11 Jul 2008

Pigouvian Libertarians

DS asks "Is Pigovian code for anti-Libertarian?"

Pigouvian taxes on activities with negative externalities are meant to internalize the costs of those externalities so that people choose the "right" level of that activity. As such, they are libertarian -- since they do not restrict action but do reduce the cost of one's action on others. OTOH, people who dislike government and taxes dislike Pigou because he created a new rationale for government intervention. In that sense, these idealist libertarians (government can do no good) dislike Pigouvian taxes...

Bottom Line: Every idea can be abused. (Marx disavowed at least one version of "Marxism"!)


Silas Barta said...

Debates over Pigouvian taxes inevitably become unproductive, as people fail to clairy exactly what their premises are: why are externalities bad? Which externalities warrant Pigouvian taxes, and why? What should people have the legal right to do?

I can do little more but repeat what I said when I set things straight on Pigouvian taxes on my blog.

Bottom line: Pigouvian taxes can be libertarian so long as they are for externalities we believe to be violating justified property rights, and the proceeds are applied to the victims. Look at the substance of the transfer, not the name.

Gene Callahan said...

The problem with Pigovian taxes is that there is no way to get the tax level right except by sheer chance -- with no market, we just don't know what the "price" for the negative externality should be.

Silas Barta said...

Gene: You mean like how there's no way to find out what the "correct" compensation due for torts is, except by chance?

Gene Callahan said...

Silas, the purpose of a tort judgment is to satisfy our sense of justice. It can do that. The purpose of a Pigovian tax is to achieve economic efficiency. It can't do that, because absent a market we don't know where to set it. If you want to turn it into a kind of tort judgment, you are making up your own language again.

David Zetland said...

Gene -- I disagree. A pigouvian tax > 0 can be a step in the right direction. 100% efficiency (as predicted by pencil and paper) is nice but probably unachievable. Even better, a low tax can slowly be raised until it gets closer to optimal. Even if it overshoots, the tax can be reduced. (None of this holds for cap and trade, which relies on the creation of rights...)

Your red herring argument ("we don't know where to set it so we shouldn't set it anywhere") appears designed to avoid all taxes instead of address the problem.

Bob Murphy said...

First, there's no "u" in it; see here. (Don't ask me why, there just isn't.)

Second, the problems are deeper than just incentives and knowledge on the part of bureaucrats. (I.e. how would they know the right amount to tax, and how can we trust them to implement it, even if they did?)

The classic reference is Coase's "The Problem of Social Cost" (pdf).

"Everybody knows" what this Coase paper is about, but actually most people are wrong; even economists who teach it usually get it wrong, because they haven't actually sat down and read the thing. (I realize how condescending I sound, but this actually happened. I had people come in and guest lecture in my history of thought class and they taught the standard version of "Coase thought transactions costs were low enough so that..." which is 100% wrong.)

Anyway, let me give you a quick example to show the kind of problem Coase describes at the end of his paper. Suppose there is a TV factory that is emitting toxins into the river, and there are people downstream getting sick. After consulting with Dr. Zetland, the local government enacts a $25 per set Pigovian tax, which compensates for the cancer that the toxins are causing. At such a high price, the factory shuts down.

But suppose an alternative would have been for the factory to pay the people downstream to move to another location, and that the people would have jumped at the chance to do so. (I.e. the people truly voluntarily accept the offer.) Coase gives an example where the numbers are such that this is the socially optimal (in Pigou's sense) outcome.

But if the government slaps on a Pigovian tax, that outcome won't occur. It's true, you could say, "Well what if the factory pays the people to move, and then asks the legislature to revisit the tax rate, given that there are now no longer people downstream." But even there, you're admitting that the Pigovian tax is screwing things up.

So besides the more specific problem of knowing the correct magnitude, there is also the broader problem that the Pigovian tax assumes that the way to reduce the externality is to declare a particular party at fault and then have it scale back its behavior. In general, that might be wrong.

It's the same principle as the gov't mandating smokestack scrubbers (rather than just taxing pollution and letting market find least-cost avoidance). The Pigovian tax can lead to the same type of inefficiency as other command and control approaches, though it's not as bad as the other kinds.

Bob Murphy said...

Ha ha I just noticed that David linked to a wikipedia entry with his spelling... Well, I'm pretty sure economists spell it "Pigovian." (Yes I know David is an economist too, but he's one of those resource economists or something. I'm talking historians of economic thought, all 13 of us.)

Gene Callahan said...

And by the way, Dave, I really don't appreciate being called a liar, which is what you doing when you accuse me of forwarding a red herring argument -- what I really want is to eliminate all taxes, but I duplicitously argue about efficiency in a stealth attack.

Bob Murphy said...

One other point: Just to make this relevant to what we're all thinking, suppose the gov't enacts a carbon tax that is "efficient" according to Nordhaus. Industries scale back their emissions, humanity survives, and everything is great. There is an opportunity cost of, say, $15 trillion in PDV, but hey we're worth it.

But imagine if nothing had been done by the government, and the growing concerns over climate change caused insurers and the Bill Gates Foundation to do some geo-engineering solution that cost $10 billion in PDV. They won't bother spending this money, of course, if there is a carbon tax that is causing emissions to come way down; no point in spending $10 billion in the private sector for a problem that the gov't has fixed (at a cost of $15 trillion).

So if this happens to be the true situation, the Pigovian tax obviously leads to massive waste.

