actual range of relevant considerations by taking the elements as discrete, perhaps because the ways in which they are interrelated are so complex and variable that they can be taken as random for practical purposes. The point is that one cannot assume that they are random without some substantial warrant for doing so.

In general there are two possibilities. One can compare the results obtained by predicting people’s preferences between aggregates on the basis of their preferences between the components of the aggregates with their actual choices between those aggregates and justify one’s assumption that the conditions for applying this form of analysis are met by displaying its success in prediction.3 Or one can attempt to show positively that the relations between the various alternatives are such that they cannot influence the aggregate result in any systematic way. The first method is unconvincing unless the number of cases in which success is demonstrated is large and homogeneous. It also supposes that we already possess a way of determining people’s preferences between the aggregates that is independent of the method being tested. Normally that comes down to taking people’s saying what they would prefer as an accurate guide to what they would in fact prefer. Since we cannot usually contrive test situations, we cannot test many predictions in fact. So the conditions under which such methods are reliable are themselves difficult to discover.

Attempting to demonstrate that there are no relevant factors that are omitted from a set of objects of choice is even more difficult. Even if there were some comprehensive theoretical framework that could ensure that the descriptions of the various alternatives were complete in some sense and independent of each other, it is doubtful whether this would help very much. It might be possible in principle to describe each alternative exhaustively in physical terms, but such descriptions would not bear any systematic relation to the description under which people form their preferences. 67These are irremediably subjective and intensional rather than objective and extensional.4 It is not obvious, for example, that it is irrational to prefer object X to object Y that is an exact replica of X simply because X is a family heirloom.

These are only some of the difficulties in constructing a satisfactory formula for deriving preferences between aggregates from preferences between their components. If one adds them to the acknowledged difficulties of expressing weights of preferences in terms of votes, the problems of deriving satisfactory decisions by voting alone are clearly insurmountable in practice. That is the case even if we assume that people’s preferences are consistent and well based. Obviously “raw” preferences rarely are. We would all like to have incompatible things, sometimes because we are caught in unresolvable conflicts, sometimes because it is very difficult to know that they are incompatible. In these circumstances analyses of possible configurations of compatible preferences may be of great assistance in helping us determine the limits and sources of incompatibility and the ways in which we might revise our preferences to make them more realistic. But this use of preference theory is dialectical, involving a process of adjustment of preferences that undercuts any employment of preferences as fixed points of reference from which to derive a firm decision about a complex of preferences.

The practical importance of negotiations

In a social context, the only way in which this dialectical process can take place consciously is by negotiations in which the precise preferences of the participants are not fixed. Each party proposes to the other various packages, attempting to show how each might gain various advantages from alternatives that are also advantageous to their proponent. Very often such negotiations will involve key components that are in some degree prisoner’s dilemma situations.5 That is to say, the situation is so structured that the result of acting solely in one’s own interest without regard to the actions of the other party will result in each preventing the other from getting what it wants. On the other hand; if each party is prepared to settle on a course of action that is less advantageous than that which it would prefer in the abstract, both can be assured of a result that is acceptable and in the circumstances the best attainable.

Clearly there are very tight limits on the number of parties that can enter into fruitful negotiations on any particular matter. A good deal of knowledge of the situation is required, a great deal of time must go into examining proposals and constructing counter-proposals, excluding irrelevancies and exploring possibilities. An enormous amount of information must be communicated to and assimilated by each participant. In practice negotiations tend to be expeditious and avoid cross purposes and irrelevancies to the extent that the 68parties come to know each other well enough to reach a number of tacit agreements about what does not need to be spelled out explicitly, what is not negotiable, what is impracticable and so on. It is also desirable, though not, of course, necessary, that the parties are in a position to trust one another to work for a genuinely optimal solution rather than attempt to conceal relevant information or deceive the other parties to the negotiations.

The result of these considerations applied to any reasonably complex society is that there is no possibility of reaching reasonable conclusions about matters in which a diversity of interests and opinions are involved by voting or by the direct involvement of all those affected by the decisions. Most decisions have to be left to negotiators who enjoy a very large margin of discretion about what to concede and what to refuse to other parties. Realistically, then, the problem of democracy is that of selecting and controlling the representatives who are to negotiate the various decisions that have to be made about matters of public policy and administration.

Referendums

There is, however, another procedure that is rarely used but sometimes advocated to ensure democratic control over negotiations, namely that the final package constructed by the negotiators should be submitted to the vote of the people affected by it. This happens regularly within bodies such as parliaments where subcommittees report and recommend but the whole body accepts or rejects their proposals. Occasionally even bodies that have the power to act submit proposals to general referendums of those they represent, as the British parliament did in the case of entry into the European Economic Community. Some argued in that particular case that parliament was evading its responsibility. Its proper function is to decide what is best for the nation, which is not necessarily what a majority will vote for.

Whatever the merits of that particular case, there are sound reasons for not regarding referendums on specific proposals as a satisfactory way of controlling public policy in most instances and as a general practice. Some derive fairly directly from points we have already made about voting. On any specific issue it is very likely that many groups of people and interests will be affected much more strongly than others. Equity would seem to require that they be given votes in proportion to the degree in which they are affected. Otherwise a relatively apathetic majority may block a proposal that is of great advantage to a minority simply out of distaste for change or because it involves some relatively trivial adjustment on their part. But of course such uneven distributions of interest vary widely from issue to issue. Determining the relevant voting structure for each issue would be, to say the least, extremely difficult. 69

Moreover, it is often the case that a relatively small deliberative body will take a much more enlightened view on a matter than the general public does, even when it is a matter in which most people have something approaching an equal interest. Penal policy, and particularly capital punishment, is a case in point. Not many people understand a great deal about criminology. It is not necessary to attribute to them irrational motivations or moral insensitivity to understand why they often support capital punishment even when their legislature does not. They are using oversimplifed analyses of the problem, often based on very unusual cases. Their judgement is not soundly based.

It might well be that if referendums on issues of importance were held frequently the standard of information and understanding of many issues in the community would rise. But there are very severe limits on the capacity of everybody to be adequately informed about every thing that is of importance to them, given even the best will in the world, and we would need very much better channels of information than our present mass media. Moreover, the significance of one vote among millions is so small that it must seem pointless to most people to bother about most issues. When in doubt the devil you know is better than the devil you don’t know. The result would almost certainly be a mass of very short-sightedly conservative decisions if most people voted on most issues. If, on the other hand, only those with some stake in the issue tended to vote in any particular referendum the result would be a large change of electorate from issue to issue, almost certainly resulting in inconsistent decisions. Inconsistent directives cannot be obeyed. They must be amended, at least in part.

The only way that is not dictatorial is negotiation. It is theoretically possible, of course, that such negotiations might be controlled by a series of iterated referendums, but the cost in time and effort and the exiguousness of the benefits even to the collectivity, let alone the individual, rule such a solution out in practice, at least in societies of the scale and complexity of ours. On most of the issues that affect us most of us have no strong opinions. We would like them to be decided by able, well-informed people with our interests at heart. We live in a world of specialisation. As Adam Smith argued it is the basis of the private wealth of nations. It is time to give full scope to specialisation in the public sphere as well.

