Is Economics Testable?

A Reply to the Theory of Anything Podcast

Bruce Nielson and Peter Johansen of the Theory of Anything Podcast, which I highly recommend, recently released an episode discussing whether economic theory is testable. I was mentioned in connection with the view that economics is not testable. My reasoning has been that human behaviour is unpredictable and that government policy rarely permits controlled, rigorous experimentation. Bruce suggests instead that economics is roughly testable, a position to which I am sympathetic.

One argument in favour of testability is that there are clear historical examples of government mismanagement. Grocery stores in the Soviet Union were frequently empty, which is often taken to indicate that socialism failed to allocate resources efficiently. A counterargument is that this assumes collectivisation itself was responsible, rather than other factors such as planning failures, distribution problems, or political prioritisation of heavy industry and military spending. The empirical lesson is therefore less decisive than it might initially appear.

I remain unsure where this leaves me. I am inclined to think that the history of the Soviet Union poses serious challenges to claims about the efficiency of collectivisation. At the same time, my view is driven more by basic economic reasoning, such as incentives and supply and demand, than by historical data alone. “Roughly testable” may therefore be the right description, in that economic theory has empirical content, though I would prefer if this terminology could be made more precise.

Anarcho Capitalism

A second topic raised was anarcho-capitalism, a hypothetical system without conventional government institutions in which laws are privately enforced. My motivation for considering anarcho-capitalism, or ancap for short, comes from issues with existing voting systems. Voters in the US, Britain, or other modern democracies are expected to inform themselves extensively, yet the benefits of doing so are shared by everyone, regardless of their effort, leading to free riders. This weakens the incentive to become informed. Moreover, gathering detailed information about legislation can easily cost more effort than the expected benefits of one vote for or against a particular candidate are worth.

In a system of private rights enforcement agencies, by contrast, individuals choose the legal framework under which they live, and many such agencies could coexist within the same region. The payoff to being informed is therefore more immediate, which would address a significant weakness in current voting systems.

Bruce raised several criticisms, some more persuasive than others. One is that anarcho-capitalism is utopian, and versions such as Rothbard’s are decidedly so. But Friedman’s model of private rights enforcement need not be read that way. It can instead be understood as a critique of existing institutions. The rationally uninformed voter problem is real, and Friedman’s proposal shows that alternative private arrangements could mitigate it. Friedman requires only basic assumptions already common in economics and social choice theory, such as assessing voting systems by considering voter incentives and conducting cost-benefit analyses. In this sense, it is a natural model to consider, especially given the success of other private endeavours. It is not necessarily a blueprint for an immediately implementable policy, but it does increase our understanding of social choice theory, which in turn helps us assess and create legislation. Hence, Friedman’s model allows us to criticise theories of government more effectively, and as a criticism, it is not utopian.

Homeowners’ Associations and Social Contract Theory

Another concern raised in the podcast is that private arrangements, such as homeowners’ associations, resemble governments, since they create rules and charge fees to those living under them. Bruce suggests that people can become subject to an HOA without ever consenting, for example, by inheriting property governed by one.

Here, I think Bruce makes two mistakes, one technical and one philosophical. HOA rules come with the property, but one is not forced to accept an inheritance. It can be declined. Accepting the property, therefore, amounts to accepting the conditions attached to it, including HOA membership. The consent here is indirect and bundled rather than negotiated, but it is not accurate to say that someone becomes subject to such arrangements without any legal consent. This marks a meaningful distinction between HOAs and modern states, where individuals are typically born into jurisdiction without any comparable act of consent.

The philosophical error is to argue about the legitimacy of government based on whether it arises from legal contracts. Bruce’s suggestion seems to be that if contractual arrangements can produce HOA-like structures at scale, then existing governments and taxation must therefore be legitimate. This resembles social contract theory, the view that society is legitimate because we are all party to an inherited agreement. It is unclear whether Bruce attributes this view to anarcho-capitalists or endorses it himself. It is true that ancaps tend to debate whether taxation counts as theft, but this issue of legitimacy is beside the point. Questions of legitimacy have historically rested on ideology, religion, or doctrine, and have offered little guidance for practical decision-making. The only relevant question is: how well do different arrangements satisfy Popper’s criterion that bad leaders can be removed with minimal damage?

A more useful line of inquiry is, therefore, whether HOA-like structures satisfy that criterion. Do they allow those subject to their rules to replace leadership effectively and correct mistakes? Arguably, they do. Members have incentives to maintain property values and challenge poor governance, and less functional HOAs tend to be reflected in property prices, assuming buyers have access to that information. Switching HOAs by moving to a nearby property is also far easier than switching governments by moving to another country. Hence, in this discussion of practical issues of decision-making and incentives, the question of whether HOAs are like governments fades into the background.

An Alternative View

As such, I would like to offer an alternative version of ancap to consider, one that is not concerned with the legitimacy of taxation or with the universality of contract (again, one has to admit Bruce is right that some real-life versions of ancap are too concerned with this, just as there exist versions of it that are utopian). The idea is simply that private arrangements generally create stronger incentives than public ones, and that they satisfy Popper’s criterion better. Systems of government that are more privatised and more decentralised, both in having a smaller public sector and in decentralising decision-making, are likely to be more effective at removing bad leaders, whether those be leaders of large corporations, HOAs, or politicians. Likewise, we have good reasons to believe that private rights enforcement would be more efficient than its public counterpart. Framed in this way, ancap is a practical claim about incentives, error correction, and institutional design. The question is not whether governments are legitimate, since many existing governments may be considered legitimate for various reasons. Instead, the question is which arrangements minimise damage when mistakes are made and allow poor decision-makers to be replaced.