Happy New Year! Since it is the new year, we have a new article about state capacity libertarianism from Arnold Kling (that’s his Substack, the full essay is here.)
For those of you who read me or Dr. Kling frequently, you will recognize the bones of the argument: like it or not, the administrative state is necessary to ordering society and never going away, so we had better figure out how to make it work better than it does, which is really bad. In fact, if you have listened to e.g. Kling on Econtalk, you have heard the entire argument before; no real changes or refinements have been made since then. I don’t see any real changes since I wrote about it in October, in fact; presumably Arnold didn’t see any virtue in my arguments. Or perhaps my arguments were opaquely written… it wouldn’t be the first time!
So let’s break down the issue a little more clearly. The key mistake I think Kling (and Tyler Cowan, who came up with the idea) make is that first part about the administrative state is necessary to ordering society. Kling describes the issue thusly:
Conservatives are reluctant to concede that there is anything fundamentally right with the regulatory state. In their view, it was a mistake to permit executive agencies to develop and coalesce into a "fourth branch" of government. Some even go so far as to describe the regulatory state as an alien tyranny foisted by arrogant progressives on an unwitting public.
While the conservative critique of the regulatory state is not unfounded, from a historical perspective, the rise of regulation — and the attendant rise in the need for institutions to develop, adopt, and enforce new rules — was a necessary response to the dramatic changes that took place in our society after the founding. Such changes led to an increase in the complexity of daily life and, in turn, a rise in the complexity of the challenges facing our governing institutions.
Four historical forces in particular help explain why the challenges of governance are more complicated in the 21st century than they were when the Constitution was written: greater urbanization, the rising importance of intangible sources of wealth, increased specialization, and the digital revolution.
So the legislature (at any level: city, state or federal) cannot be relied upon to pass laws dealing with the complexities of modern life, nor can the judiciary work out precedents case by case that sort things out, and so both need to delegate rule making authority to regulators to create rules with the force of law to deal with these problems. For reference again, these are the four problems:
Increasing city life
Importance of intangible property
Increasing specialization
The digital revolution
The trouble with the regulatory agencies as they exist is primarily that they are without any oversight or proper incentivization to not be terrible. With the proper oversight and incentives the regulators would be able to order things very well; to get that oversight, Kling recommends creating a Chief Operating Officer (COO) and Chief Auditor (CA) with organizational backing to determine what regulatory agencies are doing well and which are doing poorly, and hold the relevant people accountable. When a crisis pops up, if the “peacetime bureaucrats” are failing the COO can fire them and put in new ones. If the COO fails to see the failure, the CA can pop into his office and point out how badly the regulators are doing.
Doesn’t the president already have that power? Well… legally, who knows, but in practice, no, he does not. Remember when President Trump tried to ax big parts of the Department of Energy? Well, ok, yea, I only remember because I was lecturing at GMU at the time and had a handful of students in class with family members who worked there at the time. Short version: it didn’t go well. In effect, the president said “Get rid of groups X, Y and Z!” and in response the managers of groups X and Y changed their titles to A and B. The people in charge of firing people looked at the roster, then replied “Ok, groups X and Y don’t exist, so job done. Group Z though, we fired all them and closed it down,” and that was that1. No one in the Department of Energy has any incentive to shut down internal groups, and every incentive to protect their people, so they effectively could stall forever. Apparently the executive can eliminate positions, but not fire people by name, probably for union reasons, and so by reassigning people to different positions defined by different titles of organization, one can avoid ever reducing head count if one wants. A COO who could directly fire people could get around this, and so enforce efficiency. In fact, Kling’s COO seems to be almost defined as being able to fire people who need firing in the bureaucracy.
So, what’s your problem, Hammer?
Two problems, actually. (With Kling’s formulation, at least…) With the state capacity libertarianism thing, the problem with the administrative state being necessary has two problems:
Can the administrative state order society even in its ideal state?
Is the administrative state the only way to order society?
Then you have the problems of the COO and CA:
How does the COO and CA effect changes the president can not?
How do the COO and CA actually know who is doing badly?
How do the COO and CA get put into office?
How do we incentivize the COO and CA and hold them accountable?
I want to focus on the necessary bits, but the issues of the CooCa (which I will now be using to refer to the COO and CA idea) will be addressed. In fact, I am mostly going to gloss over/ignore “2. Is the administrative state the only way to order society?” and focus on 1.
Is the regulatory state necessary?
Kling explicitly claims that the regulatory or administrative state is necessary to ordering society. Implicitly, he claims that the regulatory state is capable of ordering society. That is, given the right organization, incentives and authority, the regulators could sort out all the problems they are supposed to be sorting out.
My first question to Arnold would be: what makes you think they can do this?
Arnold brings up the Texas power grid problems of the last year of an example of the state needing more power to regulate to keep problems from happening that make people demand more government intervention. Sure, Texas has relatively little government involvement in its power industry, and hasn’t for years. Strangely, it only had an issue with that over the past couple years during the pandemic, but still, Texas could definitely have a more secure power grid if it had more state oversight like, say, California does. California has a great power supply, right? Very reliable?
