Demarchy: A Democratic Alternative to Electoral Politics

Published in Kick It Over, No. 30, Fall 1992, pp. 11-13

Brian Martin

Go to

Brian Martin's publications on demarchy and democracy

Brian Martin's publications

Brian Martin's website

Subsequent commentary

Nils R. Connor, "Demarchy: it's not anarchy", Kick It Over, No. 31, Summer 1993, pp. 31-34: a critique of demarchy in response to the article above.

Brian Martin, letter, Kick It Over, No. 32, Fall 1993, p. 45: reply to Nils R. Connor's critique

Nils R. Connor, "Demarchy" (letter), Kick It Over, No. 34, Fall 1994, pp. 62-63: reply to Brian Martin's letter

Australian philosopher John Burnheim has invented the term "demarchy" to describe a political system without the state or bureaucracies, and based instead on randomly selected groups of decision makers.

Burnheim decided that the word democracy is so corrupted in meaning that it was better to introduce a different word for his proposed alternative. Although democracy literally means rule by the people, those Western societies commonly called democracies actually give the people little role in self-government. Admittedly, people are able to vote for political leaders, but only at infrequent intervals. They are certainly not able to vote for powerful figures in corporations and government bureaucracies.

Political scientist Benjamin Ginsberg has made a penetrating study of the structural implications of elections. On the one hand, elections allow mass participation in a way impossible under systems of bureaucratic or military rule. On the other hand, elections constrain participation. People are expected to participate only a specified intervals, and the participation takes the narrow form of voting for pre-selected candidates.

Ginsberg points out that voting is a potential threat to established elites. Certainly this is what has happened in the Soviet Union and Eastern Europe in recent years. But, just as importantly, elections serve to legitimize the system of rule by the state. Because people believe they have participated in choosing their rulers, they are much more willing to accept the system of rule itself. In fact, one explanation for the introduction of elections in the Soviet Union was to provide legitimacy to a failing system.

There are, of course, corruptions in electoral politics, including campaigning by mass media and image, special-interest contributions to candidates, and political trade-offs in parliaments. But Ginsberg's analysis goes deeper than this. Corruption actually weakens the legitimizing role of elections. Even a clean, fair electoral system inherently limits citizen participation.

Burnheim's critique of representative democracy is at this fundamental level. The power of the state is inherently antagonistic to local, participative decision making. It is also linked to severe problems, of which the foremost are war, repression and inequality.


Problems of Participation

But is there any alternative to the state? Can people govern themselves directly? There are some standard objections to greater participation, and they are not easy to overcome.

One objection is that people are not well enough informed to make decisions directly. They will be swayed by prejudice and demagogues. This is a good argument against systems such as electronic referendums, in which people might be persuaded by the latest media story.

The solution to the problem of prejudice is genuine dialogue and a focus on local issues that are meaningful to people's everyday lives. Members of parent-teacher associations and voluntary groups such as Rotary are noted for being responsible.

This leads to the next objection. People have only so much time, and there is simply not enough time for everyone to participate on all the issues. Indeed, there is not even enough time to become well informed about all the issues that might be dealt with on a local level.

This objection is usually taken as a definitive refutation of systems of direct popular self-government. The system of representation is said to provide the solution to this dilemma: citizens need only become informed about the quality of the candidates (and perhaps their stands on key issues). Democracy in a mass society is assumed to be representative democracy.

This assumption is easily shown to be false. There have been, after all, quite a number of proposed alternatives. There are the systems of initiative, petition and referendum. There is the Swiss system of decentralization, with canton-level government. There are also alternatives based on local worker and community self-management, in which delegates are selected for higher-level decision making.

But these alternatives all have difficulties with the problem of knowledge and participation. Referendums do not replace central government. Delegate systems can still produce the inequalities of power characteristic of elections.



Burnheim had a different idea. He had decided that the state and large-scale bureaucracy should be abolished. How then could decisions be made locally without running into the problem of knowledge and participation?

Random selection provides the first half of Burnheim's solution. This is a strongly participatory mechanism, since it prevents power-seekers from gaining exceptional power. But random selection could still allow centralized rule. Imagine, for example, the U.S. House of Representatives being replaced by a randomly selected "Representative House." (This has, in fact, been proposed.) Whatever the shortcomings of random selection, many would say that the result could be no worse than the present system. But the actual direct participation in national legislation would be minimal, since only one person out of half a million would be involved. Furthermore, since Burnheim sought a model without the state, random selection by itself was not sufficient.

The second part of Burnheim's solution is "functional groups." This simply means groups that deal with functions such as education, garbage collection, health services, transport, food production, manufacturing and so on. In each locality, there would be dozens of randomly selected groups, each dealing with a different function.

The number of groups could be determined to give a "desirable" level of participation. For the sake of illustration, let's say a community of 10,000 people had 200 groups each with 10 members, each serving strict two-year terms. The average person would then expect to serve 2 out of every 20 years.

Of course, right now there are bodies that deal with education, health and so forth. But they are not participative bodies, except sometimes through elections -- and elections select only certain types of personalities.

Let's look more closely at the functional groups. They are selected randomly from the local population. Doesn't that mean that there will be people selected who have no interest in the topic?

Burnheim at this point insists that the random selection be made only from volunteers. The people who volunteer for selection to the education group, for example, are likely to be those who know a lot about the issues and who feel passionately about them. Selecting from volunteers overcomes the problem of lack of interest. People who are happy with the decisions that are being made by others need not volunteer for any group.

