I have been interested in the creation of private cities for some time. A recent email exchange with Paul Romer (who I want to say was kind and clearly communicated with me which I very much appreciate as I have been critical of his work and he is a person whose time is very valuable) has led me to realize that I should write up a 15-minute pitch explaining my beliefs.

First, why care about laws at all? Romer’s TED talk is still probably the best introduction. Basically, rich countries are rich because they have good governments, poor countries are poor because they have predatory governments. A good government is one that allows private property, contract, exchange, rule of law, and organization. A bad government is one that prevents those things. With private property, contract, exchange, rule of law, and organization people learn to produce wealth, trade and have a higher standard of living. This is not just my opinion, it has become close to a consensus among economists who study the issue (see here for the best introduction, here, here, and here for econometrics, and here for a more advanced “why” analysis).

So basically, laws, or institutions, which can be thought of as the power structures which determine laws, determine the wealth or poverty of nations. So, why care about institutionally autonomous cities, institutional autonomy meaning having substantively different laws from the rest of the country? Basically, institutional change is very hard. British Prime Minister Gordon Brown has been known to quip that “the hardest part about building rule of law is the first 500 years.” In other words, getting the institutions right is very difficult.

The fall of the Soviet Union offers a useful example. A lot of well-intentioned economists tried to “fix” the former Soviet Republics. There are some success stories, Poland for example, but overall most of the countries failed to perform as anticipated. The economists were more focused on getting the prices right, than ensuring the courts were free and fair and the bureaucracy wasn’t corrupt. Another example is both Iraq and Afghanistan. The US poured hundreds of billions of dollars and thousands of lives into both, and the result is less evil governments than before, but hardly a substantive change in living standards for most of the population.

Institutional change is difficult because it first requires asking the institutions, to admit they do not function well, and second, to reform. These problems are only necessary because the institutions do not function in the first place. If a government was capable of doing so they would likely have already done it. In some sense, it is like asking coal to squeeze itself into a diamond.

Institutionally autonomous cities offer a way around this difficulty. By being built in rural land with no strong interest groups, they do not disrupt the status quo. Second, by being opt in, the acquire more legitimacy. They are not an outside body imposing new laws on an unwilling populace, but rather a new option for people trying to make a better life for themselves. Simply put, because they are regional instead of national, institutionally autonomous cities offer an easier path to institutional change.

So, I have focused my studies and energies on institutionally autonomous cities because I believe they offer the best option for institutional change, economic development, and improving the lives of the least well off humans.

Of course, there are different visions for the ideal institutionally autonomous cities. The visions can be distinguished by the governance body for the cities.

The best known advocate for institutionally autonomous cities is Paul Romer. His vision is Charter Cities. The governance body in a Charter City would be a first world country, Denmark for example. In that sense it would directly import first world institutions to third world countries. Instead of Indians or Zimbabweans needing to move to Denmark, Danish law and the opportunities and economic growth that comes with it could be brought to this country.

However, Charter Cities are not the only option. I think private institutionally autonomous cities, administered by Google for example, could produce better results more quickly than Charter Cities. However, before I discuss why I would like to say I pretty much entirely agree with the Charter Cities project. I think building a Charter City would likely be an enormous success. It would generate a great deal of wealth and improve the lives of people in the host country. In this sense, my quibbles are relatively minor. A Charter City would be a huge leap forward from the existing world, and a private autonomous city could be a small step beyond a Charter City.

So, why do I prefer a private city. The primary reason is adaptability. Building a new city is a megaproject. Private companies would be more responsive to changing conditions, new information, and unforeseen challenges.  However, before going into more depth, I would like to take a second to explain how a private autonomous city would function.

A private autonomous city would begin by buying a large area of relatively uninhabited land. They would also negotiate for institutional autonomy, basically a special economic zone. For this essay I will assume a private city state, for complete autonomy, but the arguments I make will also apply to a private city with more limited autonomy.

The business model would be similar to that of a mall. The proprietor of the land would pay for improvements, then rent out the space at a higher value because of those improvements. In this scenario, the improvements would be traditional public goods, roads, lights, water, sewage, electricity, as well as non-traditional public goods, police and courts.

The general argument is that the land proprietor would have a long term interest in the economic success of the city because such success would increase the value of the land and therefore the rents of the proprietor.

