Contra Kling on State Capacity Libertarianism

Why I think that really is an oxymoron

Recently Arnold Kling was on EconTalk discussing state capacity libertarianism with Russ Roberts, and it brought up a lot issues I have with the notion itself. I want to preface this by stating that I really respect Kling; his blog is one of the few that I check every day or so, and have been doing so nearly since its inception. I have read all his books, and Specialization and Exchange really influenced my thought and my teaching. Not so much that you should blame him for me, but if you find you like some of the ideas I write about you should definitely get a copy.

Further, Kling is an intellectual operating outside academia. I mean, ok, last I checked he taught high school economics, but I don’t think many would consider that Ivory Tower. As such, I think he is really an exemplar, doing good intellectual work in additional to his actual job, work that probably is not so lucrative relative to the effort he puts into it. For those of us outside of academia working to improve discourse and thought as effectively a hobby, he is someone to look up to.

I point out all that because I am about to be pretty hard on him for what he says, and I don’t want anyone thinking I think he is a bad person, or that he went to the dark side, or started taking Facebook or New York Times’ money to parrot their lines, etc. I believe his ideas in this case are wrong, but people should still read them.

What is “State Capacity Libertarianism” anyway?

Kling starts right out by contrasting “naïve libertarianism” and “state capacity libertarianism” (hereafter NL and SCL).

Arnold Kling: Okay. Let me start out by trying to create a contrast between what I'll call naive libertarianism and what Tyler Cowen calls state capacity libertarianism, which sounds a bit like an oxymoron. Basically a naive libertarian, as I term it, is somebody who just wants the state to be as small as possible, just protect property, keep the peace, don't do anything else. Keep the state in this small box. What I'm calling a state capacity libertarian says: modern life has many threats and opportunities that seem to call for government, that most people expect government to do something about. Better that government do those well than do it poorly because if government does those things poorly, it ends up even infringing on liberty even more. That's how it becomes less of an oxymoron.

That’s a pretty good definition I would say, and sounds a lot like Tyler Cowan’s explanation I remember from a few years back when he was promoting the idea a bit.¹ It also has some nods to Dan Klein’s distinction between direct and overall liberty, the idea that some limits on people’s direct personal liberty might result in increasing society’s overall liberty and thus be desirable. The examples Kling cites later revolve around that distinction:

• COVID: People are going to expect government to fix issue A, and if government does a poor job the results are going to be even more awful than if the government does nothing.

• Power grids: If the government doesn’t make sure power grids work all the time, when there are issues with a power grid failing we are likely to see really awful abuses by government.

So, if I may restate the SCL position, it is that there is a sequence of events to address:

A: People want government to fix certain things, so…

B: Government is going to respond by doing things to try to fix those things so…

C: Given that government will act at the behest of the people, we need a government that will do it well, because getting the bad version is really bad.

I have a bit of sympathy with that process. I think that humans are constitutionally incapable of not having government of some form; what starts out as non-government organizations providing governance starts to slip into normal government because sooner or later some group is authorized to ensure compliance with the governance by punishment up to death. Once you have that, poof, you have a government as we know it. To me the question is really what do we want government to do (what is it good at doing, what is it better at doing than other organizations, what is dangerous for it to do, etc.) and not whether one is even needed. The answers to those questions tend to put me in the naïve libertarian camp because the answers often point to government being the best at doing very little, all the while suffering the usual dysfunctions that occur when violence is introduced as a way to solve problems.

However, as Kling describes things, there isn’t necessarily any mutually exclusivity between NL and SCL. NL works on point A above, arguing that people shouldn’t want government to fix certain things, and attempting to change their minds.² SCL is what you do when changing people’s minds on A fail.

I think this unrecognized point is the first red flag that SCL posed in opposition to NL throws up: it accepts that government³ is the answer to any given question, and proceeds from there to discuss what government solution gives the best outcome.  Government is the answer; what’s the question?