Yea… it doesn’t seem the governments are very good at making sure power works right. Every US state other than TX has immense regulatory control of the power industry, and results are not amazing. Partly this seems to be because of the political focus on renewables, particularly wind power, which don’t do really well in adverse weather, but partly it is because of bad incentives when it comes to power provision in general. Replacing coal, oil and gas plants with nuclear would be a big win, but every state is going in the other direction it seems. We have better technology and rely on electricity a great deal, yet every state seems to have pretty poor outcomes here despite a lot of government oversight and control. Why is this the example of government needing to step in? They have been in since the beginning and are not doing well.
Let’s see what else Kling brings up…
COVID: Uhm… did ANY government succeed here? Sweden’s government seems to have done well by, well, not doing much. Could any state have solved COVID?
Banking: Yea… this one is regulated to hell and breakfast already and it isn’t obvious that has done anything to improve things.
Transportation: Isn’t building roads almost the monopolistic domain of the government? Right, it is… private builders need elaborate permissions to make toll roads and often have to share tolls with the local government, who usually gets to set them. Yet everyone complains about the government roads… why do we think government can do that well?
That isn’t promising, but let’s zoom out and look at the big four issues Kling cites. Maybe we can do better there.
Increasing city life: Firstly, cities have been around for a LONG time [citation needed]. All the problems of externalities Kling brings up existed in London of 1500, not to mention ancient Rome or Babylon. The issues are probably easier to deal with now, however, thanks to better technology. Yet cities are still a mess.
Importance of intangible property: Kling points out that property rights are one of the big things government is supposed to solve. Have they? Oh… right. They currently suck at it, he says.
Increasing specialization: Go to the BLS list of occupations and see if you can find yours. Yea… specialization has WAY outpaced what government can keep track of. You can see the issue more easily looking at the census data for race: 6 total categories. Asian includes "Asian Indian," "Chinese," "Filipino," "Korean," "Japanese," "Vietnamese," and "Other Asian"… yea, probably nothing more detailed is useful there, basically all the same, sure.
Arnold, you wrote a wonderful book about how economists (and by extension the government) misunderstand specialization and trade and need to rethink things. Why do you think they are capable of being wise experts creating order when they have waited till you to get the book written and have so far failed to accept it?The digital revolution: Well, thus far the internet works pretty well, and every state intervention suggested seems to make it worse, up to straight up censorship of dissent and scientific discussion. What improvements do you see the government providing?
So, Dr. Kling, what makes you think the administrative state can solve these problems? Or, to put it in a better framing, what makes you think the administrative state is the right tool for these jobs, as opposed to market solutions? If your answer is “Because people seem to think it should be” I would suggest that the goal should be attempting to change people’s minds, not simply saying that what people think is always true.
Why do I think the state cannot do the things Kling (and other people) want it to do? In part I agree with Kling: the state is not good at incentives and accountability. Kling assumes away a huge problem, however. Even if the state did have good incentives and accountability, it doesn’t know what is good!
The state has a knowledge problem.
How do we know what to do, and if our state is doing it?
What should the CDC have done? Would it have worked? What is the standard for how the FDA and CDC should have behaved, and how far did they miss that mark by?
Those are big questions, and need better answers than “They could have done a lot better than they did.” Maybe they did the best they or any other organization could. How would we know?
Typically we look at other countries and compare how their governments performed. We might compare the policies of the US, the UK, Sweden, France, etc. and see what the outcomes looked like. The trouble is, that only gives us rank orders, who did better or worse. If every country did a bad job and there was some really good but untried policy out there, well too bad, we can only see whose bad job was least bad, not find out what we should have done.
Confounding even that limited information is how different the countries compared are. People are still arguing whether Florida or California had better COVID outcomes so far, let alone Sweden and the USA.
Do you think the answer of who has better banking regulations is clearer?
Kling’s suggestion neglects the very important role of competition in markets: discovery2. One big reason to have lots of companies competing against each other instead of one big monopoly regulated by the state is that trying out lots of things allows us to find out what works. A monopoly has a hard time improving because you don’t know if something works until you try it. A business leader without competition only can compare their current organization against how it did in the past; competition lets you see how good you could be if you had made different decisions in the past.
At the government level you have no direct competition, and very little experimentation across various comparable groups, and so you have a very poor way of determining what is better or worse policy, and no one taking moon shots to try out policies that seem crazy.
In theory the USA could have state level CDCs and FDAs instead of the federal ones with different policies and we could compare how they did, but then again the comparison problem crops up. Were Florida’s policies better than California’s? Did Delaware have better outcomes than Rhode Island because residents could just go to PA for vaccinations? How do we judge?
You probably can already see where I am going with this, so I will fast forward a bit. I don’t see why one would think the state is capable of providing the rules and order Kling asks for, because it is not up to the task of generating and identifying the knowledge needed to improve itself, or even what the optimal outcome would even look like. I really don’t see how the CooCa would figure that out either.
