Before answering the question in the post title, I will provide a background. By the way, I have made this point before. One problem with blogs is that old posts are buried over time, and it’s rare that someone stumble upon an old post and revive it. Maybe that’s not so much of a problem, because it allows the point to be repeated later, but after the idea has matured for a little while longer.
Suppose we could distinguish between two polar opposite forms of governance: extraction- and exchange-based. Let the variable g represent the total number of people whose ends are served through political institutions; let n represent the total number of people within that society. On one pole, g = 1 (and n > 1), so that the political process serves only that one person’s end, and on the other pole g/n = 1, which is what I call perfect pluralism. Extractive institutions, where g/n is considerably less than one, allow a minority to fulfill their ends at the expense of other, non-represented (n–g), members of society. Under exchange-based, or liberal, political institutions, a multiplicity of ends are served — in the same sense that markets serve a multiplicity of ends.
The reason I make the distinction between extraction- and exchange-based societies is to frame my point within a world where only exchange-based institutions exist. Not necessarily perfect pluralism, but somewhere along that spectrum where we can all agree that the government in question is “not so bad.”
Would large states exist? There are some reasons that they would, and they involve internal economies of scale. Internal economies exist when marginal cost is below average cost, so that average cost decreases as “the firm” produces more output. The common example is a firm that faces high fixed costs, so that the larger total output the more finely distributed those fixed costs are. In the study of market structures, internal economies are important, because their existence implies a force that favors larger over smaller firms. That is, all else equal, the market will be characterized by a relatively smaller number of larger firms (where there are no internal economies, we expect relatively more firms, which are smaller in size). The same is true if we swap “governments” for “firms.”
Before moving on to the important part of the post, I should comment on the empirical status of public goods and, more generally, “collective action” problems. Some people think roads are public goods, many others do not. Some people think there is some public good aspect to the provision of security, others do not. This post is not about which goods are public goods and which aren’t. I am talking on a more abstract level. I’m asking you to accept that there may be such a thing as a public good, and that if these goods exist then some of them differ in the requirements to provide them (including fixed and marginal costs). Any examples I use are for the purpose of illustrating the point alone. I acknowledge that, for example, it is debatable whether roads are really public goods, despite using them as examples. And, I hope that the reader generalize the points I make using specific examples.
There are some “collective action” problems that are cheaper to solve by larger governments. Take the provision of security. To simply, let’s assume that the only form of security is a standing army. Standing armies require large fixed costs, such as military installations. Inputs, such as equipment, are also often cheaper in bulk (the suppliers also benefit from internal economies). The probability of internal economies of scale is very high. Governments that provide security, then, are bound to be large, to exploit these economies of scale.
Another example — maybe one that is a little easier to swallow — is an upper order of a justice system. Suppose that there are many local courts, whether private or public, and these interact with each other (because, for instance, one association’s client attacked another association’s client). There will be an incentive to appeal to a higher order of rules, or a common standard between local courts, if you will. These orders might be like nested circles, where there are many of them — similar to what a modern court system looks like in a developed country. Establishing these levels require some fixed cost, but the marginal cost of extending the standard to new courts is very low. And, the benefits to an overarching standard increase as the number of people who abide by it rises: a network effect. These are related to internal economies, and the consequence is the same.
When exchange-based governments increase output, the implication is that they are appealing to a larger number of peoples’ ends. If g/n is approaching 1, then close to all of these people will be involved in the decision-making process. When there are internal economies of scale, what we should expect — in the context of exchange-based governments — are large governments that involve large quantities of people in the decision-making process. Using democracy, because that’s the most inclusive form of governance that we know of at this moment, when there are internal economies we should expect large democratic governments.
The truth is that there can be a broad range of goods provided by government. Some of these might have large internal economies, others may have no internal economies. A government that has a comparative advantage in producing one good may not have a comparative advantage in producing another. What this leads to are several governments, or several levels of governance, where each level specializes in producing a certain range of goods. This is exactly what modern representative governments are moving towards. Take the United States: we have local governments at several levels — city, county, and state, for example — and a Federal government. Each of these have their own range of services they provide, and they even have their own, sometimes different forms of funding. So, abstracting from the details, institutions of governance in a world where g/n is relatively high look remarkably similar to those which exist today in much of the developed world.