If that sounds extreme, engage in a thought experiment with me. What sort of issue would not be brought into the political governmental realm with the logic that Kling puts forward? Obesity is a problem in the US; People want the government to fix this problem; therefore the government will try to fix this problem; therefore the only question is how to ensure the government does this well. Essentially any issue you care to name with enough people to looking to government for an answer such that the government wants to provide an answer is defined as something government should do, because remember, only NL suggests that the answer should be “government should stay out of it.”

At best, this is the current state of US politics, where as Russ Roberts has put it “both parties want to take my money and give it to their friends, they just have different friends.” The only limit on what government should get involved in is what the people in the government feel like getting involved in.

Does last sentence that sound too different from “People want government to fix certain things” to you? It should, because again, SCL doesn’t care about that part, but rather part B, “Government is going to respond by doing things to fix those things.” How many people have to ask the government to fix things before the government responds? Presumably very few considering how many things the US government does that no one has any idea about, not to mention how many things government does whether people like it or not. The number of people who matter are far fewer than a majority of the citizens, or even a majority of the voters, if that is what you were thinking. Getting a majority in the US House of Representatives requires far fewer than half the voters, likewise for the Senate. Presidents periodically lose the popular vote, and turnout for presidential elections is generally under 60% of the voting age population, meaning close races will see the winner getting <40% of the voting population’s votes.

So that’s the first problem with State Capacity Libertarianism: it isn’t libertarian, but rather standard political thought where nothing is to be inherently outside the purview of the state should enough people desire it, which is to say, should the state decide to put something within its purview.

Much like the “libertarian paternalism” of Cass Sunstein and co. all the work is done by the words hiding behind the smokescreen of “libertarian.” Interestingly, the two seem to approach the problem of individuals wanting bad things from opposite sides: Libertarian Paternalism says people are better off if smart people like Cass Sunstein can make the good choices the default and so create a small cost for choosing differently, so people too foolish to make the right decision on their own are taken care of. SCL says people are going to want the bad decision, so just give it to them and spend all your time figuring out how to polish that turd and stop suggesting they shouldn’t want it. Neither is quite comfortable with both letting people make their own decisions and trying to convince them that your sense of the better is in fact better. SCL assumes governmental coercion is the answer to the problem, the relevant question being just how best to apply that coercion; quite clearly not libertarian. Pretty much the antithesis, really.

That isn’t to say that SCL is bad in content! It is just misnamed in a way that would make the Ministry of Truth proud. SCL is what you do once the libertarian option of “just don’t get the government involved” is off the table. In effect SCL as advocated by Kling is just looking at how to get a government to operate better by bringing in checks and balances through clear lines of accountability and responsibility. If you need a government, which we probably do for at least some things, we should try to make it effective and efficient, which probably involves accountability. I am sure there is at least one academic field that focuses exactly on that sort of thing, although I struggle to come up with a name for it. Governmental administration? Public choice theory seems to touch on it a lot. Constitutional design? I don’t know, maybe there isn’t one and there should be, but it shouldn’t be called “state capacity libertarianism.” Maybe something like “institutional accountability systems”?

So, in a nutshell, to the extent that state capacity libertarianism is a replacement for naïve libertarianism, SCL is bad. Once you have given up the argument that government should be limited and instead come out with essentially “Ok, fine, anything the government wants to control is fine, but at least do it well,” you have given up on freedom and liberty and are interested only in suggestions as to how to make your masters more fair. Why the government would take any interest in your suggestions is another question entirely; after all, once you accept that the government can legitimately control whatever it wants, what carrot can you hold out, what stick can you threaten, that those in power will not simply take as they see fit?

---

But that’s not all…

I am going to resist going through the podcast point by point and needling at Kling’s points, and instead address some of the larger gaps in the logic, with maybe a few pointers to where those gaps appear. I do recommend listening to the entire podcast, as Russ Roberts does a good job pushing back against Kling, and Kling himself does seem to understand that his suggestions are somewhat vague and handwavey.