Voting in elections

The electoral system is supposed to fill this need. Very often the strongest single influence on the outcome of elections is not substantive issues as much as the candidate’s or party’s competence and performance.6 The swinging voter is, by definition, not strongly committed on the outstanding issues between the major parties or coalitions. He or she often regards the choice between the parties on issues as a matter of emphasizing now one set of 70factors, now another, as circumstances change. But whether such emphases are effective in practice depends on the skill and responsiveness of the executive branch of government. Better a capable and reasonably responsive and responsible executive of a colour that is not to one’s taste than a more congenial but incompetent or inflexible government. The question, then, is how well can voting perform when the task is seen not as the hopeless one of producing policy but its normal task of producing governments?

The usual way of choosing political representatives is by geographically based electorates each choosing one or more candidates to represent that electorate. There is no doubt that electoral systems are always very artificial contrivances which are sensitive to only a few of the factors that might reasonably be regarded as relevant.

Simple plurality voting with single-representative constituencies usually results in more conservative decisions than would result if minorities who are strongly dissatisfied with the status quo could express the strength of their preferences.7 No doubt a certain amount of vote-trading does go on between interest groups through the brokerage of political parties. But the power of any group to make its desires count in that context is much more a matter of their having enough of what other groups want to trade for concessions than of the strength of their desires. It is true, of course, that not all people vote to secure some specific interest of their own. Many identify their own broad interests with the interests of just and sound administration. They are prepared to support groups that are manifestly disadvantaged and also to support platforms and candidates that promise to produce a reasonable balance between conflicting interests. But their perceptions and understanding of many of those interests is bound to be very incomplete and overlaid by conventional stereotypes. The benevolent concern of others is no substitute for power to make one’s own needs count.

Another nest of difficulties is connected with the well-known problems of voter’s order of preference among alternatives. I have only one vote on a certain question and there are three alternatives, and my first preference is rejected. Should I not then have a vote about which is to be preferred between the other two? Systems of preferential voting such as those used in Australia or of iterated voting such as is used in France, are attempts to deal with this problem. Again, in most systems of voting an interest group might constitute almost half of every constituency and not win in any. By contrast a group of similar aggregate numbers concentrated in a smaller number of constituencies might emerge as the most powerful group, winning every constituency in which it is present. It is notorious that there is no general solution to such difficulties. Any solution will tend to have very different outcomes, some of them highly undesirable, in different circumstances. 71

It is often possible to design a voting system that under the given circumstances will approach more nearly than any other to ensuring that major identified interest group can be represented roughly in proportion to its numbers. For such systems to be workable it is usually necessary that the number of interest groups to be represented be relatively small, cohesive over a wide range of issues and fairly clearly distinguishable from each other. Nevertheless, the representation can never be such that it is manifestly fair and guarantees that the emerging government will be fully representative. So most contemporary discussions insist that the problem to be solved is that of producing an unequivocal and reasonably satisfactory answer to the question, Who is to form the government in a nation-state?

The point of voting is not so much to represent the variety of interests of those who have a vote but to produce a decision that will be accepted by nearly everybody as final. This comes out most clearly in relatively small committees, where the vote is essentially a device for terminating discussion at a point where some suitable degree of consensus has been reached, or where it is impossible or unprofitable to prolong the discussion. The worth of the vote is almost wholly a function of the quality of the discussion and negotiation that have produced the alternatives that finally come to the vote.

In practice, in all modern electoral systems, that comes down to the soundness of political parties as a means of forming governments. Where the choice is between two parties each with a comprehensive and well-thought-out programme, there may be little to choose between them. The vote amounts to not much more than the toss of a coin. At the other extreme, when all the alternative governments are unacceptable there is little that votes can do to remedy the situation. So the worth of electoral democracy is closely bound up with the processes of party politics.

II Electoral politics

The function of elections is usually thought to be to produce a government, whether in a nation-state, a province or even a voluntary organization. The government is produced by competition among rival parties for the requisite number of votes of the electorate. Once elected a government is normally not subject to recall until the next election occurs. According to some it is very desirable that the elected government be free to pursue its policies, within the limits of the constitution, until it must face the electorate again. According to others the elected government should be constrained by as many organized pressure groups as are needed to make it responsive to the diverse needs in the community. However, even in the latter case the pressures that can be brought to bear on a government are usually effective in proportion to their bearing on elections to come. 72

The classical view, expressed by Mill, was that the electoral process should be so designed as to encourage voters to elect suitably qualified community representatives rather than agents for their specific interests. Such a view is, surprisingly, often reinforced by the development of highly disciplined parties. A great many people judge political parties not so much by their programmes as by their style of political action. Let us grant for the sake of argument that it is more important that the business of governing be well conducted than that some specific minority interests they favour be promoted at the cost of considerable failure in other areas. The problem is whether such assessments can in fact be made on a sound basis and given effect in the electoral process.

The process itself unfortunately tends to produce distorted results in quite systematic ways. Some of these are:

Mystifying issues

There has been much complaint and misgiving about the use of advertising techniques and image-making in electoral campaigns. But the mystification that matters most begins much further back. Overwhelming pressures to lie, to pretend, to conceal, to denigrate or sanctify are always present when the object to be sold is intangible and its properties unverifiable until long after the time when the decision to buy can be reversed. Sound strategy demands that in bidding for votes one pays out as little as possible in clear commitments and plays down any unpopular aspects of policies one is committed to. Usually it pays the opposition to concentrate more on spinning their own deceptive webs than on attempting to pin their opponents down. They fear being pinned down themselves. It is safer to stick to slogans and empty rhetoric, or to play on people’s fear and prejudices. It pays to fight myth with myth.

Agglomeration of issues

The voters have no opportunity of separating issues out from each other. The result is that voters with a single dominant interest can be made to vote for a package of things they do not like in order to secure their major interest. Of course it is necessary and desirable that people have to make concessions to other interests in order to protect their own. The problem is that the packages are constructed, not with a view to getting the most acceptable compromise, but to achieving the one that suits the political strategy of the constructor.8

Professionalization of politics

The constructors of programmes and political images are increasingly professionals who organize and use power to advance their own careers. The progress of careers in an organization such as a party is determined mainly by the internal politics of the party. This would not matter, perhaps, if the processes of those politics resulted in coherent choices. But in the nature 73of the case they cannot generally do so. What concessions are made, and where, become a matter of contingencies that bear no systematic relation to the substantive problems. Some who have a weak case will be appeased because they are powerful or strategically placed. Others with stronger needs will be ignored or dismissed with mere rhetoric because they are not well organized or articulate. The needs of particular politicians to mobilize support determine which voices are recognized and which unheard.

These complaints are familiar enough. Because we are so familiar with them it is difficult to appreciate just how serious they are in practice. Clearly, in a stable and relatively homogeneous set of electorates they may not be thought to matter very much. In fact it may be argued that the interests of voters are more likely to be served by professional powerbrokers than by committed people who act on principle. People may be better served by those who are lookng for clients than by those who are looking for followers. It is not my task to adjudicate this debate. The crucial point is that neither alternative is satisfactory. The leaders will sacrifice their followers to their own visions and power-brokers will sacrifice weak clients to retain or gain strong ones. Both will attempt to narrow the choices open to their supporters, to prevent alternative leaders from emerging and to confront their followers or clients with an all-or-nothing choice.

