Thursday, September 22, 2005

Second-Best Gas Policy

[Cross-posted on The Agitator.]

Via Don Boudreaux, I find James Surowiecki arguing on NPR for a 50-cent hike in the gasoline tax.

Surowiecki bases his argument on straightforward Pigovian externality theory. Using fossil fuels creates external costs in the form of air pollution and other environmental effects. Since these costs are not included in the price of fuel, drivers tend to drive too much and too often. A tax would internalize the pollution cost, giving drivers an incentive to cut back.

Boudreaux promised to address this issue later, and focused instead on a different flaw in Surowiecki’s analysis. But I’m going to beat him to the punch. The main problem here is that Surowiecki looks only at one potential source of market failure (external costs) while failing to consider another (imperfect competition). According to the economic theory of the second-best, it turns out that when there exist multiple market imperfections, they can actually offset each other – and fixing one can exacerbate the other. Oil and gas are perfect examples. The external cost from pollution means the market price of gas is too low. But the existence of a moderately effective cartel, OPEC, means the market price is too high. These effects at least partially offset each other. It could be, for instance, that the cartel-premium is large enough to effectively internalize the pollution externality. If so, then taxing gasoline would induce drivers to drive too little.

So which of the two effects predominates? I don’t know, and I’ll bet Surowiecki doesn’t either. I do know that 50-cent figure -- on top of the taxes we already pay -- sounds suspiciously like a number he pulled out of his butt.

(I will give Surowiecki credit for this correct point: that if we assume gasoline use should be reduced, taxes are a more efficient means of achieving that goal than mandating fuel economy standards.)


Ben said...

I don't see how your example illustrates the theory of second best. The environmental effect is totally unrelated market wise to the cartel. It is true that the cartel, by its elevation of the gas price, help the environment, but this is a mere coincidence. There is no market force that links the formation of the cartel with the environmental problems. In the big picture, there is as much chance that this kind of coincidental inefficiency hurts the environment as there is chance it helps it. We are still justified at trying to tax the car drivers for an environmental cost that is not random but rather has a clear relation well understood and explained. Isn’t avoiding raising taxes because of the cartel akin to having negative levies and equivalent to subsidizing the Saudis instead of letting free trade govern prices?

Here is how I see the second best theory: Imagine a company has grown so much that it dominates the market so that there is practically no competition and as a result it is almost in a state of cartel or monopoly. Let’s call this company Wall-Mart. This company can thus charge prices that are over marginal costs and make really good profits. Without competitors we can’t say we are in the best economical state. This is a great position for Wall-Mart to be. Wall-Mart however, doesn’t have infinite freedom, because if they make the prices rise too much, it will be worth for someone else to incur the cost of starting a new company and become a competitor. Isn’t this imaginary competitor the second best thing to the real competitor?

Now in your example, you say that since the price of gasoline is already too high. Taxing it would only make things worst. But wouldn’t the price the cartel charge go down a little if this tax would be implemented? They do want to sell their oil. The effect of the tax would be to effectively take money from them and keep it in the country and indirectly remove a burden on the tax-payer.

Higher fuel price solves only half of the environment externalities. Shouldn’t the money from the higher price go towards indemnifying those affected by the environmental problems? E.g. Katrina victims? When the money goes to the government, it has much more chance of helping the victims than when it goes to Saudi kings.

Glen Whitman said...


"The environmental effect is totally unrelated market wise to the cartel." It's related in the sense that matters: both of them affect the price of gas. The environmental effect tends to make the price too low. The cartel effect tends to make it too high. Therefore, these effects at least partially offset each other.

"In the big picture, there is as much chance that this kind of coincidental inefficiency hurts the environment as there is chance it helps it." Untrue. To the extent the cartel drives up prices, there will be less pollution.

