Bryan Caplan  

Liberal Authoritarianism

PRINT
Who Caused the August 1990 Spi... San Francisco: America's Large...
Critics often view liberals as deeply authoritarian.  Most liberals naturally object to this unflattering claim.  Critics notwithstanding, liberals don't relish using the power of government.  They don't have a raw preference for forcing everyone live their way.  Instead, liberals maintain, they favor using the power of government to advance liberal aims because such policies have good overall consequences.

Are liberals seeing their collective motivations clearly?  Not really.  For starters, most liberals - like most human beings - don't know enough social science to begin to weigh policies' overall consequences.  The best they can do, as Kahneman explains, is covertly change the subject, then answer easier questions.  To evaluate the overall consequences of raising the minimum wage, for example, you need to know the elasticity of labor demand.  Few laymen even understand the concept of elasticity, so they mentally substitute easier questions like, "Would I be happy if employers gave low-skilled workers a raise?"

How does this show that most liberals aren't consequentialists?  Well, if most liberals don't know enough social science to weigh policies overall consequences, most can't honestly say, "I'm a liberal because using the power of government to advance liberal aims has overall good consequences."  If a liberal spends near-zero mental effort studying policies' consequences, something other than his beliefs about policies' consequences must be driving his liberalism. 

Such as?  Sheer love of government isn't the only possibility, but it's a good guess.  Consider: If you want lots of X, but are too ignorant to evaluate X's indirect effects, you probably just really love X.  If you want lots of ice cream, but are too ignorant to evaluate ice cream's effect on your health, you probably just really love ice cream.  If you want lots of government, but are too ignorant to evaluate government's overall consequences, you probably just really love government.

At this point, the thoughtful liberal may clarify his position: "When I claimed that liberals were consequentialists rather than authoritarians, I was only talking about liberal policy wonks like me who do know a lot of social science."  When a liberal grasps the connection between the minimum wage and labor demand elasticity - and hundreds of other esoteric policy points - his consequentialist self-portrait becomes fairly credible. 

Does this really refute critics' charge that liberals are authoritarian?  It depends.  You could define an authoritarian as "someone who relishes the use of government power."   On this definition, liberal wonks plausibly escape the authoritarian charge. 

But that's an awfully strong definition.  I'd suggest a more reasonable definition: an authoritarian is "someone who doesn't mind the use of government power."  This doesn't mean that you're an "authoritarian" if you favor using government power under any circumstances. What it means, rather, is that you're an authoritarian unless you have at least a modest presumption against using government power.

On the latter definition, "I'm a consequentialist" doesn't rebut the authoritarian accusation.  It confirms it.  Why?  Because consequentialism is inherently authoritarian! 

Suppose government forcing everyone to do A has slightly better consequences than the next-best alternative of leaving people alone.  True to his name, the consequentialist announces, "We should force everyone to do A."  A nay-sayer raises his hand and says, "What's the big deal?  I don't want to do A.  Leave me alone."  The clever consequentialist responds, "My calculations of the overall consequences take your reluctance into account.  So we should still force you to do A."  The nay-sayer nays, "The overall consequences are only slightly better.  Just leave me alone." 

In the end, the consequentialist has to either abandon consequentialism or say, "I refuse to leave you alone.  Although the difference between the best and second-best is small, you have to do A whether you like it or not."  And isn't that an awfully authoritarian attitude?

P.S. I leave the writing of the companion post on "Conservative Authoritarianism" as an exercise for the reader.



COMMENTS (33 to date)
Art Carden writes:

Challenge accepted! Look for a post titled "Conservative Authoritarianism" within the next few days.

silly sailor writes:

Isn't it simpler than this. For historic reasons socialist is a dirty word in America. In any other western country politicians who call them selves liberals would call themselves socialists with out the stigma?

Joshua writes:

I think the ice cream metaphor is a good one. I LOVE ice cream, but I choose to eat it only sparingly because I know the consequences. The difference is that I find many liberal policies not just unwise, but also immoral.

I love to give poor people money, but I choose to only give my own money instead of taking from one group and giving to another group because I know the consequences and I think it's usually immoral to forcibly take other people's money.

Greg G writes:

OK then. Time for all you authoritarians who have a presumption in favor of using government power to enforce contracts to repent.

MikeDC writes:

"Socialism" should have a stigma attached to it, just like other coercive and harmful behaviors.

Tom West writes:

Woah. Try Occam's razor.

Most people are well aware that the social sciences are way to vague, inaccurate, or occasionally flat out wrong to be a useful guide to estimating the results of policy.

