Analytical Anarchism

I’ve been remiss in not plugging this excellent collection of writing on the positive analysis of anarchism created by Michael Wiebe. There’s a list of published papers and books, and some working papers (including one by Eric and I).

Here’s how Michael introduces the site:

The purpose of Analytical Anarchism is to create an open forum for the academic community to promote and discuss research in analytical anarchism.

What is analytical anarchism? As the subtitle says, it is the positive political economy of anarchism, or simply, anarchism from the economic point of view. Anarchism here simply means the absence of government. Peter Boettke divides anarchist thought into three categories:

1. Utopian — following in the tradition of William Godwin’s An Enquiry Concerning Political Justice (1793).
2. Revolutionary — following in the tradition of Mikhail Bakunin and the First International, 1864-76.
3. Analytical — in the tradition of Murray Rothbard’s For a New Liberty (1973) and David Friedman’s The Machinery of Freedom (1973).

The analytical anarchism research program has developed out of this last tradition, and is currently being pursued by economists such as Pete LeesonEd Stringham, and Chris Coyne.

Why anarchy? Research in anarchism has a fundamental theoretical importance for understanding the mystery of cooperation among strangers, which forms the basis of modern social order. Understanding anarchy also has a critical practical importance for transition economies, Third World development, and post-war reconstruction. Economic analysis of these problems cannot assume a functioning state.

For an introduction to the subject, see Boettke’s “Anarchism as a Progressive Research Program in Political Economy.”

Scholars and students working in this field are invited to submit working papers and posts discussing the literature, general issues, potential research topics, etc.

Quote of the Day: Mises Edition

Another from the “I want to make a note of this for future reference and a blog post seems like the easiest way to do it” files. From Theory and History, chapter 7:

In the world of reality, life, and human action there is no such thing as interests independent of ideas, preceding them temporally and logically. What a man considers his interest is the result of his ideas.

If there is any sense in the proposition that the interests of the proletarians would be best served by socialism, it is this: the ends which the individual proletarians are aiming at will be best achieved by socialism. Such a proposition requires proof. It is vain to substitute for such a proof the recourse to an arbitrarily contrived system of philosophy of history.

All this could never occur to Marx because he was engrossed by the idea that human interests are uniquely and entirely determined by the biological nature of the human body. Man, as he saw it, is exclusively interested in the procurement of the largest quantity of tangible goods. There is no qualitative, only a quantitative, problem in the supply of goods and services. Wants do not depend on ideas but solely on physiological conditions. Blinded by this preconception, Marx ignored the fact that one of the problems of production is to decide what kind of goods are to be produced.

With animals and with primitive men on the verge of starvation it is certainly true that nothing counts but the quantity of edible things they can secure. There is no need to point out that conditions are entirely different for men, even for those in the earliest stages of civilization. Civilized man is faced with the problem of choosing among the satisfactions of various needs and among various modes of satisfying the same need. His interests are diversified and are determined by the ideas that influence his choosing. One does not serve the interests of a man who wants a new coat by giving him a pair of shoes or those of a man who wants to hear a Beethoven symphony by giving him admission to a boxing match. It is ideas that are responsible for the fact that the interests of people are disparate.

Incidentally it may be mentioned that this misconstruing of human wants and interests prevented Marx and other socialists from comprehending the distinction between freedom and slavery, between the condition of a man who himself decides how to spend his income and that of a man whom a paternal authority supplies with those things which, as the authority thinks, he needs. In the market economy the consumers choose and thereby determine the quantity and the quality of the goods produced. Under socialism the authority takes care of these matters. In the eyes of Marx and the Marxians there is no substantial difference between these two methods of want satisfaction; it is of no consequence who chooses, the “paltry” individual for himself or the authority for all its subjects. They fail to realize that the authority does not give its wards what they want to get but what, according to the opinion, of the authority; they ought to get. If a man who wants to get the Bible gets the Koran instead, he is no longer free.

That’s about the best summation of the subjectivist position I’ve come across.