Your Wal-Mart story is one potential application of the theory of the second-best, but it's definitely not the only one. My example is another. Here's how they're related. In general, having fewer firms is undesirable because it tends to drive up prices. But having fewer firms can also be good, because larger firms can exploit economies of scale, which tends to drive *down* prices. Which of these effects is larger? Either could be (though in the case of the actual Wal-Mart, the latter effect is clearly the more important).

"But wouldn’t the price the cartel charge go down a little if this tax would be implemented?" No. The price would go up. It's true that some of the tax would be absorbed by the producers, so the price wouldn't go up by the entire amount of the tax. But the net effect would still be positive.

"Shouldn’t the money from the higher price go towards indemnifying those affected by the environmental problems?" That would be nice, but it's not always possible -- at least not without exacerbating other problems. And is there any reason to think the revenues from gas taxes go to environmental victims? The government had already spent all its tax dollars (on a whole slew of pork barrel projects) when Katrina hit, which is why it had to go into even greater debt to provide disaster relief.

Over A. Barrel said...

Of course there are other good reasons to tax gas heavily. People would more rapidly switch to gas efficient, non-polluting vehicles such as hybrids. This sick love affair with SUVS would end except for the rich. Maybe, they too could be shamed into not driving what most others could not, and should not drive.

If the price were much higher, there would be far less congestion on our highways and more people would make use of public transportation and bicycles, mopeds etc. Also, since much of the world oil comes from bad countries like Saudi Arabia, Iran, etc., we would hurt these countries and their miserable, pampered elites financially. I don't think a 200 billion dollar calamity like Katarina will change anything. Maybe when we have multiple natural disasters all at once with a major terrorist attack thrown in for good measure, then the ensuing economic collapse will result in the government being overthrown once and for all and the people, or more likely a dictator will be put in charge. Let's hope she's benevolent.

The oil, which we will run out of soon anyway, could better be used for other purposes like manufacturing plastics or pharmaceuticals rather than being burned up into greenhouse gases and other damaging pollutants. Besides all that, the air would be a lot cleaner, and that would be a nice relief for asthmatics, little kids and dogs too.

You trivialize the matter by discussing a 50 cents/gal. tax. I'll pull $100 a gallon (NOT per barrel) out of my butt. That ought to shake things up more than the Mexico City earthquake ever did. Cannondale stock, anyone?

Ben said...

I still don’t agree. There is still something wrong with your example. I can’t quite put my finger on it. Let me see... OK This is the way I see it: Mr. Society incurs damages to his property when you pollute. If he is smart Society should not let you make this damage unless he is indemnified. Now if some random third party asks too much for one of the tools you use for polluting would you consider that you have paid your debt to Society?

The notion that there is a right amount of gas that should be burnt doesn’t make any sense. There right amount is the amount where we have paid all the parties what they asked for the use of their goods and properties. When you picked that example you either meant

a)that the government should indemnify the population when there is a cartel. This is a completely different issue and it should be dealt with separately. There is no reason for the government to indemnify people for cartels and monopolies. Not even people who thought Microsoft was a monopoly ever suggested we should get a check from the government when buying Windows 98. I don’t see how you could justify it. In my view it would just make things work be giving even more incentive to the cartel.


b) The example you gave was just meant as an example of market failure which could offset the environmental effect in which case there is no a priori evidence that the sum of the market failures other than the environment externality go one way or the other.

A postpriori when you have the information about the environmental externality and about the effect of the cartel, the government can either say I won’t tax gas more because I have decided that I should indemnify the population for the cartel by the same amount I would have raised taxes for the environment externalities.

You can’t just say: I won’t raise taxes because there must be another failure that will cancel this one, because a priori there is none. There is not more chance that there will be a failure that offsets this one than there is chance that there is a failure that will add to the inefficiency.

Over A. Barrel said...

"A postpriori"

I like that word, Ben!

New & Old Terminology :

1) a priori knowledge

2) a postpriori knowledge (your term)

Defn. (mine) - a theoretical deduction backed up with some empirical data, just in case the original theory wasn't airtight.