In these circumstances it is very questionable whether elections can be regarded as anything more than a working arrangement for bringing about changes of government. Such changes are necessary from time to time if the political process is not to become too fixed in one mould. The defence of electoral democracy is reduced to the claim that it is preferable for this reason to other systems of government. Notoriously other systems of government, whatever their merits, usually fail at this task. To look upon elections as a means of choosing the best representatives is to expect too much of them. The performance of some candidates is known. The rest merely show some promise. There is little basis for most people to make a sound judgement. Again, when it comes to policies, the set of proposals that a party puts forward is usually in response to a situation that will have changed significantly by the time they can be implemented. One trouble with politicians is that they often break their promises. Another is that they often keep them, in spite of their no longer being appropriate.

Granted the impossibility of arriving at rationally aggregated preferences by any mechanical procedure and the desirability of negotiated settlements of conflicts that exploit the possibilities of the specific situation, it is probably desirable that governments not be precommitted to very specific proposals. This is particularly so if the proposals can be implemented from above only by increasing bureaucratic control, which is likely to be costly, often ineffectual 74and always difficult to direct and contain. Politics oscillates between efforts to use the government as a means of social change and disillusion with the results of such efforts and the costs they impose.

Negative control of government

The more important side of the process of political accommodation occurs not in the electoral process but in the process of government itself. To the extent that there is a large variety of organized interest groups that are free to combine or manoeuvre among themselves in response to government initiatives the power of government is restrained. Such groups can greatly increase the cost of initiatives they dislike by refusing to co-operate, by mobilizing the opposition of other groups and by playing on the fears of the public. The task of stopping a measure is very easy where there is no onus of providing a positive alternative. A dozen conflicting interests may be united only in their opposition to a particular proposal. If their aim is simply to stop the proposal their very diversity of interests can be a source of strength. If each insists that only a proposal that meets all the interests of the parties is acceptable, the line against the objectionable proposal can be held very easily, since the coalition cannot be broken down. The costs of any member deserting are too high. To betray all the other parties is to invite retaliation, or at least strong distrust.

Increasingly the politics of lobbying those in power also becomes professionalized. The profession of politics at every level becomes one of power-broking and patronage. The vast range of appointments to various offices that governments have at their disposal are used to buy support for those in power. Even those who are not just venal careerists are forced to join in the scramble for position. Without an institutionalized power base one has no leverage within the system. Electoral politics merely reinforces this tendency.

The primacy of power-broking

In order to be elected one has to attract votes. In order to attract votes one has to be known and assessed. In most cases people do not and cannot have adequate information to assess a candidate’s policies, ability or moral qualities. So they rely on the verdict of a party that endorses candidates. Such reliance is notoriously very weakly justified. Whatever the various methods used by parties in selecting candidates, in practice the decisions are made by a small number of people, whether they be a central committee or the local branch activists or the party backers who support candidates in primary elections. These groups almost always represent a quite narrow range of interests and make their choices on strategic or tactical grounds rather than on the merits of the candidates. The faithful party hack or loyal worker for the 75faction is preferred whenever it is possible to do so without risk of electoral failure, and sometimes even at the cost of failure at the polls.

The process of selection of candidates does very little to ensure that able legislators and administrators are selected, except perhaps for some of the most salient positions. It is, I think, fairly obvious that many of those who are elected to high office in most democratic countries are undistinguished in most relevant respects. The reason many abler and better people give for not going into politics is that they are not well equipped to deal with the continual jockeying for influence and position in party life, the toadying to those already in positions of power, the necessity of discrediting others rather than co-operating with them, the subordination of issues to tactics and so on.

Of course, these features are not confined to political parties. They are characteristic of all organizations where there is competition for advancement and no firm, clear and independently sanctioned criteria of suitability for advancement. They affect parties in single-party and non-democratic regimes even more severely than democratic parties. Clearly elections impose limits on the degradation of selection criteria. Even hard-core party supporters will refuse to vote for some candidates.

Nevertheless, electoral politics does breed party politics, and party politics breeds mediocrity and corruption. Even where the local branch of a party is largely autonomous and formally democratic it is extremely easy for a cohesive group to gain control and manipulate rules and procedures to keep others out, unless they are even better organized and more determined. In practice such organization and motivation are usually found only among fringe or extreme groups who want to use the party as a means of advancing their interests rather than the ones for which it stands in the eyes of most voters. Such tactics may on occasions result in a rejuvenation of a party and of the political life of the nation. More often they degenerate into a sterile “numbers game” which soon alienates from political life all those whose main interest is not power but substantive issues.

Parties choose the policies and teams of candidates that are put to electors, not on the basis of the needs or the wishes of the electors but on the basis of the relative power of various groups in the party and their specific interests as powerbrokers. To be successful in politics one has to play this game and play it to win. There is no way of making parties much better than they are. A party is an organization designed to enforce a political line by exercising as much control as possible over its members. To gain electoral success it must confront the voters with a narrow choice between accepting the package it offers and an unacceptable alternative. To do so it must guard against splits at all costs. It must be strongly disciplined. Discipline is attained only by concentrating power to enforce decisions in relatively few hands, together 76with judgements about how to use that power. A party cannot rely entirely on self-discipline and co-operation among its members precisely because they are competing for power among themselves. There is little room for loyalty in power politics. The ruthless, the opportunists, those who are prepared to break their undertakings, very often win, and winners are rarely punished.

Inappropriate centralization and decentralization

Where party discipline is weak, power in many matters devolves on to local branches. The most frequent result is “pork-barrelling”. The local politicians attempt to promote those issues and persons that combine a reasonable electoral appeal with the greatest possible advantage to those who control the local party branch. Electoral politics tend to decentralize power in matters where more central control is desirable and centralize power where decentralized control is desirable. Obviously this effect is difficult to prove and will be very variable according to a number of other factors. However, a number of important mechanisms are at work to produce the tendency, and there are few countervailing ones.

Let us look first at inappropriate decentralization. The most obvious factor here is local chauvinism. It is always relatively easy to mobilize local support for a policy that brings tangible benefits to a local community or enables it to hold on to advantages it possesses, no matter what the effects of its doing so may be on outsiders. It is almost inevitable that chauvinistic issues will be much more strongly supported and less divisive than broader issues. So there will be a strong “democratic” pull to make the electoral system sensitive to them. On the other hand, issues that concern a more generalized devolution of power have no special appeal to any one local electorate. The more general issues are perceived as remote and less important.

Where decentralization simply means centralization on a smaller scale, bringing as many issues and powers as possible together under a single local authority, it is of very dubious value. The agglomeration of issues is, if anything, more pernicious at the local level than at the level of the nation state. At the local level it is much easier for the formal legal and institutional means of control and the informal power structures of the community to be knit together so closely as to suppress any effective challenge to them. Almost all issues come to be considered primarily in relation to the problem of preserving the cohesiveness and power of the local community. In practice, of course, that usually means the power of a small group within that community. But even if power is not highly concentrated within a local community, there are very narrow limits to the amount of variation that it can allow while still maintaining itself as a total community, an appropriate locus for centralizing almost all decision-making. It must resist changes that make for stronger links 77between members of the community and outsiders. It must deny the validity of claims that outsiders make upon it.

To sum up, decentralized democratic control by electoral processes is likely to produce a static and oppressive local chauvinism of a very conservative kind, the mirror, in fact, of the worst aspects of nationalism. I am arguing throughout that an appropriate form of decentralization would concentrate on functions, rather than localities. The organizations that supply various community needs should be independent of each other. Their geographical extent should vary according to the technical and social exigencies of their carrying out those functions. Such a great diversity of authorities with different circumscriptions could not be managed by electoral processes, as we shall see. But that is no disadvantage.

