On 23 June, the UK will vote on whether to withdraw from the European Union. This is a momentous decision, with far-reaching economic and political consequences that are now being hotly debated. If Britons vote to exit, it could lead to a break-up of the EU, and set back the global move towards greater political integration.

This raises an intriguing question: What’s the optimal size for economic performance? Are we better off with many little competing city-states, a bunch of mid-sized nations or just a few big super-countries overseeing hundreds of millions of people each? If bigger is better, what about a global government?

Actually, economists have thought about this a fair amount. In 1956, Charles Tiebout believed he had a solution to the problem. He reasoned that local governments knew more about their people’s needs than distant central governments, and so the best system was one where local governing units—city-states, essentially—offered different packages of taxes and public services. People would vote with their feet, going to the place that suited them the most.

This is a very appealing vision. Want to live in a libertarian paradise, where all the roads and schools are privately built? In a Tiebout world, there will be a city set up to accommodate you. Would you rather have lots of government-built trains and hospitals? That’ll be available as well.

A number of modern-day thinkers have suggested a return to Tiebout-style city-states as a solution to our problems of political gridlock and policy paralysis. In 2013, entrepreneur and venture capitalist Balaji Srinivasan playfully suggested that Silicon Valley could secede from the US, forming a technology-oriented utopia. A number of my Japanese friends have similarly suggested that Tokyo should secede from Japan—notably, not all of those favouring succession were Tokyo residents! Others have praised the economies of city-states such as Singapore, Dubai and Hong Kong.

Some people also claim that political fragmentation has been beneficial in the past. Anthropologist Jared Diamond, in his book Guns, Germs, and Steel, suggested that competition between small countries allowed Europe to get a head start on unified China in the Industrial Revolution. Economists Brad DeLong and Andrei Shleifer argued in 1993 that city-states helped Europe develop (though more recent evidence seems to counter this). Casual evidence would also suggest that Taiwan’s de facto independence from China helped provide the mainland with a capitalist model to revive its moribund economy in the 1980s and ’90s.

So, there are plenty of economic and historical arguments in favour of fragmentation. These imply that a Brexit would be good not just for the UK, but for all of Europe.

But there are arguments on the other side, too. Mathematician and economist Truman Bewley examined the Tiebout idea in the 1980s, and found that a patchwork of little city-states doesn’t always lead to a well-functioning system.

There are several reasons why Tiebout’s idea can fail. One is that many of the services governments provide are what economists call public goods. These are things that the private sector either can’t or won’t provide. The classic examples are national defence, police, courts and support for basic research. But many other things, like roads, electrical grids and ports, are usually in short supply when left to the private sector.

When you have public goods, one person’s economic welfare depends on how much everyone else in that city, state or country is willing to pay for things like roads, power grids and sewers. That throws a big wrench into the Tiebout model, because there are many different kinds of these goods, and the amount people want of each one tends to differ a lot. You and I might both agree that we need good roads, but you might care much more about broadband infrastructure or schools than I do. Bewley showed that even splitting the government into tons of local units doesn’t solve this thorny problem.

A second issue is that governments don’t always have the right incentives. Some governments may decide to maximize the size of their tax bases. Others might care only about the welfare of their citizens, while others might be beholden to special interests—I imagine an independent San Francisco would be ruled by local landlords even more than it already is. There’s no perfect type of local government, and so we’ll have a wide variety of them. Bewley showed that this problem also prevents Tiebout’s patchwork from being an economically efficient utopia.

There are plenty of other problems that Bewley didn’t think about. It might be very hard to coordinate between city-states—for example, one little local government, concerned about preserving open space, might be able to veto a cross-continental highway that would boost almost everyone’s income. Also, there’s the possibility that some city-states might just decide to conquer their neighbours, returning us to a world of empires. In fact, this is exactly what happened in Europe, China and elsewhere every time those regions fragment. And the only way to defend yourself against a neighbouring empire, many cities have found, is to band together into a nation. Since peace is good for growth, security considerations like this are also economically important.

So, the answer to the question of fragmentation is that there’s no ideal size. Sometimes city-states are the best, but sometimes super-states are better. The UK might be on the verge of an interesting experiment. If it abandons the EU, only history will tell us whether it was a good idea. Bloomberg

Comments are welcome at theirview@livemint.com