3) a preposteriori knowledge (created in the same vein as your new word)

Defn. (mine) - Observations have been completed, and you are on the verge of deducing probable causes. (NOT preposterous knowledge!)

4) a posteriori knowledge

This sort of reminds me of the stages of mitosis/meiosis from biology class. It shows that our acquisition of knowledge can be viewed as a sequential, reaffirming process. Nice!

Ben said...

hey let it go, it was late at night.

Untenured Sociologist (R-CA) said...


You're assuming that the only externality from oil consumption is CO2. In fact, there are also substantial political externalities. While much of the world's petroleum is in the hands of generally pleasant people like Canadians, Mexicans, and Norwegians, a good portion of it (including most of the accessible and/or low sulfur stuff) is held by far less savory people like Venezuela, Iran, and Saudi Arabia. The cartel increases the profits of these regimes and thus promotes negative externalities such as FARC, Hezbollah, and Al Qaeda. Sin taxes on petroleum products would reduce the profits of these regimes and compensate for such externalities as destabilizing Columbia and free Iraq.

Glen Whitman said...

Ben -- I think we're talking past each other, because you're talking about fairness or justice, while I'm talking about efficiency. Apparently, your position is that people have a property right to clean air and deserve to get paid for any infringement of it as a matter of justice. I, on the other hand, was talking about giving people incentives that will lead to the value-maximizing level of gasoline use (by which I mean consuming all and only those gallons of gasoline whose marginal benefits exceed their marginal costs to all parties).

I'm not saying your concern with fairness is wrong, but I do think it's tangential to the question at hand. Surowiecki made an economic externality argument, which is *not* generally understood as a claim about fairness, but as a claim about efficiency. The claim is that negative externalities lead to inefficiently high levels of consumption, not that they unfairly transfer wealth from one group to another. The distribution of wealth is an important but distinct question.

Even if we do focus on the fairness issue, you should ask whether a tax will really achieve it. Gas taxes are nominally earmarked for transportion purposes, but the reality is that governments use them as general-purpose slush funds. Even if the funds went only to transportion, that would arguably be a subsidy for the creators of the externality (drivers), not compensation of pollution victims (breathers of air in general, including pedestrians).

"The example you gave was just meant as an example of market failure which could offset the environmental effect in which case there is no a priori evidence that the sum of the market failures other than the environment externality go one way or the other."

That's right, and I admitted that in the original post. I said I didn't know which effects predominated. (It is clear, though, the externality effect and cartel effect point in opposite directions, so they at least partially offset each other.) Surowiecki seemed to think the existence of a negative externality was prima facie evidence that a gas tax -- and a higher one at that -- was justified on economic grounds. My point was that Surowiecki could not claim that degree of certainty, much less justify a number as specific as 50 cents.

Ben said...

I think I somewhat understand what you are saying—that basically your argument was that we have to look at more externalities than just the environment before determining how we should tax gasoline. That’s a valid point. I still have a few comments to add.

First, I’m not talking about fairness or justice. This is not the difference between your argument and mine. I’m talking about the government recuperating future costs. Sure the extra tax money is going to be used as a slush fund, but that doesn’t matter because the environmental cost to the government is not going to come from a specific source either. It’s going to be from reduced productivity, increased number of sick days, invalidity claims, natural disasters etc. The government will take this money from _wherever_ to pay for these things.

Second, I think the main difference between my argument and yours is in the definition of the right amount of consumption. When you say: “(by which I mean consuming all and only those gallons of gasoline whose marginal benefits exceed their marginal costs to all parties)”, you use “all parties” globally. It includes the whole human race. For you, if anyone in the world has higher benefits, in this case the OPEC, to offset environmental costs, you conclude that the benefits exceed the costs _globally_.