So, we look at what other places do or have done, decide whether we like the outcome, and base our policy desires on that. General empiricism is as accurate as your ever going to get where human beings are concerned.

(By the way, no slight to social scientists. Modelling two people saying hello is harder than modelling the interactions of every star in the galaxy. Accurate social science is an impossible task because humans are simply way too complex.)

I suspect that what statists like myself *don't* have is a visceral horror at being told what to do. It's why I like to read sites that remind me that a fair number of people have really quite different utility functions than I do.

Ivan writes:

If you define authoritarianism as a collective action perpetuated by the government to coercively alter the actions of individuals towards a specific goal, then both conservatives and liberals would be considered as deeply authoritarian.

Key differences between the two: liberals justify authoritarianism on the basis of helping the disadvantaged and using government power to balance out what they perceive to be concentrated private power. Conservatives, on the other hand, justify authoritarianism on the basis of preservation of perceived moral values that underpin the structure of society.

Technically, both view authoritarianism as a means to end, which is to bring about a certain state that will improve/protect society. Although, from general observation, it seems that conservatives frequently value authoritarianism as an end in itself. A good society, in their mind, would have a strong power base, usually in a form of an individual that can be seen as a source of moral wisdom and absolute decision-making power.

It seems to me that liberals would be uncomfortable with the above-mentioned arrangement even if it were created in the name of their societal goals. They would, on the other hand, like to see this kind of power in the hands of a much broader group.

I would certainly concede the point that neither group is very concerned about the broader implications of their form of authoritarianism. I think this is because when it comes to political beliefs, humans have very strong preconceived notions on the correct course of action. In only very rare individuals, these notions are amendable by actual rational (not rationalization) thought and analysis.

Maximum Liberty writes:

Professor:

I suspect this approach is little more than arguing from and about definitions, so it probably fruitless. Nonetheless ...

Would a tie-breaker rule against government action be sufficient to count as a "modest presumption"? Examples in institutional decision-making might include:

  • The need for a positive majority vote in a nominally representative body to pass a new law.

  • The need for a jury to find for a civil claimant by a preponderance of the evidence.

  • The need for a jury to find a criminal defendant guilty beyond a reasonable doubt.

If so, then I would say that mainstream democratic socialists (whom you call liberals here) do not qualify as authoritarian because they accept these institutional limits on the exercise of governmental power.

Max

jc writes:

Many, if not most, people: (a) want to think of themselves as good people, (b) have tribalist tendencies and associated sacred tribal norms, beliefs, and values, and/or (c) don't mind, or even enjoy (openly or secretly), dominating others and compelling heathen nonbelievers to "do the right thing", conforming to their tribe's notions of morality and decency.

But, yes, many of these folks will pretend - fooling even themselves - that they're knowledgeable and have thought things through, or that answers are obvious, and that they are thus, indeed, enlightened consequentialists.

The truth is that they're merely human, rationalizing and justifying urges that they seem wired to follow, while signaling to others (and themselves) that they belong to an enlightened faction of their tribe.

"The first principle is that you must not fool yourself, and you are the easiest person to fool." - Richard Feynman

Philo writes:

We all give at least considerable weight to consequences (even those who *say* they don't). By your lights we are all "authoritarians."

Suppose a friend of yours, with tendency to depression, after experiencing a series of bad events tries to jump off a high bridge. You physically restrain him, hoping his impulse will pass and he will carry on with his life (which, so far as you can see, is not *all that* bad). Have you just revealed your *authoritarian* disposition, by interfering with the voluntary action of another person? No, that is a silly way to talk!

In practice it usually has best consequences to leave people uncoerced, so long as they are not harming others (admittedly, the definition of 'harm' will be rather tricky); a sensible consequentialist will be 99%, or even 99.9%, anti-authoritarian. And that is all he *should* be.

Kevin writes:
Most people are well aware that the social sciences are way to vague, inaccurate, or occasionally flat out wrong to be a useful guide to estimating the results of policy.

So, we look at what other places do or have done, decide whether we like the outcome, and base our policy desires on that. General empiricism is as accurate as your ever going to get where human beings are concerned.

This is essentially replacing (possibly) well-done social science with social guesstimating, though. I mean, all of the complaints you could levy at the social sciences are actually worse using the method you're advocating.

Butler T. Reynolds writes:

Is "consequentialist" the polite word for what we used to call a "commie"? :)

Noah Smith writes:

With all due respect, Bryan, this post is perhaps the most spectacular Ideological Turing Test failure I have ever seen...

