Mancur Olson, in “The Logic of Collective Action,” proved that small groups with concentrated interests always win the coordination game against a diffuse majority due to the free-rider problem. Three CEOs will agree over dinner; a million customers never will, because each rationally waits for others to take the risk first.
What specific mechanism in ancap prevents a cartel agreement between two or three of the largest protection agencies to divide territories?
Konweni
This is a question that also concerned David Friedman. In the chapter “The Mechanics of Freedom,” concerning economies of scale, he reasons that in the law enforcement market, there seems to be a more pronounced positive economy of scale than he initially thought, which in turn increases the likelihood of a cartel agreement.
It seems to me that the cause of this excessive pessimism is the overestimation of the importance of direct armed protection in the security market. Literary descriptions of the security market under ancap are filled with some kind of universal protection agencies that are responsible for security in general across the most diverse aspects and scale easily. Of course, such an initial monopoly, already given in the problem statement, leads an honest analyst to conclude that such agencies will tend to grow significantly and enter into cartels, and their impact on people will be completely total, making the sale of such security services terribly dangerous for clients, as it is only one step away from total enslavement.
But where will these protection agencies come from? Who will be their clients, and what services will they buy from them?
Suppose I don’t live in peaceful Montenegro, where every house has a legal shotgun plus a couple of black guns left over from the war, but, say, in some place like Brazil. I have reasons to fear street attacks, car theft, and break-ins at my home. And then, a startup opens in the city that is ready to serve me, even if I wander into the favelas. What does such a startup need, within the technologies relevant at the time of writing this post? To reach an agreement with the company that owns the lamp posts to place cameras on them, as well as landing pads for drones with a charging function. Then, AI tracks suspicious actions, such as breaking into a house door or a car; a drone takes off from the nearest pole and demands a justification for the legality of the actions. In the case of an aggressive response, it applies measures of influence to the suspect, for example, firing a taser, after which it calls a live squad to process the detention. Similarly, in the case of an attack on a person: it flies in, demands they stop, reasons with those who ignore the instruction, waits for the squad, and returns to the nest. And then comes the court proceeding regarding how much to recover from the aggressor.
After working in this market for six months, the startup is surprised to find that its services are becoming less in demand because the city has become noticeably safer. And what, should they now start harassing me clandestinely so that I renew my subscription? No, of course not; they will simply make the subscription cheaper, and the drones that spend most of their time idle will be used, for example, for delivering pizza.
And what about peaceful Montenegro—do I not need protection agency services there at all? Of course I do. I need it so that when I break my leg while hiking in the mountains, brave rescuers in a helicopter fly to me and pull me out of that hole. Or pull me out of a car that flew off a cliff. Or find me in the sea when a storm carries me far from shore on a SUP board. Or get a cat down from a tree.
You realize that a Brazilian security startup won’t be able to enter the Montenegrin market because that market is completely different? Protection agencies are doomed to locality due to the local nature of threats, and are also doomed to a small size because the nature of threats also tends to change. It is harder for a large company to change, which means inefficiency, and inefficiency means saying goodbye to market share.
But surely there are functions for protection agencies that would be in demand globally? Of course there are, but these are specialized functions. For example, to find someone anywhere on planet Earth who owes me a vast amount of money and refuses to return it. Or a person who killed my mother and knows that I would like to burn him on her funeral pyre because she will need a servant there. Such agencies will not sell a subscription for service; they will have to take on rare one-off orders, as this market is too specific. It is impossible to build global dominance with the prospect of bending anyone to your will and subjecting them to your sinister market power based on this.
And what about the prospect of an external armed invasion? Shouldn’t private military companies be the backbone of any respectable social order if it wishes to call itself ancap? Shouldn’t such companies unite into a mega-cartel dictating its will to all of humanity? To the great regret of the adepts of this branch of human development, in peacetime, the peaceful average citizen is not inclined to finance the military. He doesn’t need war; he doesn’t buy it. An invasion army can only be equipped if you establish taxation on your territory. However, a militia for protection against an invasion army organizes itself in the presence of a free market—quickly, efficiently, and not too expensively, as shown by the experience of at least the Ukrainian war.