Power Corrupts Rational Thought

Wray Herbert reports research showing that a sense of power gives people the illusion of control over random events:

One recent study may offer some insight into the connection between power and hubris and delusional thinking. Stanford University psychologist Nathanael Fast actually started off exploring the positive effects of power. His idea was that power creates a false sense of control over life’s events, and that this feeling of control in turn boosts self-esteem and optimism. But his findings apply just as well to prideful overconfidence. Here’s the experiment:

Fast and his colleagues used a well-tested laboratory technique to prime some volunteers’ sense of power. Then they used a clever test to see if these feelings of power influenced their sense of control over random events. They had all the volunteers play a dice game to see if they could win a prize, and they were allowed to either roll the dice themselves or to let someone else roll the dice. A roll of dice is random, no matter who rolls them, so those who chose to roll the dice were displaying an unrealistic sense of control over random events.

The results were unambiguous. As reported in the journal Psychological Science, each and every one of the volunteers who was primed for power (compared to controls who were not) grabbed the dice. They had the delusional belief that, by rolling the dice themselves, they could control the outcome.

Herbert suggests this as an explanation of political corruption: due to a higher sense of control, people in positions of power see themselves as more able to evade detection and punishment. While that’s probably true to a certain degree, I suspect the the impact of illusions of control are much greater on well-intentioned policy.  A false sense of control will turn even the most altruistic rulers (though power does corrupt morally, as well) into men of system apt see society in easily-influenced, mechanistic terms.

Evolution, Entertainment, and the Socialist Calculation Debate

Eric Crampton is justly worried that the God-Game genre, typified by SimCity and Civilization, fills people with technocratic hubris, giving rise to the fatal conceit that a benevolent planner can actively improve the lives of citizens through direct intervention. He quotes a review of the new game Dawn of Discovery which complains that the complexity of running an entire society doesn’t make for interesting gameplay:

Naturally, as your population grows, so too does the amount of each good that the population consumes, but there’s no clear way to determine just how many tons of a particular good your residents require. This makes it needlessly difficult to anticipate upcoming shortages, and it’s easy to get frustrated when you find yourself in the midst of a dairy crisis or similar shortage that could have been avoided with clearer information regarding supply and demand.

As Eric points out, any planner in the real world would face even more serious problems. The sheer complexity of social and economic life makes large-scale planning impossible:

If you as central planner don’t build things like airports, ports, libraries, universities, temples or a colosseum, they just don’t get built. If your workers don’t build farms and mines, no entrepreneur steps in to do it. In SimCity, or at least the version I played more than a decade ago now, you have to specify rigid zoning and can’t just let the city evolve. Unfortunately, any realistic game that requires the central planner to make all of these decisions will require that we encounter the calculation problem; it’s neat to see the game reviewer complaining about it. Of course, the gaming would be a bit more boring for the player if he could just set some basic laws, a low tax rate, and try to stay on good terms with the other civilizations out there: the game is designed to maximize fun for the player, not to maximize utility for the simulated persons within the game. The more that games disguise the inefficiencies caused by the “economic planning” approach, the less will today’s players appreciate Hayek.

I think that’s dead right, and also applies to any form of entertainment with a macro-social setting. We enjoy playing games with relatively simple cause-and-effect dynamics because our minds evolved to deal with simple cause-and-effect situations. Most of the challenges we faced in our ancestral environment, and continue to face in our everyday lives today, involve overcoming a particular problem without much need to be concerned about unintended consequences.

A society at large, though, doesn’t work this way, but is instead a complex system in which results emerge from the interaction of many individuals pursuing their disparate goals. Any attempt to improve outcomes through top-down intervention is just as likely to make matters worse. There is conflict between our folk economics and reality and, when we want to be entertained, folk economics always wins: nobody wants to play a game in which the only winning move is not to play, or read a book with no protagonist.

The availability heuristic means that consuming stories of active characters changing the macrostructure of the world singlehandedly through sheer grit and determination is likely to bias our perceptions of reality. The ease of imagining some situation, for example, seems to affect its perceived likelihood. Since entertainment does our imagining for us, it will likely bias our models of the world.

While many fictional biases are likely to affect our everyday lives (by making us afraid of terrorists, for example), technocratic hubris and the “Great Man” theory of history probably don’t matter most of the time. Where they could plausibly have an effect is in the political realm, biasing both voter preferences towards statism and giving policymakers too much faith in their own technocratic abilities. Unfortunately, politics provides no incentives for the correction of our biases.