Josiah writes:

Well, if most liberals don't know enough social science to weigh policies overall consequences, most can't honestly say, "I'm a liberal because using the power of government to advance liberal aims has overall good consequences."

That doesn't follow.

charlie writes:

The Left has for a long time used the language of "rights" as a rhetorical tool, just as much as the Right does. It's not clear that, in non-communist countries, one side of the spectrum is more consequentialist.

For example, social-democrats are obsessed with "positive rights." Abortion is justified as a right. So is gay marriage. Ronald Dworkin's influential legal writings in favor of left-leaning policies are based on the idea that "rights are a trump" over consequentialist justifications.

The Right has Locke, the Left has Kant; both are deontologists.

Matthew writes:

This post is just silly. Raising the minimum wage isn't a liberal thing--it's a policy with massive support from liberals, independents, and conservatives. A majority of republican-identified voters want a minimum wage hike, as do a sizable percent of the GOP congressional caucuses. Do they intrinsically love big government? http://www.gallup.com/poll/165794/americans-raising-minimum-wage.aspx

UserGoogol writes:

I think it's a category error to call consequentialism authoritarian like that. Authoritarian implies some sort of actual concrete authority which is imposing its will onto society, not just the abstract moral authority that one thing is better than another thing. And any consequentialist worth their salt should realize that individual hedonic calculations aren't really enough to justify restructuring society, so even if some particular calculation points in the direction of restricting people's freedoms you shouldn't just run out and do it right away.

And conversely, I think liberty and consequentialism go more hand in hand than you imply. Freedoms clash. "Your right to swing your fist ends at my nose" but why should I be able to oppress your desire to punch me? There are various deontological answers to that about self-ownership and such, but the more consequentialist answer "because hitting their nose causes more suffering than your benefit of being able to swing your fist around" seems pretty reasonable too, and it hardly seems authoritarian to say that you should consider other people's feelings.

Brian W writes:

Something I would add to the many great comments here and the post is that both sides seem to use language to make their ideas seem less authoritarian and probably to justify this truth to themselves. They will say if you don't support X then you hate Y even if the two have nothing to do with each other. This allows the policy to seem to be what is collectively good for everyone and the idea of the government using force to enforce the idea doesn't seem to be as harsh.

Dan Hill writes:

Imagine how much ice cream would get eaten if:

a) For various perfectly valid reasons it was almost impossible to isolate the effect of ice cream eating on health from myriad other factors, even though logic tells us ice cream makes you fat

and

b) the proven way to get elected was to convince people that eating ice cream doesn't make you fat, and even if it does have adverse health effects, it will be people you don't know / don't identify with who will get fat instead of you ... which is easy to do because ICE CREAM!

Pajser writes:

People do not use the word "authority" in the meaning "government." Those two are not synonymous. Typical use of the word may be "legitimate or socially approved use of power"(Wikipedia) or "the power to give orders or make decisions, the power or right to direct or control someone or something."(Merriam Webster) Importantly, "a property right is the exclusive authority to determine how a resource is used, whether that resource is owned by government or by individuals." (Alchian, Property rights, The concise encyclopedia of economics, econlib.org)

Using these meanings, authority is more restricted in liberal than in libertarian society. In liberal society, everyone is guaranteed to have freedom to walk, sit, swim, maybe sleep on public open spaces; eat some food in public kitchen and sleep in some shelters; attend to university, use medical care etc. It is guaranteed that authority will not step on his way and liberals want to extend these rights. In libertarian society, authority is not restricted; whenever one steps, it is someones private property, and one is on mercy of owner's authority.


Tom West writes:

Kevin

This is essentially replacing (possibly) well-done social science with social guesstimating, though. I mean, all of the complaints you could levy at the social sciences are actually worse using the method you're advocating.

I'd disagree. The problem is that well-done science is that it only has a few variables. Once you scale up, all the fuzzy factors come into play that often end up dominating or at least heavily influencing the outcome. Better not to even pretend that you can predict accurately.

It becomes like saying you can predict drug effectiveness by using our knowledge of biochemistry. Sure knowledge of biochemistry might help up your predictive abilities a bit, but woe betide the company that assumed it can predict outcomes without looking at what *really* happens when a thousand patients take the drug.

I will agree that observation of others is social guesstimating. Honestly, I think that's the best that we can do.

As for economics, it does pretty well, but the fact that anything much beyond econ 101 (and even some of that) has economists arguing with each other makes it pretty clear that rational people can't come up with a single explanation for what we think is simple phenomena makes it clear that reality is just too complex to model with any degree of certainty.