That electoral control leads to inappropriate centralization is less controversial. As we saw in discussing bureacracy, it is often the case that higher layers of bureaucratic control have no other defensible function than to bring a number of functional units under central control, not in the interests of efficiency but simply for the sake of control. Electoral politics involves putting the responsibility for the production of common goods in the hands of unitary central authorities. These authorities must find means of controlling the agencies for which they are responsible. Control has to be centralized irrespective of efficiency or flexibility. Putting each of these various agencies under its own popularly elected board would not solve the problem. In practice most electors could know very little about the candidates and the issues in the host of institutions by which they are affected. In most cases they would have to vote for party tickets, which is really a way of returning these bodies to centralized control.

Voting systems

A great deal has been written about the properties of various systems of voting, especially plurality, majority and proportional systems. A fully adequate dismissal of electoral democracy would have to show that none of these systems can remedy the defects we have discussed. There have been numerous discussions of the relative merits of the various methods of computing votes both from highly theoretical and very practical points of view. But these discussions normally assume that the problem is that of providing a government for a nation-state, which is just the assumption that is under attack in this book.

Very roughly, plurality and majority systems tend to produce two large parties or stable coalitions that in most circumstances compete for the middle ground and offer alternative administrations which differ only in emphasis. It is possible, however, especially in a plurality system, for the parties to diverge, each pursuing a fairly extreme policy and hoping that the middle ground will 78be more afraid of their opponents’ extremism than of theirs. Proportional representation seems to offer less possibility of government being dominated either by the middle or by an extreme of the political spectrum. In fact, however, proportional representation does not ensure a multiplicity of parties corresponding to the various hues of the spectrum. In many countries it works in what is virtually a two-party manner, sometimes to bring those parties into convergence, sometimes into confrontation. Even when it does result in a multiplicity of parties these tend to have more destructive than constructive power. It is much easier to immobilize the governmental machine than to set it going in a specific direction. The fact is that there is no voting system that cannot be manipulated by a large, well-organized and well-placed political party to ensure that it gets a great deal more of the power than its proportion of the vote would warrant.

In effect, the achievement of electoral systems is that they make it possible peacefully to eject a government and replace it with a government that is at least less disagreeable to the largest cohesive group of voters. But this gain is a gain only when compared with other systems of tenure of office that rest on heredity or co-option or military force. It probably has few disadvantages that are not shared by those other systems, but equally it shares most of their disadvantages. It inspires the loyalty that it has won mainly because the known alternatives are so repulsive. We can do better.

III The alternative to electoral democracy

Negotiation and size

I have already argued that most issues of any complexity need to be settled by negotiation in a co-operative framework. The most advantageous feasible solution for each party is often possible only if it is prepared to make concessions to other parties who are in a position to block many of the possible outcomes. In any case there are almost always long-term advantages in encouraging reciprocity by striving to achieve solutions that are at least seen as fair by all involved. Negotiation involves attempts to construct comprehensive packages through a process of exploration of possibilities, expanding or contracting the scope of the packages, trying to get clear about what is important and what is negotiable for each party.

Such work must be done in committee. The point of the small committee, the reason why large assemblies often refer problems to them, is that the members of the committee can spare the time to familiarize themselves with all the details of a matter in a way that is not possible to most members of the assembly, granted the range of other problems demanding their attention. Moreover, in informal debate in a committee, the members can explore and 79construct a wide variety of possible solutions to a problem and attempt to find ways of conciliating divergent interests. The larger assembly, by contrast, is usually capable only of accepting or rejecting a given proposal, perhaps with some minor amendments.

Committees, precisely because they cannot claim to be fully representative of all the interests involved may come under strong pressure to demonstrate that they have made every effort to take them all into account. Committees strive to achieve as much unanimity as possible in their final proposals because unanimity enhances the force of their decision. Committees tend to accommodation, large assemblies to confrontation. In the large assembly there is neither the time not the capacity for detailed negotiation, and partly because of this each agent is usually reduced to representing a single interest. Each member of a large body can have some impact on the outcome only by making some single clear and telling point. By and large in such situations it is easier to be effective in stopping proposals than in proposing constructive alternatives.

The large assembly is usually forced to bring the question to a vote whether or not it has been discussed adequately. In modern parliamentary situations party discipline ensures that government sponsored legislation goes through with the minimum of amendment. The committees in which the real work of negotiation is done conduct their meetings in secret, sometimes because they are in fact clandestine, sometimes behind a veil of officially sanctioned “confidentiality”. Clearly it is desirable that negotiations on public policy decisions should be conducted publicly. At the same time, however, the negotiators should be free to consider alternatives without being called to heel by those whose interests they represent. The process of exploration of possibilities cannot proceed if pressure groups are in a position to lay down in advance what is negotiable and what is not.

Representation and function

Two main reasons why one ought to entrust the furtherance of one’s interests to somebody else are the opportunity costs of pursuing them oneself and one’s lack of the appropriate knowledge and skill. My representative should have at least as strong an interest in advancing my interests as I have, should be in a position to devote more time and effort to the task and bring to it superior knowledge and skill. It is very unlikely that anybody can meet these requirements in regard to my interests generally, or even my interest in public goods. But once one specifies a particular interest, the local library, for example, it is quite likely that there are people who share my interests, have a stronger motivation to work for them than I do, and are better equipped to do so. 80

If the range of interests that I have, many of them conflicting interests, are to be properly represented, it is most unlikely that any one person will be appropriate.9 What I need is a host of specialized representatives each of whom has the same interest as I in the relevant respect, and the appropriate motivation, knowledge and skill. How is it possible for me to have such representation?

In most democratic countries a large number of common goods are provided at local level by municipal councils, which are elected by the residents (or property-holders) in a certain area and draw the bulk of their revenue from taxes, especially on landed property, that they levy on residents or propertyholders in that area. Among the functions of these bodies the salient ones are usually to lay down and enforce rules about and provide services in areas such as town planning and building regulations, public health, garbage disposal, parks and recreational facilities, local roads and drainage, libraries and some educational facilities, environmental protection and the promotion of local industry. The basic reasons why these different functions are vested in the same body are: (1) that it is the source of finance for them, and (2) that it is the means by which the exercise of those functions is controlled in some degree by the governed.

There is no particularly strong relation between the functions themselves. Many of them have stronger practical links with similar functions in neighbouring municipalities or in higher levels of government than with other functions at the local level. It is not always the case that the geographical area of a municipality is the most suitable unit from the point of view of efficiency of operations. It is not very likely that a library authority and a garbage disposal authority for example would break up into the same geographical units, particularly in the large conurbations in which most people in Western countries now live. The city, the town, the suburb no longer exist as communities in the way that they did when the structure of our patterns of local government was formed. People’s activities take place in a variety of localities and they belong to many overlapping communities, most of which have no precise location or boundaries or membership.