1: We just need better government behavior.
One of the key tenants of Public Choice Theory is that you are not allowed to say “We just need better people or rulers.” Saying “Communism would work if you just had the right people in power” is just sidestepping the twin problems of “The people who want power are not the right kind of people to have that power,” or “Unrestrained power turns people into monsters.” In other words, if your plan for governing human societies only works if humans are different from what they are, it is a bad plan for governing human societies.⁴

Likewise, saying “We just need better behavior from governments” is sort of dodging the real question: How do we know government is a good answer to this problem? There are many well documented cases of government failure, and simply stacks of reasons to believe that government is inherently unlikely to be able to perform certain functions well. Good government behavior in some situations could well be “Don’t try and fix this.” All the capacity in the world can’t help if the solution is inimical to the goal.

In a similar vein, Bryan Caplan wrote an excellent post along these lines over at Econlib, State Priorities, Not State “Capacity”. Maybe you should just stop here and go read that, after clicking the subscribe button there…

… but come back!

I would sum up Caplan’s point thusly: governmental behavior and outcomes depend on capacity (ability to act) and priorities (choice of actions). To do X you need both the capacity and desire to do X, but just because you have the capacity to do X does not mean you will, in fact do X. If anything, recent developments have demonstrated that the US government has immense capacity to do what it wants, but wow… the results are seriously disappointing.

So if you have a strong state like the US, to get the better state behavior that Arnold Kling wants, you have to either have politicians and bureaucrats with better priorities or systems of institutional accountability that align politicians and bureaucrats priorities with the general populace’s. The former is magic hand wavy nonsense. The latter is also rather difficult, perhaps more difficult than Kling realizes.

Let’s just poke our heads into this rabbit hole⁵ by asking the question “What are the priorities of the general populace?” How do we figure that out? How do we rank them? When in conflict, how do we resolve those conflict and trade off priorities? When there has to be one rule for everyone, can one size ever fit all? What if their priorities are bad?

Short answer: social utility is meaningless, humans just want too many different things and want to try out too many different things for us all to agree on any given end state. Everyone has a different rank order of priorities, or more appropriately, different preferences across trade offs for various things, such that there will never be 100% agreement on a best choice state. As a result some number of people are going to be harmed, their priorities ignored and violated, by any government action. Making a decision for the entire group necessarily means government priorities are going to be different than the priorities of some, many or most of the governed. That is even assuming that the priorities of one or the other are even possible to achieve.

Short example: Suppose the people of the USA unanimously prioritize “Make COVID not be happening!” The government hears the people crying out with one voice, and since they all individually want the same thing, being members of the populace themselves, the government stands up as one and loudly proclaims “WE WILL MAKE THIS NOT BE HAPPENING!”

But how? What if the government can’t agree on how to do that? Some press for lockdowns to limit spreading until vaccines are available to immunize everyone. Some oppose lockdowns and press for treatments to limit deaths until natural immunity can assert itself. Which is better? Will vaccines be ready before the locked down economy craters and everyone dies of starvation? Will treatment options keep enough people alive to have an economy after the virus rips through everyone? Is there maybe some middle ground? Some entirely different option? Can that knowledge problem be solved?

And… what if you simply can’t make COVID stop happening? What if vaccines can’t wipe out the virus and people can still transmit and periodically get sick from it? It happens with flu⁶ vaccines all the time, and really we humans have only eradicated one single disease in the wild.⁷ What if having had COVID and recovered just fine you are still likely to catch a variant and pass it on, so herd immunity is more like “moderate herd resistance”. The disease will never stop happening, just stop being quite so new and horrifying, settling into being a version of the flu that is more deadly to the elderly and less to the very young. What is a government to do when number one priority of the governed is impossible to achieve regardless of capacity?

A scarier question is “What couldn’t a government with massive capacity do when its citizens demand the impossible?”

2: The Constitution doesn’t mention power grids.
Ok, this is just sloppy thinking. So sloppy, here’s the quote; it is right at the beginning of the podcast if you want to hear the words.