Dan Hill, that's a nice, elegant counter-argument! Well done.

Handle writes:

That's not how people understand or use the word 'authoritarian', which has a lot more historical baggage and force and negative connotation of tyrannical abuses.

Consequentialist Liberals understand their worldview as being very distant from sanctioning the kinds of extreme coercive policies that constitute their picture of an 'authoritarian' regime.

It's really inappropriate to play the linguistic guilt-by-association game. Well, if you don't have a modest presumption against using government power, then 'authoritarian' and Stalin and Mao. If you have any objection to open borders, then you're a Nazi Nativist Nationalist and Hitler.

But people aren't really like that. The world isn't really like that. It's no use playing word games.

Psmith_in_the_city writes:
a visceral horror at being told what to do.

I resemble that remark.

With all due respect, Bryan, this post is perhaps the most spectacular Ideological Turing Test failure I have ever seen...

Sure is. Most liberals I know tend to claim (and, as far as I can tell, sincerely believe) that the real authoritarian threat is from unchecked corporate power (or words to that effect) rather than from the state, and that they, the liberals, are the true anti-authoritarians.

Tracy W writes:

Tom West:
So, we look at what other places do or have done, decide whether we like the outcome, and base our policy desires on that. General empiricism is as accurate as your ever going to get where human beings are concerned.

Hmm, the social sciences are too liable to be wrong, so we're not going to base our policy analysis on this, instead we'll base it on an even-more-likely-to-be-wrong basis!

It's like "well, medical research is often wrong so I'm going to treat my cancer by just going to a chemists' lab and taking things at random." This is NOT a sensible solution to the inherent messiness of research.

The trouble with "just looking at what other places do or have done" is that other places do an awful lot of things, and working out which caused the desired outcome is NOT easy. Especially since time lags can be significant, eg a law that encourages running down natural resources in favour of immediate production might take decades to show its bad consequences. Allowing people the liberty to experiment often results in bad outcomes at first and only eventually better ones (eg it took a long time for medicine to accept that it was really better to do control groups where some people didn't get the best-thought-of-treatment, because in the long-run that lowered deaths).

We can't get away from the need for theory as well as empiricism and the need for sensible empiricism, not moronic "let's just do what vaguely looks like it might be useful."

Robert Nielsen writes:

Basing your entire argument on calling your opponents ignorant is certainly satisfying for you, but contributes nothing to the debate. I could just as easily write a condenscending post on how conservatives are too ignorant to understand the consequences of their beliefs, but what good would that do?

Chris Wegener` writes:

Robert Nielsen says, very well, what I feel about reading this post.

It is impossible to live without government as long as there are more than three people involved, libertarian beliefs to the contrary.

The illusion that it is possible to organize a country the size of the US without compelling people to follow some rules is just that an illusion.

At a daily local level life is anarchy. In the immediate moment in most places anyone is free to do whatever they want to anyone else present. To believe otherwise is to have never stepped outside your house.

Any group requires a mechanism to control individuals who do not conform to the societal norms that everyone has collectively grown up with in any particular culture. We ignore the homeless mentally ill person raving on the corner. We tolerate the individual who speeds and does not stop for pedestrians. (Though occasionally they may receive a ticket.) We restrain and punish individuals who steal or assault other people as long as they are not married to attacker. We punish, or at least pretend to punish people, particularly corporations, who pollute or otherwise misbehave. We ignore the wholesale purchase of our political system.

Yet individuals, like myself, who believe that everyone should have a sufficient income to live a barely tolerable life are "authoritarian."

Yes I understand labor elasticity. I am sure that the country would be far better off paying a livable wage because we need only look at Europe with its socialist system. Everyone is far better off, There is less crime, less economic suffering and far less social strife.

Only someone who believes in the fantasy that the money they earn is some how morally blessed and untouchable would call a desire to improve everyone's life "authoritarian."

Mark V Anderson writes:

Sorry, Bryan, your post doesn't show anything other than the tautology that someone who doesn't place anti-authoritarianism as the ultimate guide to behavior will sometimes be authoritarian.

I consider myself a consequentialist libertarian. That is I mostly believe in libertarianism because minimal government generally results in the best society. Are there cases where this isn't true? Of course! In those cases, authoritarianism is the best approach. Somehow I don't think I am an authoritarian because of the minimal cases when I do believe in it.

Also, as Philo points out, EVERYONE is authoritarian on occasion. He brings up the issue of a suicidal friend, but I think an even better example would be your children. Do you let them do everything they want to do? If not, your post implies that you are an authoritarian.