What about the problem that some types of people might volunteer more frequently -- men more than women, for example -- leading to biases in the composition of the groups? This is easily handled by specifying appropriate sampling procedures for the random selection. For example, if 100 men and 50 women volunteered for a 10-person group, it is simple to pick 1/20 of the men and 1/10 of the women, resulting in a gender-balanced group. Similar procedures could be used for any other desired characteristic. In this way, those chosen would be people who both want to serve and are statistically representative of the community.

Another problem: how would decisions be made about these sorts of rules, such as the number of men and women, the size of the groups, age restrictions and so forth? Burnheim envisages what he calls "second-order groups" whose sole task would be to adjudicate on procedures for the "first-order groups," which are the ones that actually deal with community issues.

The second-order groups would be composed of people who had already served on first-order groups. They could be chosen randomly, or by consensus from their peers. The second-order groups would be like mediators or judges -- they would not have executive powers.

This raises the question, what power would the groups have, anyway? This is where the difference between demarchy and representative democracy really becomes apparent. Remember that there is no state: no central executive authority. That means, among other things, that there would be no military forces. (There might be local police.)

The demarchic groups would not be able to implement their decisions by force. Their power would rest on the persuasiveness of their decisions. One reason they would be persuasive is the process of random selection itself. This is the same reason why, in a court of law, a randomly selected jury is perceived as less corruptible and more representative of community values than a judge, and hence gives greater legitimacy to a decision.

Also, the decision-making process would have to be an open one. Submissions would be invited, presentations could be made and demonstrations of community feeling mounted. All hearings of the groups would be open, and a full record of deliberations made.

If, in spite of this, enough people didn't like a decision, they would have several ways of taking action. They could write letters, hold public meetings and form groups to promote their own views.

Actually, this is not so very different from present society. Most government decisions are successful only when most people obey voluntarily. When there is mass opposition, laws are left unenforced or repealed, as in the case of prohibition.

Unpopular policies would be unlikely to persist. The composition of any group would change over a period of a year or so. (For example, half of a group might be replaced each year, thereby providing both continuity and turnover.) A different group could easily come up with a different decision.

On intensely debated issues, such as abortion, there would be a strong incentive for partisans to mobilize as many supporters as possible to stand for the relevant group. But these supporters would have to be truly informed, since otherwise they might change their minds on hearing evidence and participating in group deliberations. Demarchy would encourage community education in the best sense.

The community would also have the benefit of thousands of former group members. These people, with their experience, would take a special interest in ongoing deliberations. Also, because most group members would expect to live in their local community afterwards, there would be a strong disincentive against appearing to gain any special personal benefit from the decisions made.


Unanswered Questions

Demarchy at the moment is a general concept. There are many issues that remain to be analyzed. What about implementation of decisions, such as a decision to set up a rail system? In Burnheim's model, there are no government bureaucracies. So the demarchic group would implement decisions themselves. How exactly would this happen? Would anyone have responsibility?

There is also the problem of coordination. The numerous local decision-making groups would need to coordinate their activities, not to mention coordination with groups in other localities. This could be done on a network basis, without any central supervisory executive. But what exactly would the network look like, and how would it work?

What would be the economic system be like? Burnheim thinks that demarchy is compatible with either small-scale private enterprise or more collective forms such as worker cooperatives. It is even possible to extend the demarchic process to the economic sphere, for example by having randomly selected groups to make decisions about land, money and labour. How would this operate?

These sorts of questions are well worth exploring, but it may be more fruitful to ask a more practical question: how can demarchy be achieved? When Burnheim wrote his book, he had little hope for this. He saw all the trends, on both sides of the mainstream political spectrum, moving toward greater centralization and bureaucracy.

The obvious way to move in the direction of demarchy is to try out decision making by randomly selected groups on a small scale. It so happens that this is exactly what the Jefferson Center in Minneapolis has been doing with policy juries and what the researchers at the University of Wuppertal in Germany have been doing with planning cells. In these experimental projects, randomly selected citizens deal with controversial social issues such as energy policy or water pollution from farm run-off. The groups listen to testimony from advocates of different positions, discuss the issues themselves and decide on their preferred course of action. It is found that participants take the exercises very seriously, draw "sensible" conclusions (who is to say whether they are right or wrong?) and become firmly committed to the process of participation. These experiences are the best available evidence that people selected randomly from the community have the interest, capacity and responsibility to make decisions on important issues.

When Burnheim wrote his book, he actually knew nothing about this practical work which meshes so nicely with his ideas.. Perhaps because he was writing at an abstract level, as a political philosopher, he was able to develop both a critique and an alternative that challenge fundamental assumptions about present political decision making.

Another avenue for promoting demarchy is through by changing decision-making methods at workplaces. Fred Emery, who has studied and promoted workplace democracy for decades, makes a good argument for this approach.

It is not difficult to guess that most politicians will be intensely opposed to demarchy. Indeed, many people will have a hard time understanding how it would work. As Ginsberg points out, we have been taught from an early age that elections are democracy and that this system is the only fair and efficient way to run a country.

Furthermore, even the leaders of "alternative" political groupings, such as green parties, may be unreceptive. After all, people who are leaders of political movements, including alternative ones, are not guaranteed a position of official influence by random selection. In my opinion, that may be the best argument for it.



John Burnheim, Is Democracy Possible? The Alternative to Electoral Politics (London: Polity Press, 1985).

F. E. Emery, Toward Real Democracy and Toward Real Democracy: Further Problems (Toronto: Ontario Ministry of Labour, 1989).

Benjamin Ginsberg, The Consequences of Consent: Elections, Citizen Control and Popular Acquiescence (Reading, MA: AddisonWesley, 1982).