Of the two reasons given above why these various functions are united in the same body, the first, that it provides the bulk of the finance for them is historically important but not very cogent, since there is no particular reason why each of these functions should draw its revenue from the same source. Even if they did, the function of the fundraising body might simply be that of apportioning funds to a variety of agencies, leaving the responsibilities for deciding on what use is to be made of those funds to the agencies themselves, provided there is appropriate control of those agencies by some other means. 81

So the decisive reason for making a single body responsible for a variety of functions is that it is a way of keeping some degree of democratic control over their operations. What this form of organization does is to add an administrative level to the levels that are required to perform the specific functions and attempt to achieve democratic control of all of them through this level. The obvious alternative would be to have specific elections for the controlling bodies in each function. The advantages would be numerous. Representatives with particular interest in one or other of the bodies would nominate for it. The number of opportunities for people to take an active role in public life would be increased enormously, and many people who would not be willing to set themselves the whole range of problems of a municipal council would probably be willing to take a strong interest in a particular function, especially if in doing so they had a fairly free hand.

The central difficulty, of course, is that there would be too many elections. Each voter would belong to many different electorates, and most would lack the time, information and motivation to make any informed personal judgement about most of them. In practice most people, if they bothered to vote, would tend to vote for party lists or “tickets”. The net result might well be simply to strengthen the party apparatuses. The game of power-trading and building a power base would tend to take precedence over the specific issues involved. The value of functional representation would be lost.

Representation and sampling

The most reliable way of getting a group that is representative of a particular population is to take a statistical sample of the population. The theory and practice of sampling is now highly developed. Such a representative group could be trusted to act as representatives if they had some stronger than average motive for devoting themselves to the interests they represent and acquiring adequate knowledge and skill for the task. If they were compensated for the time and effort involved the mere fact of their being chosen to be representatives might be sufficient motivation for many. For others it would, like jury service today, be an unwelcome imposition.

It seems preferable, therefore, that representatives should be volunteers. Once a statistical characterization of the various interests to be represented is established there should be no problem in selecting from those who are willing to serve a group that is representative in the statistical sense.10 Granted that they volunteer to work in this particular area rather than some other or none, it is likely that they feel more strongly than most people about the issues involved and that they have or are prepared to acquire superior knowledge of the problems. So it would be reasonable for the group that these people represent statistically to accept them as representatives in decision-making. 82One could envisage a well-based convention granting authority to such bodies and appropriate procedures for selecting them.

The first thing that would need to be clearly defined is the function to be supervised and a reasonably comprehensive analysis of the various interests that people might have in the supervision of that function. In general the main groups affected would be those who work at providing the good in question, those who are consumers of it and those who are affected by its side-effects. But within these groups there might well be very substantial conflicts of interest and differences of opinion.

So the question immediately arises whether it is interest or opinions that are to form the basis of selection. There is no insuperable difficulty in either case. One could take a broad opinion survey, classify groups of opinions and choose representatives on the basis of their own answers to the survey questions so as to match the strength of the various opinions in the population. There would be some incentive for those with unpopular opinions to disguise them in the hope of being chosen. It is statistically unlikely, however, that this would affect the result substantially unless a very large proportion of those who volunteer to serve entered into a conspiracy to manipulate the sampling process. There could be safeguards against this possibility. Candidates might be required to declare publicly their beliefs and the interests they claim to represent and even be liable to legal actions for gross misrepresentation or fraud.

Such a system would still leave representatives a great deal of room for negotiation, and even for changing their opinions whenever they had defensible ground for doing so. It would approximate more closely than any other system to a fully participatory democracy. The main objection to it is that if people are not very well informed about a matter and express opinions about it in relative isolation those opinions are not well-grounded. There is no good reason for taking them seriously as the basis for structuring the body that is to make the decisions.

The objection could be met by making the survey not a matter of “off the cuff” answers to a questionnaire but an educative and exploratory exercise. A demographically representative sample could be chosen and the people in the sample could be given special information and opportunities for discussions about the issues before giving their answers. Perhaps they might even be paid to take the time to study the problems. Such measures might be expected to improve the quality of the basis for choosing representatives, the quality of the representatives and the general quality of on-going debate in the community about such matters. They would not abolish conflicts of interest or differences of opinion, but they would create an atmosphere of rational open 83and constructive discussion in which optimal solutions would be most likely to emerge and gain acceptance.

In most situations, however, such an elaborate method of choosing representatives might not be warranted. It would be simpler to choose them on the basis of an analysis of the interests involved. Interests are more important and even more cogent in the long-run than opinions. Rationally I should prefer my interests to be safeguarded rather than have my more or less shaky opinions prevail. In practice, of course, I may often set a higher priority on getting my own way than on securing my long-term interests because of pride, spite, impatience, stupidity or other human failings. But most of us in most matters that affect us do not have such a definite will. Even when we know what we would like to happen we are well aware that if we knew all the relevant facts our practical choice could well be different.

Moreover, in many cases there is no “right” decision. Our long-term interests are not fixed. Different and largely incommensurable possibilities exist. Choosing one or other of those possibilities means entering a different situation in which some of our tastes, hopes and relationships will change. There is no unchanging point of reference, no fixed set of needs or preferences, against which such possibilities can be so measured and evaluated as to produce a right answer. There are indeed many wrong answers. It is not the case that we are infinitely malleable. We need to be assured that clearly wrong choices are not made. But the quest for a way of finding the right answer is radically pointless.

In general it is more in our interests that a variety of possibilities be explored than that attempts should be made to discover and impose what purports to be the best way of dealing with a problem. The great strength of the case for the superiority of the market over planning is just such a point. However, allowing great opportunities for diverse solutions to problems, diverse projections of needs, and differing opinions to be tried is not an exclusive property of the market. If representatives in charge of various functional enterprises had a free hand to try whatever solutions they thought best there is little doubt that a great variety of ways of doing things would be tried. I shall argue that this would be an advantage and that it need not carry with it any serious disadvantages.

Representation and responsibility

In some circumstances we may sensibly consent to be represented by a person over whom we have no control. In some circumstances we may reasonably be presumed to consent to being represented by somebody without even knowing what is going on. But the relationship of representation is undoubtedly stronger in proportion to the degree of willingness of the person represented to accept the representative. In political contexts it has commonly 84been assumed that the degree in which a polity is democratic is proportional to the degree of control that the citizens have over their representatives. Clearly acceptance is not enough or any legitimate regime would be democratic. Some authoritarian regimes may have enjoyed as high a degree of acceptance as many a democracy.

What control involves is not as simple as it might at first seem. If one is in a position to decide quite determinately what one wants the degree of control one has over a representative is just a matter of one’s capacity to ensure that the representative acts as effectively as possible to implement one’s decision. Corporate bodies, especially large bodies of people with diverse interests in relation to a type of activity, are not in a position to have a fully determinate will about the practical matters to be decided. They may, of course, have many areas of agreement and share a number of aspirations, but a host of issues remain to be resolved. There is in general no correct resolution of these issues and no correct procedure of arriving at a resolution of them. In other words, there cannot be anything like Rousseau’s general will, as he himself recognized, except in the very simplest societies. There can only be conventions to accept certain results or decision procedures for the sake of getting things done. So, for example, there is no right system of voting. As long as the results are not too disastrous we agree to abide by the results of the system we have simply because it is accepted. It is the convention.

