What's right and what's wrong with libertarianism

Peter A. Taylor
July 8th, 2020

Advocates of particular political doctrines...should beware of denying the context in which their doctrines can operate....

— Bernard Crick, In Defence of Politics

By "libertarianism", I refer to a broad, classical liberal tendency in (mainly) American politics. A very weak definition of "libertarian" would be a set of people with some vague attitudes about limited government and leaving people alone who want to be left alone. Libertarians reject what the Texas LP platform used to call "the cult of the omnipotent state" or what Dan Klein calls "The People's Romance".

I am deliberately avoiding the use of the term, "neoliberal", for reasons that Glenn Reynolds explains (hat tip to James A. Donald):

I find "neoliberal" to be the new "fascist" — a largely meaningless term, except as a catch-all for what the speaker dislikes. But the unhappiness with additional freedom on the part of the populace seems to be a constant.

What I'm getting at is that the libertarian movement as a whole has been strongly influenced by economic theory (Murray Rothbard, David Friedman, et al), Public Choice theory, and the intellectual and rhetorical legacy of the Libertarian Party. This intellectual and cultural legacy gives the movement certain characteristic strengths and weaknesses.

While this movement includes anarcho-capitalists, I am not much interested in them here, apart from the influence they have had on the broader movement. As Will Wilkinson complained, "limited-government libertarians tend to internalize more of the anarchist framework than they logically should." Also, although objectivists typically disown the "libertarian" label, much of what follows also applies to them.

Libertarians usually give both moral and practical arguments.

The practical case for libertarianism is heavily dependent on some economic "efficiency proofs". Basically, "economic efficiency" means trying to bake the largest possible pie, while simultaneously trying to dance around the question of how to make interpersonal utility comparisons. (See David Friedman's Price Theory.) These proofs are known elsewhere as the first fundamental theorem of welfare economics. The theorem depends on some assumptions (e.g. property rights being defined and easily enforced) that may be more or less true, depending on various contexts. Libertarian economists (e.g. Bryan Caplan, Tyler Cowen) tend to advocate legislation that makes sense in terms of this theorem, but which I claim often has side effects which they do not take into account. I claim that these side effects often undermine the context in which their arguments make sense.

Libertarians' moral arguments are typically appeals to natural law or liberty. These arguments carry some weight for most people, but their degree of cogency varies a lot with different audiences. Most people have other concerns in addition to liberty. See Jonathan Haidt's moral foundations theory. Too often, libertarians respond to these other concerns by trying to delegitimize them rather than by trying to understand and address them. Part of the problem is that these other concerns don't fit easily into the libertarians' worldview.

I try to illustrate the strengths and weaknesses of the libertarian movement in the following flow chart. The focus of libertarianism is on legislation: What kinds of legislation are consistent with (A) liberty, as understood in the "natural law" tradition, and (B) economic efficiency? The default color is black. I indicate areas where I think the libertarians are particularly strong in green, and areas where they are particularly weak in red.

 
Libertarianism in Context

 
Technological context

Suppose you're a cattleman in the old American West. There are "tragedy of the commons" problems with open range land. You want to solve these problems through privatization. Whether or not it is practical to do this depends on technology: whether barbed wire is available to make fencing affordable. Some libertarians get this, but some seem to take it for granted that suitable technology, something equivalent to barbed wire, will always be available. I regard technology as an exogenous variable; government policy can affect technology, but can't really control it. Maybe the technology libertarians need will be available and maybe it won't.

 
Distribution

I want to ask two related questions here. The first question is how much incentive one particular individual has to work just a little bit harder. The second question is how much difference there is between the ways two different people are rewarded for similar levels of effort, but different amounts of luck or talent. For the first question, I want to plot income vs. effort. For the second question, it makes more sense to plot income vs. the product of effort times luck or talent. But the curves I want to talk about have the same shapes in both graphs.

Consider several possible graphs of individuals' income vs. effort.

 
In an ideal world, I would like income to be more or less proportional to effort. (Warning: There is no theoretical agreement on what constitutes a "good" distribution of wealth.) Unfortunately, some people may be disabled severely enough that no amount of effort will enable them to support themselves. Libertarians don't have a coherent theory of charity, public or private, to address this, other than Objectivists, who seem to be flat-out against charity in general, which seems untenable.

A Marxist might want income to be uniform, independent of effort ("From each according to his ability, to each according to his need.") This seems crazy to me because, if effort and income are too loosely coupled, it's hard to motivate people to work except through coercion.

A third possibility is suggested by Ed Leamer's question of whether a computer is more like a forklift or a microphone. A forklift increases worker productivity, but it is also an equalizer. A 5th percentile Japanese woman and a 95th percentile Scandinavian man can lift the same weight if they are driving the same model of forklift. A microphone, on the other hand, creates a "superstar" effect. One Enrico Pavarati with a microphone can put thousands of second-rate singers out of work.

Every town had its supply of entertainers in the Middle Ages but now a single entertainer can serve the whole world.

Modern technology, modern intellectual property and anti-trust law, and possibly several other factors (tax law?) seem to be producing a lot of "superstar" effects. Libertarians should be worried about this. We don't understand it nearly well enough. Typical libertarian approaches to distributional issues are

1. to attack various actual government policies that arguably make matters worse^*,
2. to hope, without much theoretical or empirical basis, that the policies we advocate are at least distributionally neutral in the long run, and
3. to try to delegitimize distributional issues as moral or practical concerns.

^*For example, many welfare programs have misfeatures such that the benefits "fall off a cliff" (the recipient suddenly loses a large benefit) if he works more than a certain number of hours or his earned income crosses a certain threshold.

