Julian Sanchez header image 2

photos by Lara Shipley

Understanding Nozick Perfectly Well, Thank You

February 20th, 2008 · 5 Comments

Randy Barnett writes at Cato Unbound:

In his reply to my post, Tony repeats a common mischaracterization of Nozick’s argument. In Anarchy, State and Utopia, Nozick made no prediction about the emergence of a dominant protection agency. Nor did he describe a mechanism for that emergence. Rather, his argument was that a dominant protection agency could emerge morally without violating anyone’s rights. This moral argument was countered by several contributors to the first issue of the Journal of Libertarian Studies, which you can access here (scroll down to the bottom).

This is just plain wrong. While Nozick does rapidly move between positive and normative arguments in the relevant portion of Anarchy, State, and Utopia, he is quite clear when he is, indeed, making predictions about the emergence of a monopolistic dominant agency, and he does describe a series of mechanisms that would tend to produce that result. For anyone who feels like reading along, that argument can be found in Chapter 2 under the heading “The Dominant Protection Association” (pp 15-18), in Chapter 5 under the heading “The De Facto Monopoly” (pp 108-109) and in Chapter 6 under the heading “Stopping the Process?” (pp 120-125), where the big clue that he is making a predictive rather than a moral argument comes in the form of a series of huge honking game-theoretic matrices intended to illustrate why rational clients of protection agencies would seek to prevent others from joining competing agencies.

For anyone too lazy to click through to the linked pages, here’s a representative quotation:

Why is this market [for coercive protection and rights enforcement] different from all other markets? Why would a virtual monopoly arise in this market without the government intervention that elsewhere creates and maintains it? The worth of the product purchased, protection against others, is relative: it depends upon how strong others are. Yet unlike other goods that are comparatively evaluated, maximal competing protective services cannot coexist; the nature of the service brings different agencies not only into competition for customers’ patronage, but also into violent conflict with each other. Also, since the worth of the less than maximal product declines disproportionately with the number who purchase the maximal product, customers will not stably settle for the lesser good, and competing companies are caught in a declining spiral.

While I always love to see a creative interpretation, I really do not think there is any way to parse this as being about what might permissibly emerge from a state of nature. It is an argument for why convergence on a single, monopolistic, state-like agency would occur. You may not agree with the argument offered here (I find it compelling enough), but it seems futile to pretend that it isn’t offered.

Update: Prof. Barnett has posted a reply. As he avers not to have any real dog in the fight and allows that my characterization may be right, I scruple to pick nits, but I’ll throw out a couple observations.

The three passages I link above really need to be taken together; they’re dispersed, but they do constitute an interlocking argument. In the first, Nozick argues that competition between protection agencies will yield either (1) victory for one over the others, (2) geographical division of the market, or (3) some unified “federal” system under which conflicts are resolved. That the first two amount to states is clear enough. But Barnett quotes a paper by Roy Childs arguing that the third scenario cannot be properly characterized as such, as it is merely a form of emergent order distinct from real state-created “federalism.”

At the risk of being excessively semantic here, though, this is not a disagreement about what Nozick is trying to argue; it is a disagreement about what Nozick’s argument really establishes. That is, Nozick manifestly wants to claim that all of the plausible outcomes of competition for coercive rights enforcement yield scenarios functionally equivalent to the monopoly state. Childs is arguing that no, one of the three possible scenarios is not really like a monopoly state. Bracket for the moment which of them is right: The whole premise of Childs’ argument here is that Nozick has sought to establish the inevitability of a monopoly state. That is why Childs is picking a fight.

Now, on the merits, my sympathies are with Nozick. Childs’ analogy to farming, wherein all food producers jointly constitute a “monopoly” in some purely verbal sense, strikes me as missing the point. And the process by which federalism emerges here does not seem nearly as important as the power dynamic that results: Once the protection agencies have arrived at a mechanism for detente or dispute resolution or whatever we care to call it, why should we not expect this same mechanism to serve to permit them to exercise monopoly power as a cartel, whether or not the agencies remain nominally distinct? It’s important to notice that the third scenario here does not contemplate genuinely open-ended competition, but only a kind of modus vivendi between agencies who are, in practice, precisely strong enough to compel their inclusion in the federal system, but not strong enough to dispense with it. And the argument in the third passage linked makes the case that the participants in such a system, once it is established, are unlikely to permit new entrants to accumulate the power base to force their inclusion. But we don’t actually need to wade into these details, since the fact that Childs takes himself to be disagreeing with Nozick proves my point, I think.

Leave that for the moment. Barnett correctly notes that the DPA as characterized by Nozick at this stage of the argument is not a full state monopoly. Barnett then argues that Nozick’s “move from the DPA to a monopoly minimal state involved a normative claim about the right of the DPA to protect its clients from the procedures of rivals that it deems to be too risky.” And Nozick does indeed make such a claim. But there are several balls in the air here. One concerns whether the kind of ban on “self-help” punishment Nozick contemplates is morally permissible. Whether we agree with Nozick on this point is, for present purposes, immaterial. Another, conceptually distinct argument, is offered in the third linked passage above, and this argument goes to the question of whether dominant agencies and their clients will, in practice, rationally attempt to prohibit self-help punishment and (crucially) the formation of new coalitions for the purpose of rights enforcement, given inevitable information asymmetry about both the intentions and the reliability of new agencies.