But do we, in order to have democracy, have to find a way in which the demos first makes up its mind what is to be done and then controls its representatives in the process of carrying it out? What I want to suggest is a different conception. Let the convention for deciding what is our common will be that we will accept the decision of a group of people who are well informed about the question, well-motivated to find as good a solution as possible and representative of our range of interests simply because they are statistically representative of us as a group. If this group is then responsible for carrying out what it decides, the problem of control of the execution process largely vanishes. Those directing the execution process are carrying out their own decisions. They may need a little prodding to keep them up to the mark, but there is no institutional basis for a conflict of interest between bodies responsible for making decisions and those responsible for execution. They have an overriding interest in showing that their decisions were practical and well-grounded.

Granted that this process may be expected to generate effective social decisions and actions the question comes down to the sense in which the people control the decision-making process. It is useful to look at this question first statistically and then in terms of the dynamics of representation. If the group making a decision is statistically representative of the group 85on whose behalf it is made then it is very likely that the decision will be in accord with the result of some reasonable decision procedure for that group. The statistical selection procedure controls the distribution of the interests represented and so controls the decisions that are likely to emerge by rational negotiations among those representatives. Granted a sound statistical procedure the people automatically control the broad outlines of the result simply by being what they are. The mapping ensures a correspondence between the character of the representatives and the represented.

Again, the group of representatives are all themselves members of the people. Nobody is selected because of professional knowledge or skill or prestige or privilege or institutional position. What the representatives are expected to bring to the decision process is that sensitivity to the interests involved that comes from having those interests. They are supposed to be typical of those affected by the decisions to be made. The decisions that the representatives will arrive at will differ from the actual wishes of those they represent because better information and the results of negotiations will make a difference. However, this is precisely the divergence from one’s actual views that a rational person is normally willing to accept in the actions of a delegate.

Naturally, the mere fact that representatives come from a certain social group does not guarantee that they will cherish the interests of that group. They may be quite deluded about what those interests are. They may put some other interest they have ahead of the interests they are supposed to represent. They may be bribed or bought off. What safeguards are there against these things? One safeguard against delinquency is to suppress the causes that tend to produce temptations. Most of these fall within a wider context of property arrangements and social rewards that we shall look at later on. The crucial question, however, is that of public scrutiny. Everything that these committees do must be open to inspection. It is not difficult to ensure this if there are enough people around who follow closely what is done and have a strong motivation for making public anything that others would like to conceal.

If political office is attained by lot there are no professional political careers. Nobody has to acquire debts to party organizations or patrons in order to gain office or to hang on to it. The usual pressures to keep quiet are absent. On the other hand, the greatest reward that a person is likely to get from public service is recognition of his or her ability and integrity. In any committee, granted that there will be a diversity of interests and loyalties there are almost certainly going to be people who will expose any attempts to suborn members of the committee.

Moreover, since membership would change in a regular pattern there would rapidly grow up a body of experienced, informed people who would have an interest in the doings of their successors and try to expose their errors 86or delinquencies, if only to enhance their own record. More positively, it seems very likely that high standards of presentation of proposals for public discussion and responsiveness to the results of such discussion would emerge. Since the representatives would not be up for re-election they could afford to look at the various objections and suggestions raised on their merits. In some cases they might do things that were quite unpopular with a majority in the interests of a disadvantaged minority. In others they might assert a majority interest in spite of the objections of a powerful minority. The cheap rhetoric of political discourse would lose in representative committees. Ultimately their motivation would be to be recognized as having done a good job. Whether or not to be held in high regard is the strongest of motivations for people generally, there can be little doubt that it is always a strong motivation for the sort of people who are attracted to public office.

In a large and complex society very few people can be well known to the general public. Very many people, however, can be well known in some specialized activity. If that activity becomes the focus for a partial community, fame within it can be a significant reward, especially if the activity is in turn seen as important to the wider community or complex of partial communities.

Representation and co-ordination

It is not too difficult to envisage the solution I have proposed to the problem of democratic control of functional units producing specific public goods and services. The problems of co-ordination between these units are in many respects much easier than is often imagined. For the most part they can be left to negotiation between the bodies concerned, once the obstacles that bureaucratic structures place in the way of cooperation are removed. To a large extent the common good consists in the good functioning of a number of services that conjointly meet our various needs. The substantive criteria of good functioning are derived not from some higher-order set of concepts but from the specific tasks and the social realities in which they are embedded.

The point of democracy is to govern society from within the specific groupings that constitute the relevant population for each of the different public activities. The whole grows by the growth of the parts. The whole is integrated by each part adjusting itself continuously to the others with which it interacts. The illusion that democracy can be assured by so-called democratic control of the state is disastrous. The state cannot be controlled democratically. It must (ultimately) be abolished.

Of course, it is desirable that general movements of social thought and aspiration in the community should find expression in the overall direction of public policy. There is no reason, however, why these emphases and preferences should not be developed through their influence on the decisions of the representatives who share them. In this way they are likely to find 87more appropriate, diverse and direct expression than if that expression were mediated by electoral and bureaucratic processes. The spirit of public policy should in these circumstances reflect the spirit of the most articulate and responsible members of the community.

Higher-level bodies

Nevertheless there remain important problems in setting up the various functional bodies, hearing appeals about their structure, restructuring them to meet changed circumstances, adjudicating their disagreements and dividing up resources between them. Determining these matters would be the task of higher-order bodies which would seek to adjudicate conflicting claims in the light of generally accepted criteria. Such bodies, on the view I am advocating, would not be empowered to initiate policy, much less to dictate to various functional bodies, but to provide a legal framework within which productive bodies operate. There are important questions about how such bodies are to be prevented from exceeding their legal role and becoming administrative authorities.

For the moment, however, I am concerned about the sense in which they ought to be representative. Very roughly, what I want to argue is that it would be appropriate for them to be representative neither of interest nor opinions but of the social experience of the community. The people chosen to staff such bodies should have substantial firsthand experience of the problems and practices of first-order functional bodies. They should be able and trusted. The best way to meet these requirements would be to choose them by lot from a pool of candidates nominated by their colleagues on first-order bodies as having shown the special skill, knowledge and dedication that would fit them for a judicial role. They would constitute a sample of those judged most suited to the task by those in the best position to know.11

In their decision-making such bodies would work by a case-based rather than rule-based procedure, arguing by analogy for the relevance of various considerations and moderating precedent in the light of new circumstances. Since they, too, would have only a limited term of office, there would be plenty of opportunity for variation within a reasonable continuity. No professional class would emerge but a body of interested, informed critics would be formed who would be at liberty to criticize judgements. There would not be any need, I believe, for protection of these legal processes by the permanent threat of sanctions. I assume that responsible public functional bodies would not be nearly as likely as private citizens or corporations to reject a reasonably constituted judicial body’s authority. In most cases they would be brought to heel by sanctions from other bodies if they did.

In extreme cases, however, it might be possible to defy such judgements with impunity. Their being open to this threat would, I believe, operate as 88salutary restraint on any tendency of judicial bodies to step beyond their proper role. The society would cohere only as long as the conventions on which it was based were respected. Ultimately, of course, that is almost trivially true of any society. However, in a society where there is a centralized and overwhelmingly powerful police and military force, control of a sort can be maintained as long as the conventions that ensure control over that force hold. In the society I am envisaging the cohesion of the whole is dependent on a much more dispersed and varied set of conventions holding, simply because of the variety of social pressures that tend to maintain them.