None of these approaches are satisfactory; the third is especially counter-productive.

 
Economic efficiency ("allocation") and the first fundamental theorem of welfare economics

This is the part of libertarian thought that seems to be nailed down the best. If you aren't familiar with this I recommend reading about the "efficiency proofs" in David Friedman's "Should Medicine be a Commodity" paper. I also highly recommend the lead essay in Thomas C. Schelling's Choice and Consequence, which is in large part a plea for people thinking about political economics to think about economic efficiency or "allocation" (How big a pie can we bake?) and "distribution" (How do we slice the pie?) as separate issues. If libertarians could confine political arguments to discussions among honest, competent economists to the relationship between legislation and economic efficiency, I expect that something resembling moderate libertarian positions would win 90% of the arguments.

 
Sociological context

I am dividing the sociological context into two parts, liberty and (group) loyalty, which I think of as analogous to micro- and macro-economics, or perhaps the difference between psychology and sociology. Jonathan Haidt, in The Righteous Mind, describes human nature as "90% chimp, 10% bee". I associate "liberty" with the chimp side and "loyalty" with the bee side. Haidt also talks about WEIRD (Western, educated, industrialized, rich and democratic) people who don't understand the bee side, which normal people understand. Other words for the bee side are "groupish", "hivish", and "asabiyah" (roughly, social cohesion).

The reason I think of chimp vs. bee in terms of micro- vs. macro-economics is the idea that some interactions can be broken down into a series of one-on-one interactions, and some can't. Some things can be understood using the individual as the "unit of analysis" and some behaviors seem to demand a larger unit. I may guard my reputation because it will be important in my future interactions with others (micro), but people seem to find most of the meaning of life in promoting things that are greater than themselves (macro).

The sociological context is also called "the moral fabric of society".

 
Liberty

Liberty ("natural law", "the traditional rights of an Englishman") is both part of the sociological context that makes libertarian policies possible and a result of those policies. The desirability of liberty as a result of healthy political and social life is intuitively obvious for libertarians. Mencken described this in terms of trust, and indeed, social scientists often talk about high-trust vs. low-trust societies.

What avails it for a man to have money in the bank and a Ford in the garage if he knows that his neighbors on both sides are watching him through knotholes, and that the pastor of the tabernacle down the road is planning to have him sent to jail? The thing that makes life charming is not money, but the society of our fellow men, and the thing that draws us toward our fellow men is not admiration for their inner virtues, their hard striving to live according to the light that is in them, but admiration for their outer graces and decencies--in brief, confidence that they will always act generously and understandingly in their intercourse with us. We must trust men before we may enjoy them.

— H. L. Mencken, Notes on Democracy

Libertarians get this.

 
Group loyalty

What motivates suicide bombers? Libertarians don't get this. I certainly don't.

Would you rather buy an apple from a cousin or from a stranger? How much extra would you be willing to pay to support your cousin? We know this is important, but we don't know how to model it. If your cousin is a crook, will you rat him out to the police because it's the righteous thing to do, or do you put family above abstract principles of justice (Garett Jones' "amoral familism")? Will members of prospective immigant group X assimilate to Anglo-Dutch ideas about natural law, and just be regular folks, or when enough of them get together ("peer effects"), will they become predatory?

The practical case for libertarianism (i.e. the First Fundamental Theorem of Welfare Economics) depends on secure property rights, and whether we like it or not, the security of your property rights depends on what kind of scruples your neighbors have. In order for libertarians to be able to carry on an intelligent conversation about immigration policy, we need some good theories about the nature, origin, and evolution of scruples in a multicultural society.

 
Legislation and political context

I have "Legislation" in green, but there's a problem. We are good at figuring out how legislation affects economic efficiency (at least in theory—see Ellickson), but how does it affect the political context? One often sees articles by libertarians arguing that the government not do something because that power is likely to be abused. For whom is this advice intended? If libertarians controlled the government, we wouldn't need these "constitutional" restraints, except as reminders not to do what we don't want to do. If libertarians don't control the government, how do we get our policies implemented?

It seems to me that we have to talk of influence rather than control, and that influence means thinking probabilistically. If we have a 60% chance of getting a bill passed that we like, and bill C only makes sense if bills A and B pass, the odds of all three passing are less than 22%.

The questions we need to ask are things like, "Does this 'tax cut' or exemption create a conflict of interest that lowers the probability that we can influence subsequent legislation? If significant numbers of Venezuelans immigrate to the US, will they or their children increase the probability of electing politicians who tend to resemble Hugo Chavez rather then Calvin Coolidge? Will technology, trade, or immigration policy alter the distribution of wealth in ways that will lower the probability of libertarians being able to attract the support of working-class voters?" We need to take probable future government-mediated violations of our property rights seriously as something we are justified in using government to prevent.

In short, libertarianism only makes sense within certain contexts, which have been variously described by Nick Szabo:

If politics could be deduced this might be called the Central Theorem of Politics—we can't properly respond to a global initiation of force without local initiations of force.

Jack Crassus:

Libertarian principles sketch out an idealized system of property rights (basically 18th century British common law) as a moral optimum, but they lack convincing mechanisms as to how those rights will be enforced without the power of the crown behind them. Furthermore, they consider the power of sovereigns to be illegitimate/immoral.

Absolute property rights without sovereign force is the libertarian ideal. A unicorn. While I agree that property rights are attractive, the zero aggression principle has to go if libertarianism is going to travel from an abstraction to a reality.

and Mencius Moldbug: From Mises to Carlyle: my sick journey to the dark side of the force.

Comments?