Tracy W writes:

I am sure that the country would be far better off paying a livable wage because we need only look at Europe with its socialist system. Everyone is far better off, There is less crime, less economic suffering and far less social strife.

Like in Greece or Spain or Portugal. No economic suffering there! (Also, FYI: Germany doesn't have a minimum wage).

And, while the USA has a very high murder rate, on violent crime more generally the picture is more complex.


Tom West writes:

It's like "well, medical research is often wrong so I'm going to treat my cancer by just going to a chemists' lab and taking things at random." This is NOT a sensible solution to the inherent messiness of research.

Before one would even *think* of prescribing medicine, it requires that it be tested on thousands of people in tightly controlled circumstances. Why? Because what comes out of the labs before testing has a massive failure rate.

If you suggested ditching the test or did a test with only a handful of people, any competent researcher would be horrified at the idea that the product of her research being prescribed. Researchers understand how little they know and can reliably predict.

If the social sciences had a lab of thousands or millions of societies, they could do much the same. Unfortunately, they don't. They are hamstrung by an incredibly complex subject (far worse than medicine as there are even more moving parts) and a very few lab subjects.

Let me try a different analogy. If we have a war and I have a deep knowledge of the rifles used by both sides, I have useful information, but I should in no way be confident in my ability to prosecute the war.

In fact, I'm likely better off looking at other wars and making guesstimates about policy than pretending my knowledge of rifles scales up and the other factors don't matter.

Eliezer Yudkowsky writes:

I have trouble with defining "authoritarian" as "not specifically prejudiced against government intervention" because it makes a chair an authoritarian.

I have trouble with defining "consequentialists" as authoritarians based on the reasoning that consequentialists will do things that involve force if it leads to "better" outcomes. If I have a large penalty term in my utility function for forcing people to do things, and a large positive term for self-reliance and so on, then it seems facile to describe me as authoritarian. Agents like this are not central examples of authoritarians.

Now admittedly, an expected utility maximizer with a penalty term for force and a positive term of self-reliance, who also cares about *other* things, will sometimes accept small losses of self-reliance for large gains somewhere else. If this is not true then you have an absolute deontological prohibition against authority. I have trouble believing that anyone lacking this absolute deontological prohibition ought to be called an authoritarian.

We might think that human beings are systematically prejudiced in overestimating the probable benefits and underestimating the probable costs of regulation, so that those claiming to be consequentialists may often act like authoritarians in practice. I agree. But the correction we need to apply for this factor is not infinite.

Dave writes:
Tom West writes: social sciences are [not] a useful guide to estimating the results of policy.

So, we look at what other places do or have done, decide whether we like the outcome, and base our policy desires on that.

Your second sentence contradicts your first. Why should we believe that scholars who dedicate their lives to understanding such phenomena get it wrong, when the correct answer is so obvious?
General empiricism is as accurate as your ever going to get where human beings are concerned.
Actual empiricism would involve allowing solutions to compete and observing the result, scaling up the successes, not making one-size-fits-all decisions based on fads.
I suspect that what statists like myself *don't* have is a visceral horror at being told what to do. It's why I like to read sites that remind me that a fair number of people have really quite different utility functions than I do.
Is that so you can actually take others' wishes and ideas into account, or so you can pat yourself on the back, while ignoring them?
Peter Gerdes writes:

Umm, you do realize you've given a clearly absurd definition here.

Consider the following consequence. Imagine an omniscient government bureaucrat with no other aim than the maximization of overall human welfare and that in his omniscience this being realizes that in actual fact government action/power is incredibly harmful in virtually every instance though in some apparently random situations they favor it, e.g., they send out a government death squad to kill hitler . Your definition of authoritarian renders this individual authoritarian as they have no presumption against government action (they perfectly happy to encourage it when it good and discourage it when bad) just a huge number of judgements about the outcomes

Indeed, it renders Adam Smith, and many other great proponents of liberty tyrannical because they only happen to oppose government action because they have seen compelling evidence/arguments that it is usually harmful in cases like those at hand.

Essentially you are saying you are authoritarian if you don't have the sadistic view that there's some amount of real human suffering you'll simply ignore before making a fully informed (you know ALL consequences) choice for government action that overall mitigates that amount of suffering.

Steve Roth writes:

Tom West: "I suspect that what statists like myself *don't* have is a visceral horror at being told what to do."

This reminds me of a great line from a friend of mine:

"A conservative would rather be hit by a semi than have a liberal tell him not to play in the freeway."

Comments for this entry have been closed
Return to top