Customer-Owned Protection Agencies

I suggested yesterday that protection agencies which credibly commit to not joining any nascent cartel are likely to attract more customers than those which don’t, potentially nullifying Cowen and Sutter’s critique of market anarchism. One obvious possibility is customer ownership of protection agencies. Cowen makes this suggestion in the final paragraph of his 1992 paper:

In the above scenarios, the network becomes a government because network shareholders are able to exploit successfully conflicts between network profit maximization and the interests of network consumers. If consumers are sufficiently far-sighted, they may prefer dealing with agencies that precommit to never becoming collusive or coercive. Consumers may attempt to control the network by owning the member firms; under this scenario, the protection agencies would become mutuals or cooperatives. Protection agencies could then be bound by democratic procedures, according to consumer vote. Collusion could not occur unless approved by agency customers (shareholders).

He expands a little in a footnote:

In mutuals, the corporation’s customers are also its owners. A mutual life insurance company, for instance, is owned by its policyholders, who serve as residual claimants. If the company makes money, the profits are refunded in the form of lower premiums; conversely, losses imply higher premiums. (Not all of the mutual’s profits are rebated to customers, however, as managers retain perks for themselves.) In so far as mutual shareholders succeed in controlling their company, their dual roles as owners and customers diminish conflicts of interest. Policies that deliberately defraud customers, for instance, would not be approved by mutual shareholders. Shareholders of traditional corporations, in contrast, will maximize profits at the expense of consumer interests, when possible. Cooperatives and nonprofit organizations are other possible organization forms for protection agencies. Although these forms differ from mutuals with respect to many details, they also eschew direct profit-maximization and allow managers to maximize the flow of perks, although subject to different institutional constraints.

Customer ownership of protection agencies is probably the simplest and most effective way of avoiding a despotic cartel emerging from libertarian anarchy. The problem with such arrangements, though, is that they introduce many of the same problems which currently plague democratic politics.

Customer-shareholders need some way of making sure management acts in their best interests. The most obvious way of doing this, as Cowen suggests, is to have shareholders periodically vote for the CEO, or directly vote whenever particularly important decisions arise. As in any firm, this won’t entirely prevent managers from exploiting their position, but it will place fairly tight limits on the extent of corruption.

Whenever a moderately large group of people vote to decide some course of action, no individual is faced with a genuine choice of which path to take. They can express their preference, but, except in the case of an otherwise tied election, the outcome will be unrelated to their choice. This means that nobody has an incentive to think carefully about their decision, and have every incentive to vote expressively and indulge their irrational biases. This has been well-documented with respect to ordinary democratic politics, but is relevant to any large-group voting situation.

This doesn’t seem to be particularly important in ordinary shareholder voting (though as far as I know, nobody has looked into it and I can imagine it having some effect), presumably because voting rights are allocating by the share rather than the person – giving those with the most at stake the most say – and because the activities of corporations don’t tap into expressive preferences or cognitive biases to the same extent as democratic politics.

Unfortunately, collective choice within customer-owned protection agencies more closely resembles political than shareholder voting in this respect. While customers with more expensive premiums may be given extra votes, the inequality of voting power will be nowhere near that of an ordinary corporation. Further, law enforcement and the definition of rights seem like areas in which expressive preferences are likely to dominate.

Expressive shareholders will not only make protection agencies run inefficiently, they will also be more likely to violate the rights of others and engage in destructive conflict. People are more bigoted and bloodthirsty when freed of cost considerations. These are problems we live with under democratic rule today, however, and it’s hard to see why they would get worse under anarchy. While anarchy with customer-owned protection agencies will be far from perfect, it should be considerably better than centralized government.

Protection agencies will initially behave like lots of little governments, with all the inefficiency and bigotry we see in politics today. The crucial difference, though, is the option of exit. A thousand nations will bloom and efficient protection agencies – those managing to minimize the harm of foolish voters and corrupt managers – will gain more customers than inefficient ones. This could result in many small agencies which give each customer a significant voice, or agencies with supermajority rules and other limits on strict majoritarianism. Of course, the potential for innovation will be lower than in a market with entrepreneurs making profit-seeking decisions. People will flock to efficient agencies, but agencies’ decision rules will be unresponsive to consumer demand.

There may be ways for an ordinary shareholder firm to credibly commit to avoid a cartel, and the market would provide every incentive entrepreneurial discovery. I can’t think of any entirely plausible way, but that doesn’t mean one doesn’t exist. We might see a customer-ownership equilibrium eventually give way to a shareholder equilibrium once commitment mechanisms are devised.