Say the COO decides the CDC and FDA did a crap job and wants to fire all the peacetime bureaucrats. Hooray, I say. The COO doesn’t have to convince me, however, but rather everyone else. What does he say? More specifically, what does he say that the president cannot currently say? “Fauci, over 700,000 people have died with COVID since 2020! Anything over 200,000 is unacceptable because…” what? How many should have died? Sure, you can say “Uhm, less, asshole,” and I can reply “Fewer… fewer,” but we don’t really have a decent way of arguing how many fewer. We can look at different countries and say “Well, only X of 100,000 people died here, why couldn’t we get that?” and Fauci could argue that the US is different because of this or that reason, and then… what? We are right back to where we were before, where whether or not Fauci did a good job is a political decision based on vague feelings. If he were going to get fired for that, he would be already.
When the COO and CA look at the bank regulators, how are they going to measure success? I’d wager it would look a lot like “When things are fine, the regulators get gold stars. When things go south, they all get fired, except when no one could have seen the problems coming, which is all the time. So, status quo.”
Pick any problem the government is supposed to solve, and you will find the same problem: how do you measure success, and how do you identify how well the agencies should be doing?
Hell, we can’t even agree whether or not the police are preventing crime well or what they should be doing. Is there a crime wave happening right now? Who the hell knows…
Would a COO and CA at least be better than what we have now?
I don’t see how. The CooCa has exactly the same incentive and knowledge problems of current government actors, and no obvious way of removing the hurdles currently faced by the president.
Unions: If you can’t say “Look, this specific person should be performing at X, and is only performing at Y, so they are going to be fired” you apparently can’t fire a government employee. How does the CooCa get around that? If you say “Well, change the rules around state employee unions” I say “We can do that now, why don’t we try that first?”
Who appoints the CooCa? Does the Congress do this, or the President? Why should this work better than the existing Secretaries of X and the Congressional Budget offices? Why wouldn’t the CooCa be just as politicized as the Fed?
Who holds the CooCa accountable? If the CooCa is accountable to an elected official or group thereof, won’t they have a lot of incentive to evaluate policies as good or bad based on the preferences of the elected officials? “The EPA is doing great! Couldn’t be better! Biden’s CooCa says so!”
If it is a post outside political appointment, don’t we then have the same problem the bureaucracy normally has of officers unaccountable to the public? Do we even want officers accountable to the public? If not, then to whom?
It is not obvious at all how the CooCa idea solves the current problems, or even addresses them. In many respect it is like the idea that we need a monarchy, because then you would have someone above the system, free from all the bad incentives that limit proper behavior in our current government system, and the monarch can then just cut through all the crap and get people to snap into line. Except we know better, at least those of us who studied public choice theory. Monarchs have their own incentives and problems, and at the very best their reforms just result in a bureaucracy that matches the monarch’s sense of how things should work, not necessarily how things should actually work. Neither would a CooCa face better incentives, or have better knowledge; they would simply be another layer of bureaucracy.
So, where does that leave things? If I were going to offer constructive criticism, I would recommend that Dr. Kling give some serious thought to addressing the following concerns:
How do we know the administrative state can actually perform the ordering of society he believes it is necessary for it to do?
How does the COO and CA he is recommending judge the bureaucracies? What does it judge them against? How does it know what they can and should do, since the executive and legislatures can’t figure that out?
What incentives do the COO and CA face? Who puts those incentives into place?
I think the knowledge problems in particular are damning, but without a good structure for incentives to hold the COO and CA accountable I don’t see how it solves any problems.
Another improvement to Kling’s idea is to devolve the US’ federal regulatory structure to the states, and then have the states experiment with different policies, CooCa, whatever. You still run into problems like “Does FL or CA have better COVID outcomes?” but at least there are some comparisons that are more apples to apples. I don’t think this really solves the problems, but it would lessen them, largely by the same mechanisms that exist for interstate competition now. Not great, but better than one agency at the federal level that can only be compared to agencies in other countries.
I had a few students with parents (and one aunt) who worked in those groups; one great thing about 3 hour long lecture sessions is that you have time to stop and interrogate things like this for 10-15 minutes. Otherwise, goddamn, I wouldn’t have believed it.
That Cowan, head of the Mercatus Center, seems to miss this point as well is baffling to me.
Quoting he article you link to, "Liberty without order is not a sustainable state of affairs." Really, that sums it up. The statist cannot imagine any order but control by government regulators. Self-organization is merely chaos.
As for the "fourth branch of government", a very compelling case that the permanent deep-state has always been illegal is made by Cornell law professor Phillip Hamburger in his book _Is Administrative Law Unlawful?_ "The Constitution’s barriers to any subdelegation of legislative or judicial power may sound merely technical, but they were expressions of an old and crucial principle against subdelegation, which underlay the efficacy of constitutions. The logic was that once the people had delegated different powers to the different branches of government, any subdelegation of such powers would allow the government to evade the structure of government chosen by the people. Alas, this has happened.”
The New Deal Administrative Practices Act amounted to a coup, replacing the Constitutional representative government with a different kind of system.