Anarchy and community

In Community, Anarchy and Liberty Michael Taylor argued that anarchy in the strict sense, a social order without a state apparatus, is possible only if there is community. The three “core characteristics” of community in Taylor’s sense are: (1) The set of persons who compose a community have beliefs and values in common; (2) Relations between members should be direct and many sided; (3) There should be a high level of reciprocity, covering such things as exchanges, mutual aid, co-operation and sharing, as well as mutual enforcement of these arrangements.12

What I am suggesting, in effect, is that we can envisage a community of organizations rather than a community of individuals as constituting a modern society. It seems to me very doubtful whether the variety and capacity for change that are characteristic of modern mass technologically advanced societies can be preserved in small communities based on individuals. In modern societies individuals belong to a multiplicity of partial communities and organizations that are continually changing. Within various specific contexts they may have relations with other individuals that are communal, but their relations with the society as a whole are much more formal, legal, inflexible and impersonal. Taylor and many other anarchists want to restore small, face-to-face, relatively comprehensive communities. I am very doubtful whether this is either possible or desirable, though I cannot argue the matter in any detail here. The point I am trying to make is, in any case, independent of the possibility or desirability of community among individuals under modern conditions. Rather it is that there can, even on the modern scale, be community among certain kinds of organizations and partial communities, without their being a total community at any level.

Obviously a set of organizations devoted to producing a variety of public goods governed by representative members of the communities affected is likely to have many shared values and beliefs, particularly values and beliefs concerning the standards and procedures by which the organizations should operate. They will deal directly with each other and since each is highly specialized it will in practice have relations of interdependence with a number 89of other bodies. Finally, the pattern of exchanges between these bodies will be a matter of mutual trust and co-operation backed by sanctions of withdrawal of co-operation. The individual men and women responsible for carrying out the various functions that constitute these organizations would find themselves constrained to act in accordance with a communal pattern. Everything in their social relations, all the pressures on them, would conspire to make them conform to their roles.

Although that would not necessarily produce an atmosphere of community among those agents in matters outside their roles, it might well do so. It would be possible, pleasant and usually advantageous that there be an atmosphere of politeness, kindness, consideration, respect and friendliness among people. Such an atmosphere sustains a great deal of communal life even between strangers. It may not be strong enough to ensure the enforcement of social order. For that you need the tighter controls of the small community, as Taylor argues. But granted that the substance of social order is guaranteed on a non-repressive and co-operative basis a generalized sense of social co-operation can flourish fairly easily. In turn, such an atmosphere makes the task of ensuring social order very much less difficult in the residual areas where there are temptations to crime.

In other words, a sense of community, interdependence and reciprocity is not restricted to small communities or to relatively self-contained communities. If there is a sufficiently strong sense, diffused through a network of partial but overlapping communities, that doing favours for strangers one will never meet again, acting considerately towards everybody or being open to other people’s points of view is itself pleasant and likely to be reciprocated, then the conduct of public affairs on a basis of reciprocity will appear natural. The principal condition for achieving such an atmosphere is to eliminate zero-sum conflicts from the structures of personal and social relationships as far as possible. In that way co-operative behaviour can become more than a veneer of morality and courtesy imposed on relations that constantly tend to confrontation. The ways in which entitlements to resources and rewards are generated and appropriated are crucial in this respect.

Protection of person and property

In a society in which people are free to change their jobs, their residence, their friends and their activities very easily, it is not difficult to kill, steal or injure without it being known who committed the crime. If it is known only in a small group it is easy to evade the sanctions that group can bring to bear. It is to be expected that many people will have strong motives for crime, and need to be deterred. In a complex and large society there seems no substitute for police forces. It is by no means clear, however, that there needs to be a 90single agency with global power to enforce the law in all areas, much less that such an agency should be the custodian of public morals. From the point of view of detention and investigation there can be a number of specialized agencies, as there are in effect within modern police forces. These agencies could co-operate rather than be under a single control.

The type of policing that seems to require an open brief is that of “the cop on the beat”, whose presence is supposed to deter criminals and ensure public order. That role went out with the motor car and the changes in the pattern of urban living. In any case, it was in its heyday primarily a means of controlling the poor, and especially the young, preventing the emergence of street gangs and the like. In so far as these or analogous functions remain in contemporary societies, there are better ways of dealing with them by provision of specific services and opportunities for those who have nothing better to do than fight or “disturb the peace”.

Once again, the point is that a local authority is likely to be too total and too repressive, a wider-based authority too powerful and too subject to corruption. Specific law enforcement agencies dealing with specific kinds of problems are much more likely to be effective, provided the problems of controlling them can be dealt with and the structure can be changed to meet changing circumstances. Once these problems are clearly distinguished, they can be met at different levels by different kinds of bodies. The problems of supervision of specialized agencies can be handled by representative committees with managerial and policy-making powers. The problems of restructuring must be dealt with in a way that avoids giving to the body that decides on restructuring proposals any overall executive power. In no area is it more important to avoid centralization of power. The decisions must take the form of arbitration between conflicting proposals to extend the jurisdiction of existing agencies or create new agencies. The arbitrating body must have no power to engage in police work of any kind. It must not direct the policy of bodies that do so except by way of settlement of disputes.

It seems likely that, given reasonably active organs of public opinion, there would be little difficulty in holding higher-order bodies to their adjudicative role. The lower-order bodies would be jealous of their autonomy, and to some extent jealous of each other. It is not quite so clear, however, that in some other contexts, such as police work, a particular agency might not come to occupy such a strategic position in the complex of agencies that it would effectively dictate to the rest in crucial matters. It could be, for example, that an agency set up to guard against computer crime might, in a highly computerized society, gain the power to control any information transaction and thus any formal social action in that society. Its operations, moreover, 91might be extremely difficult to monitor, and virtually impossible for a lay person to understand.

There is no doubt that this kind of development would be controlled only by constant vigilance by people involved in the ordinary running of the dangerous agency.13 These people would need to be given every incentive to keep the operations of the agency fully public, to discuss their problems and dangers and make proposals for change. There is, of course, an inherent tension between the need for a certain amount of guile in criminal investigation and the need for public accountability of the investigation. But the major deterrent to crime must always be the difficulty of concealing it in the first place. Publicity about the kinds of actions being taken to prevent crime is a sound deterrent if the actions are well conceived. If they are not well conceived the criminals are likely to be aware of it before the public in any case. Public scrutiny may detect problems if there is adequate discussion and it is certainly likely to encourage critical thinking within agencies. Such publicity about policy is quite compatible with discretion within the context of particular investigations.

In the day-to-day running of police agencies, as with other agencies, a balance will have to be struck between reasonable autonomy for the full-time professionals and adequate access to information for the lay supervisory bodies. There is no recipe for such a balance, and it cannot be denied that statistically representative bodies may from time to time be dominated by people who are so intrusive that they make sound professional work impossible, or so complacent that they do not exercise proper supervision. Nevertheless, compared to the present arrangements in most countries, it is hard to believe that such citizen committees would not on the whole constitute an enormous improvement. The links that at present exist between the people who control policing and the rest of the government powers, both constitutionally and informally through the party system, constitute a standing invitation to the abuse of police power for oppressive ends.

