Mancur Olson

Power and Prosperity

In the 1920s, China was in large part under the control of various warlords. They were men who led armed bands with which they conquered a territory and then appointed themselves lords of the territories they had conquered They taxed their subjects heavily and used the proceeds to serve their own interests. The warlord who I was reading about, Feng Yu-hsiang, was noted for the exceptional extent to which he used his army for suppressing thievery and for his defeat of the relatively substantial army of a notorious roving bandit called White Wolf. Apparently, most people in Feng's domain wanted him to stay as warlord and greatly preferred him to the roving bandits.

At first, this situation was puzzling: Why should warlords who were simply stationary bandits continuously stealing from a given group of victims be preferred, by those victims, to roving bandits who soon departed? The warlords had no claim to legitimacy and their thefts were distinguished from those of roving bandits only because they took the form of relentless tax theft rather than occasional plunder.

There is a good reason for this preference. As we have seen, there is little production in an anarchy and thus not much to steal. If the leader of a roving bandit gang who finds only slim pickings is strong enough to take hold of a given territory and to keep other bandits out, he can monopolize crime in that area—he becomes a stationary bandit. The advantage of this monopoly over crime is not mainly thai he can take what others might have stolen: it is rather that it gives him an encompassing interest in the territory akin to that of the Mafia family considered in the previous section. He actually has a stronger encompassing interest than the Mafia family, since the bandit leader who takes over an anarchic area does not have competition from any government's tax collectors: he is the only one who is able to tax or steal in the domain in question.

This monopoly of theft changes incentives dramatically. We have seen that the individual criminal in a populous society has such a narrow or minuscule interest in the society that he rationally ignores the damage he does to it, which is obviously also true of a gang of bandits passing through. These socially perverse incentives make anarchies work badly. The encompassing interest of a stationary bandit leader who can continue to keep out not only other criminals but outside tax collectors as well gives him an incentive to behave very differently.

First, it leads him to reduce the percentage he takes from each victim of his theft. As we have seen, the criminal who is only one among many will take 100 percent of the money in any till he robs. By contrast, the stationary bandit with continuing control of an area wants to make sure that the victims have a motive to produce and to engage in mutually advantageous trade. The more income the victims of theft generate, the more there is to take. A secure stationary bandit, by making his theft a predictable tax that takes only a part of his victims' outputs, thereby leaves them with an incentive to generate income. If he cuts his rate of tax theft from 95 percent to 90 percent, he doubles his subjects' posttax reward for production and trade, which might well increase output and tax receipts by a large multiple.

The stationary bandit keeps on gaining from reducing his rate of tax theft down to the point where what he gains (from tax theft on a larger output) is just offset by what he loses (from taking a smaller share of that output). He is left at the revenue—maximizing rate of tax theft. If the stationary bandit cut his tax rate from 51 percent to 50 percent, thereby raising output of his domain from 98 percent to 100 percent, he would essentially maximize his tax collections: he would receive half of the increase in output, which would be approximately offset by the reduction in his share of total output. That is, the stationary bandit, because of his monopoly on crime and taxation, has an encompassing interest in his domain that makes him limit his predations because he bears a substantial share of the social losses resulting from these predations. If the stationary bandit in the example above increases his tax rate from 50 percent, he bears about half of the social or 'deadweight' loss from the distortion of incentives that this higher rate of predation brings about, which is enough to keep him from taking more.

Generally speaking, the greater the loss in production from taxation at any given rate of tax, the lower the rate of tax theft at which the stationary bandit's take is maximized. Though the deadweight losses and how they vary with tax rates—and thus a stationary bandit's rate of tax vary from case to case—every stationary bandit has a rate of tax theft that is always lower than 100 percent-and usually much lower—which maximizes his collections.