Arnold Kling: …I also think that the modern life has created many new threats and opportunities that did not exist 250 years ago when we wrote the Constitution. The Constitution does not mention the words 'electrical power grid,' amazingly enough, given when it was written.

Russ Roberts: Go figure.

Arnold Kling: But, it does have the words, 'promote the general welfare' in the Preamble, which of course can be used as an excuse for almost any government activity.

But, I think the power grid example is legitimate. And, I think there are more and more legitimate uses of government power these days. A few weeks ago, I guess by the time this appears it'll be longer than a few weeks ago, but you had the podcast on the challenges of property rights nowadays. I think those have just become much more ambiguous for a variety of reasons. You have of people living closer together so that more of the things I do with my property affect you and your property. I can block your view. I can put up a noisy establishment next to you, whatever.

The more of our assets are intangible--and intangible assets give rise to a lot more ambiguity in terms of property rights. If I'm on a farm I know what my land is, I know what my farm animals are, I know what my machines are. But, when you have something, let's say as complicated as an iPhone, based on a lot of inventions and ideas, who really owns what has to be negotiated a lot. Things like spectrum, the use of the spectrum for communication, that has to be settled somehow. The use of ideas, trade names, reputation.

So much wealth is now intangible, and this intangible property is just not as clearly delineated as to who owns what. That makes modern life more complicated, and you get a lot more regulation.

Financial transactions are way more complicated than they were 250 years ago, and that probably requires a lot more regulation and government involvement. I think all these trends--all these threats and opportunities for government involvement. As an opportunity I could cite maybe research into nuclear power or research into longevity. Health-related research, whatever. They've just multiplied over the last couple hundred years.

And, life has become very complex, harder to regulate, and so there's just going to be more governance. Now, I think what we're going to see more of is more governance by private sector entities. I mean, the classic ones, the things like underwriters laboratories setting standards and so on, or the Internet engineering task forces--that's basically private sector governance. But, I think some of it is going to be more additional, more formal governance from government. I think that that's just the reality that we have to deal with.

You know what, the Constitution doesn’t mention horses, either. Or watermills. Water is mentioned exactly once, in the context of giving Congress the power to make rules concerning wartime captures on land or water. Roads get a single mention, in the context of “Establish Post Offices and post Roads.” Food is never mentioned, nor medicine. Even thousand year old technologies, even the very basic necessities of life from time immemorial go without special mention in the Constitution. So what possible import does the lack of the words “power grid” make? Am I to take away that since the Constitution does not include “Congress shall make no law infringing on the right to create power grids” that we should just ignore the document?

Charitably one could say “Well, the point is that there are many new things that are super complicated and need government that the framers couldn’t have foreseen.” Sure, there are new things, but there are always new things. The Constitution was drafted during the industrial revolution, not shortly after the fall of Rome. The Constitution wasn’t written in the context of peasant farmers and knights. Massively complex financial institutions, intellectual property, global trade networks, international corporations, all these things already existed. The immense legal complexities of the modern age compared to a few hundred years prior is largely a modern conceit. What has increased exponentially is the amount of legislative and regulatory complexity, blame for which lays as much at the feet of government itself as anyone else. As ever, before we start asking “How can government be used to address this problem?” we need to be asking “Should government be used to address this problem?”

Kling himself should be asking that NL question, considering the first sentence of that last paragraph I quoted: “life has become very complex, harder to regulate, and so there's just going to be more governance.” [emphasis mine] Is his argument really that as aspects of life become more complex and harder to regulate that implies we should task the government to regulate it more? Governance does not need to come from government, and government is often not the best way to get good governance, especially for highly complex and hard to regulate issues. Kling clearly understands that. So why, when the issues become harder to regulate, should we just assume that government regulation is the solution and just try and figure out how best to get government to do a good job? ⁸

3: We need a Chief Operations Office / Office of Accountability / some group that assigns responsibility, blame or praise to the bureaucracy and specific actors to get them to behave better.