I do have some niggling concerns over collective action problems (one shareholder-only firm would be more efficient than its competitors, would have no peers with which to form a cartel, and would therefore be attractive to customers when all other firms are customer-owned), but it seems that customer-owned anarchy would be preferable to the status quo, and would improve over time.

This is why I am now a tentative anarchist.

Protection Agency Cartels and Organizational Innovation

Every liberal worthy of the label is a philosophical anarchist. The presumption of liberty – i.e. the notion that everything is permissible unless shown otherwise, with the burden of proof resting with those who would restrict liberty – means that we should only resort to politics after all peaceful means of solving social conflict have been exhausted. Force will always be required to stop bad people raping and killing, but I think it’s an open question whether force against the innocent (taxation) is required to fund protection against the guilty.

Market anarchists such as the Tannehills and David Friedman have argued that it is not. They imagine private protection agencies and courts voluntarily competing for customers on the open market. The current monopoly on violence would be replaced by many firms filling diverse market niches and being forced to respond to customer needs. As in any other market, they say, competition will do a far better job of satisfying customer preferences than a coercive monopoly.

If these arguments are correct, it’s pretty hard to see how anyone valuing individual preferences above all else could support government. Anarchy won’t be utopia, but it will make people far freer than they are under government.

Many anarchists will claim that since markets work better than government for every good they’ve been tried, we must conclude that they’ll also work well for policing and court services. While there’s definitely some merit in this argument, I think we need to be very careful when comparing the market for organized violence with other markets.

Tyler Cowen and Dan Sutter have had me convinced for some time that orderly anarchy is unstable and would result in something very similar to a highly predatory government. To avoid perpetual war, competing protection agencies will need some way of resolving disputes when customers of different agencies come into conflict.

Suppose Alice is a customer of Agency A and Bob is a customer of Agency B. Since agencies compete partially on the basis of offering different rules tailored to customer preferences, some things viewed as a crime by Agency A will be deemed permissible by Agency B. What happens when Alice demands compensation for some action Bob has performed which she has paid Agency A to protect her against? In a world without interagency arbitration, Agency A will punish Bob. Bob will demand that Agency B protect him from this, and the two agencies will go to war. We end up with Hobbesian anarchy.

Of course, this would be very expensive for everyone concerned, and there is a strong incentive to provide for peaceful means of dispute resolution. This is where private arbitration firms come in. To avoid conflict, protection agencies will agree to be bound by the decisions of independent courts whenever conflicts arise.

Arbitration is a natural monopoly, since it’s very important to customers of protection agencies that they be able to resolve disputes with customers of other agencies. This requires considerable cooperation between competing protection agencies, including the ability to exclude or otherwise punish agencies not playing by the rules. The capabilities required of competing firms to cooperate (“cooperative efficacy”) to avoid Hobbesian anarchy – communication, ability to punish defectors, etc – are the very same things required to form a profit-maximizing cartel capable of excluding market entrants and ignoring customer preferences. From the customer’s point of view, it would be hard to distinguish this from an extremely predatory state.

So: if protection agencies can’t cooperate, we end up with chaos. If they can, we end up with a de facto state. Orderly libertarian anarchy is not a stable equilibrium.

Bryan Caplan and Ed Stringham point out that refusing to deal with harmful outlaw agencies (e.g. those allowing murder, or always willing to defend their customer with violence without concern to their guilt or innocence) is a self-enforcing constraint. The arbitration network can declare such agencies rogue, and it is in every other agency’s interest not to deal with them. It will remain profitable, however, to continue dealing with agencies which behave decently but refuse to abide by the rules of the cartel.

It therefore requires more cooperative efficacy to enforce a cartel arrangement than it does to exclude genuinely rogue agencies: there is a zone of stable libertarian anarchy between Hobbesian anarchy and despotic government. Cowen and Sutter admit that such a zone exists, but insist that it is a narrow one. We have no reason to think that this is where we’ll end up and should not wish to experiment with anarchy.

As far as the argument has gone so far, I basically agree with Cowen and Sutter. I don’t think the debate has gone far enough, however. Stringham and Hummel attempt to reply to Cowen and Sutter by arguing that ideology can change the payoffs facing protection agencies: positive material payoffs would be outweighed by negative psychic payoffs of cartelization if only those running protection agencies would internalize the libertarian worldview. That may be so, but I’d feel very uncomfortable if the only barrier between me and a despotic cartel was ideological.