For this essay I will focus on the non-traditional public goods, police and courts, and conclude with my argument that private autonomous cities would be more adaptable. Traditional public goods are often supplied by private contractors to governments, suggesting it would be easy for a private city to contract with similar companies, if not do it themselves. Further, to the best of my knowledge, no one has argued against private cities on the grounds of higher utility costs.

The strongest argument against a private criminal and civil justice system is that it would favor the powerful. A private city, especially a small one, would likely have a small number of large employers. Several big companies would pay a disproportionate share of the city’s budget, and the city would be somewhat beholden to them. Combined with the more direct link between rent and profit this could lead to a criminal and civil justice system that favors the powerful. Especially because residents in a private city would not have the same sense of ownership as they might in a democratic city. Rather than citizens, they would be more akin to customers, which could embolden the city to favor the powerful and delegitimize the voice of the residents.

This argument applies to civil justice, police, and criminal justice. I will discuss them in that order. However, first I will make several general arguments which apply to all three.

First, all cities will have some companies which pay a disproportionate amount of taxes. The specific claim against private cities, is that private cities will be more likely to favor the powerful than other types of city governance. It is not obvious that private cities will necessarily be worse in this regard. Reputation will be very important to a private city hoping to convince residents and companies to locate there. A perception of injustice, either in the civil system, police, or criminal system, could hurt them in the long run.

As an example, many companies, Amazon, MasterCard, Geico, and Ebay for example, likely have a small number of sellers which account for a disproportionate amount of revenue. However, there companies are generally perceived as fair in their dispute resolutions. Few people are going to buy car insurance from a company that favors large clients, or use the credit cards of a company that has unfair dispute resolution mechanisms.

Second, such critiques of private autonomous cities often underestimate how bad civil justice, police, and criminal justice is in undeveloped countries. For a quick sense skim the World Justice Project’s report on the Rule of Law. For example, 21%, 23%, and 24% of Nigerians believe that police, follow the law, respect the basic rights of suspects, and are punished for breaking the law respectively. As a personal anecdote, several Honduran friends have told me they fear the police more than they do the gangs, even though Honduras is the murder capital of the world. More generally, it is common knowledge that justice systems in the undeveloped world rarely deliver justice. Unfortunately, that knowledge is often forgotten or ignored when considering the possibility of a private replacement.

There is strong evidence that private civil justice mechanisms can perform as well, if not better than public civil justice mechanisms. The best book on this is “Private Governance”, by Edward Stringham. One particularly telling example is the emergence of the Dutch stock exchange. Not only was short selling not enforced, it the Dutch government actually banned it. Nevertheless, people continued to write short contracts and they continued to be paid. Those who refused to pay contracts, even ones that were technically illegal, were simply banned from future participation in the stock market.

In a more modern example, international trade is often “lawless.” Companies frequently specify that contracts be settled under private adjudication for both speed and accuracy, as government courts sometimes fail to keep up with complex commercial contracts. The prevalence of international arbitration is reflected in detailed guides to drafting contracts, exemplified here, and here. Such contracts show private dispute resolution is often better equipped to handle complex contracts.

To some extent, international arbitration clauses can be said to exist in the shadow of the state. The New York Convention ensures participating governments enforce private arbitration agreements and rulings made in other countries. However, Peter Leeson found that state enforcement only accounted for 15-38% of international trade, a substantial amount, but not nearly enough to suggest private mechanisms were not working.

There is less evidence that private police can perform as well as public police. This is not to suggest there is suggesting private police would be worse than public police, merely that there is relatively little evidence regarding private police at all.

Edward Stringham has an article about private police in San Francisco. When people and businesses who hired the private police force were asked why they did not use the free public one, responses ranged from, “they take too long to arrive” to “they scare me”. The best response was “that’s a joke right? I have little confidence in SFDP”.

The best argument for private police is the general untrustworthiness of public police in the undeveloped world. I think most honest well-traveled people would admit they trust security guards at restaurants or hotels more than the local police. Unfortunately, improving on public police in the undeveloped world is not a very high bar.

The last, most controversial point, and the one with the least evidence to support privatization is criminal courts. This would be an extreme example, as it is highly unlikely that a private city would gain sovereignty. If a private city did manage to gain some autonomy, the criminal courts would likely be the last thing a host country would abdicate control over. While there is very little evidence on this point, I would like to make several overarching statement.