Now yes, feel free to say engineers are dangerous etc. I'm just pointing out the conceptual dangers of the Pigovian approach. It assumes it knows how the answer will look and then forces it. (Literally, as in, men with guns come to your factory if you don't obey.)

David Zetland said...

1) Thanks for the spelling correction. I'll try to remember for later.
2) Your factory example is not applicable to global warming for the reasons Coase identified, i.e., transactions costs are too high.
3) It seems easier to tax a thousands of carbon wholesalers than pay billions of people. I like a carbon tax because it is simpler.
4) I don't understand how you think the Gates foundation has some super-powers that makes it more effective than governments in solving the GW negotiation.

@Gene -- I didn't call you a liar. At worst, I am saying that your point is irrelevant.

David Zetland said...

ps/re spelling: "Pigovian tax (also spelled Pigouvian tax)" -- so let's avoid that transaction cost and forget about a right way...

Bob Murphy said...


I don't think David was accusing you of lying, I think he was more saying, "Gene, your argument would basically apply to any tax, but we're here talking about Pigovian taxes." I think if we all agreed that there needs to be some taxes, then David's point would be valid. However, since we don't all agree with that...

I'm not talking about 6 billion people paying factories to switch to low-carbon technology. I'm talking about 30 insurance companies and Bill Gates chipping in $10 billion total in order to put some panels in orbit that block incoming sunlight and solve global warming. (Or whatever.)

You say: I don't understand how you think the Gates foundation has some super-powers that makes it more effective than governments in solving the GW negotiation.

Hang on a second. I could do the same thing with your arguments about water. "Why would homeowners have magic powers to conserve water in ways that the gov't couldn't just order them to do during a shortage?"

It's the same thing here, and that's Coase's point. If the gov't follows Nordhaus' advice and implements a $35 (or whatever) per ton tax on carbon, then that short circuits anybody else's attempt to solve the problem. Given that gov'ts around the world are taxing trillions of dollars from everybody on the planet, it's not in any individual's interest to step forward and solve the problem in a cheaper way, because the costs are dispersed.

But if the governments sat back and let people come up with voluntary solutions, they might hit upon a much cheaper way to do it.

So to repeat, I think your confusion here is that you're assuming any point that I derived from Coase must involve tons of negotiations. And no, that's not necessarily true. It doesn't take too much negotiating for a bunch of beachfront insurers to kick in a few billion dollars if it saves them more than that in payments over the next ten years.

(There are free rider problems, I grant you.)

Anyway, my main point here is to show that Gene's points are not the only flaws with Pigovian taxes, from a free market perspective.

David Zetland said...

Bob: Water is not the same as climate because water is a private good and climate is a public good. The transactions costs of water negotiation (or pricing and conservation) are far simpler than those of climate, where costs and benefits do NOT accrue to the same people and do NOT accrue at the same rate. These goods are VERY different and I do NOT suggest the same solution for both.

Let me emphasize -- water problems CAN EASILY be solved by markets/private negotiations. Climate problems cannot -- because they are different goods. This point is basic, important, fundamental, unavoidable (and many other words!)

Bob Murphy said...


Obviously we're not going to convince each other right here on whether gov't should levy Pigovian taxes on carbon emissions.

All I'm saying is that it is theoretically possible that the numbers are such, that a carbon tax would lead to a huge waste of resources. So in case "DS" is reading, I wanted him to see why a libertarian could oppose Pigovian taxes, beyond a simple knee jerk hostility to any tax.

Again, just for the record, in the example I gave above, it wouldn't require huge negotiations, and it would be self-financing (except for Bill Gates' philanthropic contribution).

Obviously you can say the real numbers don't work like that, but that's an empirical question. You're right, maybe they don't work out like that. But if they did, then the carbon tax would be wasteful, even if there really were an externality, and the gov't properly measured it. To enact the carbon tax is to assume that the least-cost solution to the externality is to cut back on emissions.

I'm not saying you're crazy for assuming that this is indeed the answer, but if it isn't, then the gov't will force us to adopt the costlier scheme with the carbon tax. It is the same idea as other command and control solutions, like CAFE standards etc.

Actually, that's a pretty good example. You can imagine someone saying, "I support increased CAFE standards. Under a carbon tax, that's what would happen anyway in the industry, and so I don't object to the gov't mandating what the efficient outcome will be anyway."

If the guy is right, then he's right. But if he's wrong--and the efficient tax wouldn't lead to an increase in mileage--then the CAFE increase is wasteful.

So same thing with your support for a carbon tax. You might be right that the least-cost way to fight climate change is to reduce carbon emissions, but if you happen to be wrong...

David Zetland said...

Bob said: You might be right that the least-cost way to fight climate change is to reduce carbon emissions, but if you happen to be wrong... I think I am seeing your argument now -- let climate change happen (get the benefits?) and then mitigate the impacts. I do NOT agree with this perspective, and I'll post on it.

TokyoTom said...

Bob, if we move generally from taxes on income and capital gains to taxes on consumption/inputs would we be better off? McKitrick and others say so.

As for the setting the level of carbon taxes, you and Gene keep assuming that there is a global government that sets taxes in a vacuum. Instead, we have a multi-player game, where any politically sustainably prices are set at levels that the chief emitters are willing to agree to.

This is analogous to ranchers, lobstermen or shrimpers deciding to close a range or fishery. No single one of them is setting a price.

David Zetland said...

TT has an interesting point about global taxes, and I'm creating a new post to discuss it.

Post a Comment

Note: only a member of this blog may post a comment.