Notes

1. The classical work on voting procedures is Duncan Black’s The Theory of Committees and Elections (Cambridge, Cambridge University Press: 1958). More recent developments are summarized in D. C. Mueller, Public Choice (Cambridge, Cambridge University Press: 1979), which contains an ample bibliography. A great deal of this work is made up of analyses of models where a very small number of voters are involved, as the title of Black’s book indicates. Some of the more obvious problems about voting procedures can be 92handled if alternatives are suitably subdivided and there is a highly developed trading of votes among voters. But this sort of interaction soon becomes too cumbersome and costly once more than a few voters are involved. In general, then, most of this work has interesting applications where relatively small committees are looking for suitable decision procedures, but is not very relevant to mass voting. What it suggests is that the committees in which, in my proposal, decision power on public goods should be vested, might adopt a variety of voting procedures, depending on the sort of problems they had to deal with.

2. See K. J. Arrow, Social Choice and Individual Values (New York, Wiley: 1951; rev. edn: 1963). For a lucid presentation of the present state of the question see Mueller, Public Choice, p. 184 ff.

3. The best-known advocates of the view that economic analyses should be assessed solely on their predictive power is Milton Friedman. See his Essays in Positive Economics (Chicago, University of Chicago Press: 1953). For a critique of this position see M. Hollis and E. Nell, Rational Economic Man (Cambridge, Cambridge University Press: 1975).

4. The terms “intensional” and “extensional” are philosophers’ jargon. A context is extensional if one can substitute for a name any other name or description that refers to the same thing without affecting the truth of what is said. So, if Tom Jones ran into Bob Hawke he ran into the present prime minister of Australia, or the man standing at the bottom of the steps in front of Parliament House at such and such a time, or the person who satisfies any number of descriptions that pick out the same person. But of course Tom Jones may know that he ran into the man standing at the bottom of the steps without knowing he ran into Bob Hawke. Similarly, where preferences are concerned it is very often the case that our preferences are directed towards objects as seen in a certain light rather than as they are “in themselves”. So the connections, equivalences and differences between objects qua objects of preference become extremely complex and incalculable. It used to be thought that these complexities could be handled by psychological theory, the simplest candidate being a hedonist analysis explaining preferences in terms of the ways in which objects cause pleasure and pain. Bentham’s utilitarian calculus depended on this assumption. If there were some independent way of measuring pleasure, and if the relations between objects and the pleasures they give were simply causal, such a calculus might work. But in measuring pleasure we fall back on preferences which relate not to the raw object but to the object as mediated by our cognitive relations to it.

5. The prisoner’s dilemma is a situation in which two prisoners, unable to communicate with each other, are offered incentives to incriminate each other, there being no other evidence against either on the most serious charges. If 93A incriminates B but is not himself incriminated he will be set free and B severely punished, and symmetrically so for B. If neither incriminates the other both will receive a light sentence. If both are incriminated both will receive a heavy sentence. So if each tries for the best outcome for himself both will achieve the worst. Only by settling for second best can either hope to avoid the worst if the other acts in his own interest. This and analogous situations have given rise to a branch of mathematics known as games theory. See R. D. Luce and H. Raiffa, Games and Decisions (New York, Wiley: 1957). Many of the most interesting developments concern complex games in which various situations are iterated in diverse patterns. The development of the theory is severely limited by the complications that are introduced once the number of players goes beyond two. Indeed, many not very complex problems appear to have no solution even in principle.

6. Voter behaviour has been extensively studied. There is a recent survey in Iain McLean, Dealing in Votes (Oxford, Martin Robertson: 1982). However, I must confess that my assertions in this section are based on everyday knowledge rather than methodical studies. The issue of competence is very difficult to disentangle from policy issues. A politician may be acknowledged to be “very clever” but still regarded as incompetent because the policies he or she espouses are unsound. Again, the qualities that are seen to be required, particularly in a president or prime minister will be quite different in different situations and under the sway of different mythologies and emotions. Obviously historical, regional, and economic causes affect these things, and there is no way of disentangling the diverse factors and the levels at which they operate, much less quantifying their relative inputs. Recent studies seem to me to have been strongly biased towards showing that the electorate acts rationally. No doubt there will be a reaction towards socio-psychological explanations in the near future.

7. The problems of systems of voting are well brought out in the context of an argument for proportional representation by Victor Bogdanor (Multi-party Politics and the Constitution, Cambridge, Cambridge University Press: 1983).

8. It may well be the case, for example, that a party obtains power by buying the votes of a number of diverse minorities, and proceeds to enact a package of measures that are contrary to a number of significant majority interests. This may well be possible, because the majority interests may be too widely diffused to organize effectively. What the politician is virtually forced to do is to construct policy packages with an eye to strategically placed parcels of votes. In different systems this has different effects. So, for example, in the US there is much “pork barrelling”, buying off of powerful local interests. In the UK this is not very significant. On the other hand parties in the UK rely 94on class loyalty to keep most of their followers faithful and woo the middle ground.

9. So, for example, I may have conflicting interests as a producer and a consumer, or more generally as a person who wants many things but cannot afford them all. It is not desirable that I settle in advance for some one balance between these conflicting interests. That may commit me to a range of consequences I could not foresee. It is far preferable that I be in a position to adjust the balance between my preferences in relation to the relative costs of alternatives as they emerge in the process of interplay with others. That is best achieved by having separate adjustments to each factor, and this is best achieved by my diverse interests each having its own representation. Each representative tries to do the best for a specific interest in the circumstances. If they do their work well the result ought to be a richer totality than would result from trying to meet the average voter’s preferences, since very often the problems do not take the form of zero-sum games.

10. I hasten to add that the point of statistical representation is to get a group that is sensitive to the interests of those affected. This does not necessarily mean that it is a micro-model of the relevant community strictly to scale in every respect. In fact I think it is preferable to have a fairly open-textured sampling procedure and rely on statistical variations over time to achieve balance. It is possible, however, to ensure a high probability of representativeness in quite small samples by using formulas to select the sample that score people along very many dimensions. In particular, there is not much difficulty in getting well-constructed samples of users and producers of goods and services. What is much more difficult is to identify and represent suitably those who are directly and indirectly affected by the production and consumption of those things. Assuming that the question of opportunity costs is covered by a process of adjudicating conflicting claims to resources, the problem comes down more or less to direct effects on others. The crucial question is one of differentiating legitimate from illegitimate grievances. In some cases it is an offence to pull down a tree on one’s own property or paint one’s own house a different colour without consent of the local authority in the interest of preserving the visual amenity of the neighbourhood. In some circumstances almost anything one does can constitute an affront to other people. What I suggest in this connection is that the formulas by which representatives are chosen should be open to challenge before a tribunal which would change them in those aspects where it considered that the legitimate interests of people affected by decisions were not adequately represented.

11. It would be possible, and, I believe, obviously desirable, for members of committees to be replaced one by one rather than all at one time. Each member coming on to a committee would be inducted gradually into its 95operations and build up relationships with existing members. There would be a sense of continuity of knowledge and purpose. It would not be possible for a new group to come in together and overturn existing arrangements without understanding them. A person who wanted change would have to build up allies in the committee, rouse public pressure and produce realistic proposals. Such efforts would be more likely to succeed in an atmosphere of constructive discussion than in an ideological confrontation. There would be a very high premium on presenting proposals as optimal for most of those concerned rather than simply as best for one’s constituency. In the absence of party discipline there would be no guaranteed support. Different people would need to be won over on each matter. It would pay to be attentive to the needs of others as a matter of general policy.

12. Michael Taylor, Community, Anarchy and Liberty, pp. 26-32.

13. John Mackinolty has suggested that most policing and punitive functions could be allocated to citizens in a way similar to jury service or military service.