First, there are lots of examples of private actors acting badly in criminal justice, honor killings, feuds etc. However, I think these examples have as much bearing on a private city as the Saudi Arabia beheading teenagers has on arguments for social democracy. While they are in some ways the same, private criminal justice, and the actions of a state, they strip out the context that makes them different. Saudi Arabia is still a primitive state and honor killings are a primitive form of private justice. If Google were to build a city it is hard to believe it would try to build the honor killing, teenager beheading, primitive criminal justice system.

Second, a private city could remove sovereign immunity. Wrongful actions taken against citizens, including in the criminal justice sphere, would penalize the city.

Of course, that private cities could provide better civil justice, police, and criminal justice than undeveloped countries does not mean they would do a better job than a Charter City, merely that they would be better than the status quo. A Charter City, assuming it is implemented well, would provide civil justice, police, and criminal justice, at a level consistent with the Charter country. In other words, a Charter City would have low variance of its justice system.

A private city on the other hand, would have high variance. It could outperform a Charter City, but it could also greatly underperform a Charter City. Compared to a Charter City, a private city would have a slightly higher ceiling, but a much lower floor.

So, given the expected value of a private city based on civil justice, police, and criminal justice, is lower than a Charter City, why prefer a private city? The answer is adaptability, the ability to respond quickly to rapidly changing on the ground circumstances.

Cities, especially institutionally autonomous cities, are complex. The problems they have at 10,000 people are not the same problems they face at 1,000,000. Difficulties faced by the developer will not scale linearly. This requires a decision making body that is equipped to respond rapidly and effectively to such problems. Simply put, that body is not government.

One illustration of the difficulty of complex systems is massive multiplayer online games. Some of the more popular online games have millions of players trading with each other. While companies can test the games with thousands, even tens of thousands, of players before release, the dynamics will fundamentally change with millions of players. Often times the initial release is followed by several months of trying to rapidly respond to consumer feedback about the game. For comparison, take the difficulty state governments have had in establishing Obamacare exchanges online. Hawaii spent $205 million dollars developing a website far simpler than most big budget modern games, and it doesn’t even work.

Another example is the refugee crisis in Europe. Naguib Sawiris, an Egyptian Billionaire, is actively trying to buy a Greek island to house refugees. The EU, which has the resources, as well as the political clout to create a refugee city on a Greek island is standing on the sidelines.

For another thought experiment, think of a major institution of the US government that can respond quickly and effectively to new problems, the FDA, the Department of Homeland Security, the Department of Energy? The Department of Homeland Security is the newest cabinet level department created in the US. It’s most visible program, TSA, is widely considered a failure. It was recently revealed that they had a 95% failure rate in a test, failing to detect 67 out of 70 people trying to sneak fake bombs and guns on a plane. Is the political body that created the TSA really going to create an organization capable of building a Charter City, especially as the problem faced by the TSA is far simpler than problems faced by a Charter City. In this sense it is possible for a Charter City to end up a large boondoggle, a new Brasilia built but mismanaged to the point of vastly underperforming expectations.

In a Charter City there is the added problem that the decision making body would not necessarily be close to conditions on the ground. If Denmark, for example, is the Charter Country, would we expect all the Danish workers to move to sub-Saharan Africa? Requiring them to move would likely eliminate many married or elderly workers. If they are not required to move, they will be out of tune with the local conditions.

That being said, there are historical examples of public cities rapidly growing, Shenzhen being the most prominent example. However, these cities had different institutional arrangements than what is likely to arise in a Charter City.

To conclude I’d like to reiterate my main points. Institutionally autonomous cities are currently one of the best ways to improve the lives of the world’s poorest. The two competing visions of institutionally autonomous cities are Charter Cities and private cities. The primary advantage of Charter Cities is low variance and better provision of justice systems. The disadvantage is adaptability. The advantage of private cities is adaptability. The disadvantage is the low floor for the justice systems (though even a low floor would likely be an improvement in many undeveloped countries).

Ultimately, the success of a Charter City or a private city will depend on the organization with decision making authority. It is certainly possible, though unlikely, that business, growth oriented organization can emerge from negotiations between two governments to build a Charter City. Similarly, it is possible that a business with few ethical scruples can take advantage of a country granting institutional autonomy, preying on those who move there. The potential success of either project will depend the governing body, and those of us interested in such a project should do what we can to ensure the process for choosing the governing body is fair, open, and transparent.