The place the debate needs to go, in my view, is looking at forms of organization other than a simple subscription service protection agencies might take to pre-commit to not cartelizing. Cowen mentions this at the end of his 1992 paper “Law as a Public Good,” (not on the web, but available in Stringham’s excellent volume Anarchy and the Law) but doesn’t go into any detail and the idea has not been pursued by anyone else.

Cartelization is a market failure, and like any market failure provides incentives for its own solution (see Coase on the lighthouse, for example). A firm which can guarantee its customers that it won’t enslave them in the future will receive more custom than one which cannot make such a guarantee. This would generally require forward-looking consumers, but there don’t seem to be any collective action problems to overcome. If we ever get to the situation where the state has withered away, I would bet on people being reasonably far-sighted in their choice of protection agency.

Of course, just because there’s an incentive to provide something doesn’t mean it can actually be provided. Anyone inventing a perpetual motion machine would become very rich. Unfortunately, such a machine seems to be a physical impossibility. I don’t see any reason, however, to suppose that credible commitment to non-cartelization is impossible, and there are some plausible candidates for how it might be achieved.

This post is getting long, so I’ll leave discussion of the particular ways firms might pre-commit for another day (Update: see here). By the way, this detailed overview of public choice theory and anarchism by Powell and Stringham is great stuff for anyone wanting a broad understanding of the literature.

Anarchist Prisons

David Skarbek’s recent paper Self-Governance in San Pedro Prison provides evidence for the possibility of orderly market anarchism and, when combined with past research, against that of orderly non-market anarchism. The abstract:

The inmate-governed community in the Andersonville Civil War prison camp resulted in a state of violence and disorder. Past research argues that self-governance in prisons results in a dominant group comprised of the most violent inmates preying upon other members of the community. This paper examines the inmate-governed San Pedro Prison in Bolivia, and it argues that order within a situation of prison anarchy is possible when inmates can engage in economic exchange and have access to well established markets that they expect to persist.

Looks very interesting. My to-read list is getting out of control.

Holcombe on the Foundations of Welfare Economics

I haven’t yet read it, but this paper by Randall Holcombe in the latest Review of Austrian Economics, “A reformulation of the foundations of welfare economics looks like a very important contribution. The abstract:

Neoclassical welfare economics takes an outcome-oriented approach that uses Pareto optimality as its benchmark for welfare maximization. When one looks at the remarkable improvements in economic welfare that have characterized market economies, most of those improvements in welfare have been due to economic progress that has introduced new and improved goods and services into the economy, and innovations in production methods that have brought costs down, leading to higher real incomes. Pareto optimality is only peripherally related to actual economic welfare, and no economist would argue that people are materially better off today than a century ago because the economy is closer to Pareto optimality. After analyzing the actual factors that lead to improvements in welfare, this paper suggests a reformulation of the foundations of welfare economics to replace the almost irrelevant outcome-oriented concept of Pareto optimality as the benchmark for evaluating welfare with a process-oriented benchmark based on factors that generate economic progress. The paper then explores some implications of this reformulation.

Nerdy History of Economic Thought/Metal Quote of the Day

Eric Crampton:

Is it strange that I always hear the Sepultura lyrics as “War of Methodology” rather than “War for Territory”?

Yes, Eric, but in a good way. (Read about the methodenstreit here)

Claudia Williamson on Informal Institutions

I haven’t yet read this paper from the latest Public Choice, but it looks very good. Williamson, C. 2009. ‘Informal institutions rule: institutional arrangements and economic performance.’ Public Choice 139(3), pp.  371-387.


Institutions are widely believed to be important for economic development. This paper attempts to contribute to our understanding of how institutions matter by examining the effect of formal and informal institutional arrangements on economic progress. Formal institutions represent government defined and enforced constraints while informal institutions capture private constraints. The findings suggest that the presence of informal institutions is a strong determinant of development. In contrast, formal institutions are only successful when embedded in informal constraints, and codifying informal rules can lead to negative unintended consequences. This suggests that institutions cannot be easily transplanted in order to spur economic development.

Hat tip: Division of Labour. Claudia blogs at The Perfect Substitute.