These points do get blurred together to the extent that Nozick supposes that agencies and their clients are, generally speaking, making a good-faith effort to adhere to moral constraints—the idea here is to grant the anarchist his best-case scenario. But it is important here that the relevant standard is not whether, in some sub specie aeternitatis sense, the firms are acting morally. It is whether they are acting rationally within ethical boundaries as they see them given their asymmetrical epistemic situations. Or, more briefly, Nozick is not quite saying that a monopoly state (one that prohibits competitors) will emerge even if everybody acts morally. He is making the somewhat weaker claim that this will (not can, but will) occur even if everyone is doing his best to act morally given the available information.

Now, again, it is surely open to debate whether Nozick’s argument here succeeds. I don’t think there is a great deal of wiggle room to be found as to what Nozick’s argument is.

Tags: Libertarian Theory



5 responses so far ↓

  • 1 The Whaler // Feb 21, 2008 at 8:41 am

    Julian, you’re being a tad misleading here. Nozick’s descriptive arguments work inside his model. (He’s describing how the model would work). But, in terms of argumentative strategy, the reason he’s using the model is for the very reason that Professor Barnett mentions–namely, to show that a monopolistic state *could* emerge from the state of nature without violating anyone’s rights.

    After all, (pg. 4), he opens the section by claiming to take the moral assumptions of anarchist theory seriously. His argument clearly is not this frankenstein (as you seem to be suggesting): anarchy is impossible, a monopolistic state *will always* emerge, therefore anarchy ought to be abandoned.

    If you’re doing the back work on Barnett’s point, consider pg 5: “If one could show that the state would be superior even to this most favored situation of anarchy, the best that can be realistically hoped for, or would arise by a process involving no morally impermissible steps, or would be an improvement if it arose, this would provide a rationale for the state’s existence; it would justify the state.”

    To distance himself even further from what you’re saying here, Nozick adds a footnote to that last sentence: “This contrasts with a theory that presents a state’s arising from a state of nature by a natural and inevitable process of deterioration, rather as a medical theory presents aging or dying. Such a theory would not ‘justify’ the state, though it might resign us to its existence.”

  • 2 David // Feb 21, 2008 at 8:48 am

    For what it’s worth:

    M. Rothbard takes a similar view (to Childs and Sanchez) of what Nozick is trying to argue.

    “The Nozick contention that a dominant agency would develop in each geographical area, then, is an example of an illegitimate a priori attempt to decide what the free market would do, and it is an attempt that flies in the face of concrete historical and institutional knowledge. Certainly a dominant protective agency could conceivably emerge in a particular geographical area, but it is not very likely. And, as Roy Childs points out in his critique of Nozick, even if it did, it would not likely be a ‘unified federal system.’ Childs also correctly points out that it is no more legitimate to lump all protective services together and call it a unified monopoly than it would be to lump all the food growers and producers on the market together and say that they have a collective ‘system’ or ‘monopoly’ of food production.”


  • 3 Brian N. // Feb 21, 2008 at 9:20 am

    Mr. Sanchez, I don’t know how familiar (if you’re familiar) with the Mises calculation thesis on socialism. Rothbard extended its basic point into the market itself, and pointed out that in a free market, large firms tend to be unstable, and the ‘one big cartel’ bugaboo is totally unstable; just as unstable as a socialist system. It’s one reason why I see Nozick’s thesis as being bankrupt. The other is that he confuses what monopoly is, and always has been, in the real world, and what it takes to build them and maintain them.

  • 4 Julian Sanchez // Feb 21, 2008 at 11:30 am


    I’m familiar with the calculation debate, but I think there’s a fairly straightforward two-word possibility proof here: States exist.


    I’m not denying at all that there’s a normative dimension to Nozick’s argument. My point is that he is not just making an argument that it is *possible* for a monopolistic state to emerge within this model where everyone makes a good faith effort to act morally. He’s presenting an argument for why, even under those circumstances, a monopoly state *will tend to arise*.

  • 5 Micha Ghertner // Feb 21, 2008 at 11:36 pm


    It’s true that states exist, but to my knowledge, no existing state arrived at statehood from anarchy through a process even remotely similar to what Nozick describes, which I think gets us back to Brian’s point about the calculation debate and the likelihood of stable cartels. The existence of existing states tells us very little about the likelihood of collusion among private defense agencies since no existing states arose that way.

    Your point about the unlikelihood of perfect competition in the third scenario as a result of increased barriers to entry is well taken. I’m reminded of that scene in The Godfather where all the various mafia dons are sitting around the table trying to come to some workable solution. I haven’t seen the movie in a while, but I don’t think the meeting ended well.

    It also seems quite similar to the role of the United Nations. Different countries get together and try to resolve their differences amicably. They don’t always succeed. And it’s certainly difficult to start a new country, let alone gain recognition for one in the U.N. But only the sort of conspiracy theorist who goes on and on about sovereignty truly believes that the U.N will inexorably lead to an all-encompassing world government, no?

    Then again, perhaps the difference between the U.N. and the above mentioned third scenario prediction turns on the issue of geographical monopoly. I’d have to reread my ASU to understand this difference; it’s been a while.

    On a related topic, have you read John Hasnas’s Reflections on the Minimal State, 2 Politics, Philosophy, and Economics 115 (2003)? Hasnas neatly skirts the present Nozickian controversy, arguing that even if the market failure critique of fully private law is correct, the minarchist solution of government provision of law doesn’t necessarily follow. It’s a neat argument; I think you’d like it.

Leave a Comment