If we are in a world where the problem is the priorities and incentives of state actors are not lined up with citizens, what good does adding another layer of priorities do? You can’t solve the principal agent problem by adding some more unaccountable agents to report on the other agents. If anything, that makes it worse, increasing the chances that your agents collude against you, or start treating the reporting agents as the principal. If the government of the US is now too large, complex and opaque for citizens to keep an eye on and judge its behavior, making it larger and creating an unreliable narrator of its actions is not an obviously good move. How do you trust the word of the Watchers when you can’t independently confirm what they are telling you?

---

That all isn’t to say this line of inquiry is pointless or negative.

We definitely need to work on how to align the incentives of government actors better. The Founders seem to have been wrong about a few things, such as how jealous of their powers the states, federal, and federal branches would be amongst each other. I don’t think they saw the regulatory or administrative state coming, or that the various states would give up immense power to the federal government in exchange for grants of national taxation. These are much bigger problems derailing the beneficial behaviors of the state than intellectual property or the internet. Any assessment of the problems of modern governance by the state needs to keep in the forefront “How do we keep the state from getting involved in areas where its presence is negative?” in addition to “Now that we have the state involved, how do we keep it performing properly?”

Those two questions are inextricably linked. We need to have at least a vague idea of how well the state could perform some task before we decide whether or not it should get involved, and so we need to have a rough idea of what sort of performance we can expect from the state. When asking how to organize the state institutions so that they perform better, we need to know what realms of human action should be kept out of state control, not only to allow alternate sources of power to help constrain the state, but also what powers must be denied the state entirely, lest they become the tool of its corruption.

Because you know what the Constitution does specifically mention? Not establishing a state religion. Freedom of association. Not regulating speech. The people’s right to keep and bear arms. Protection from unreasonable search and seizures. The right to due process. Quite a few more in there, too, all of which are the sorts of things that once government starts infringing upon ruins the whole enterprise. No matter how many people want the government to act, or how well the government uses its capacity to do so. The founders definitely got that right.

So in conclusion…

Listen to the podcast. While doing so, ask yourself how much of the problem is really that no one is pointing out the problems of government behavior, and how much is that we can’t really decide on what the government’s behavior should be, along with how much is that we can’t punish the bad actors no matter what, anyway. Then consider whether some of the problem wouldn’t be solved by taking some decisions out of the hands of distant unaccountable actors in the first place.

1. I also highly approve of defining just what it is you are talking about right off the bat… I hate it when people talk about their idea without actually defining it first.

2. Notably, point A is where the Direct vs Overall Liberty idea is relevant as well, focusing on whether we should want to limit some specific direct liberty for the benefit of overall liberty, prior to the discussion of how best to do that once we have decided to.

3. I am going to use state and government throughout the essay pretty much interchangeably. Sorry.

4. This seemingly is, incidentally, why the “blank slate” model of human nature, the claim that humans have no nature and are just whatever they want or are told to be, is so popular with utopians past and present. If humans can be anything, any reorganization of society can work with sufficient changes to the people. You just need that “New Man” they talk so much about. If humans are pretty much humans all the way down, though, you have some serious limitations on what you can get them to do. That cuts the other way too: if humans have pretty set natures you can predict their behaviors in various settings pretty well and plan for that, whereas if they are perfectly malleable you have no real idea how they are going to behave as the contents of their slate can change at any time due to any impulse. Strangely, the utopian social planners never seem unduly worried about that fly in their blank slate ointment.

5. Social welfare functions, group decision making, utility functions of groups… when I say rabbit hole, you have to imagine the rabbit is using apparatus appropriate for offshore drilling to make his holes. Spend enough time descending and you will start to notice the rocks taking on a rather warm, rosy hue and the temperature becoming decidedly toasty.

6. “The other coronavirus!”

7. Smallpox. Everything else is still out there, with periodic outbreaks.

8. Especially considering his example of iphone intellectual property has been a complete shit show. Should we really assume the same state that brought us IP law resulting in that mess, or e.g. copyright that ends only when Disney wants to stop paying off legislators, is the answer to IP concerns?