I, however, have limited my cost benefit analysis to the effects inside the country. I consider that if the US incurs a cost by the burning of fossil fuel, it is not enough that some other country benefits from it to offset these costs. Perhaps I have wrongly assumed you were doing the same, but I think it was a justifiable assumption because this is the usual way of making tax policies.

Working on the global level is a whole other thing, where ideally a world government would heavily tax OPEC and other oil producers to internalize environmental costs. This would replace the national taxes. There are obvious political and practical hurdles to that solution.

Glen Whitman said...

This debate has probably outlived its usefulness, but I'll respond to a couple of points.

First, with respect to the costs of environmental damage, you've slightly shifted your argument. Earlier, you were talking about individuals as victims of pollution. Now you're talking about the government as a victim, inasmuch as current policies obligate the government to make certain expenditures related to environmental damage. The idea, then, is that the government needs to collect enough revenue to pay for these expenditures. That's a fine argument, but it *is* an argument about fairness, because the necessary revenue could be collected through any kind of tax. Your advocacy of a *gas* tax presumably rests on your sense that drivers ought to bear the burden. This is distinct from an *incentive-based* argument for a gas tax, which is what Surowiecki's externality argument is all about.

Second, it's true that efficiency arguments take everyone (not just American citizens) into account. But when I mentioned the benefits of gallons of gas, I wasn't primarily talking about benefits to OPEC -- I was talking about the value consumers place on driving. From that gross benefit, we need to subtract two kinds of cost: costs of production and external costs (like pollution). OPEC experiences the cost of production, and they receive a portion of the consumer benefits via the price of oil. So here's another way of putting my point: when OPEC successfully drives up prices, it gives drivers an incentive to drive less, which reduces the environmental cost of driving to the general (American) public. Thus, OPEC arguably does a service to American air-breathers at the very same time it does a disservice to American drivers.

Ben said...

Your last comment confuses me a little. I don’t feel we have solved this yet. First I don’t see where I have shifted my argument. I was treating the individual as part of a society under a government. I think my arguments were always meant to be applied at the government level; otherwise I would have suggested something like giving individuals rights to sue oil companies to get remedy for damages from pollution instead of taxes-- which might be another legitimate solution to this problem, given that the legal overhead is not too much of a burden. This is a whole other discussion.

Now, I don’t really see how we can separate incentive and fairness. I just listened to the Surowiecki mp3 and I don’t think his arguments are purely incentive based. The separation is unnatural and it feels weird. Yes I admit it. Your trying to separate incentive and fairness makes me feel weird. I never heard anything like it (albeit I am not an economist) and I don’t see how it would be a useful abstraction. Isn’t the whole free market utilitarian economic theory based on the tight relation between those two? A pure _incentive based_ economy means it’s ok to steal as long as the thief pays the Right Price (TM) to _someone_. This Right Price then has to be fixed (by the government??) because offer and demand doesn’t work anymore when the money goes to a third party. Basically its total economic chaos, everything falls apart.

To the US, the higher price from the cartel is just like any other cost of production. I wholeheartedly agree that the cartel helps the pollution problem, in the same way that high production costs would. But the barrel of oil doesn’t pollute more because its cost of extraction is lower. It does not make more sense to reduce taxes because of the cartel than it does to increase taxes when there are technological advancements that drive the price down (away from the Right Price?) Working on dissolving the cartel and augmenting competition is like working on new and cheaper technology for extraction.

What do you mean by “the value that consumers place on driving”? Gas is not a service provided for free! The value the consumer places on gas is paid for at the pump. I’m sure I don’t have to tell an economist that: It changes with supply and demand and alternatives etc. Integrating a pollution tax in the price of gasoline lets the consumer decide by himself if:

The cost of production and profit margin
+ the cost incurred by the cartel (to the US consumer this *is* a cost)
+ the pollution costs

is worth it or if he’d rather keep his money to buy something else.

But I must be missing something since I am explaining supply and demand and to you :-S

Glen Whitman said...

Well, if you want to continue the discussion, I'm happy to do so.

"I think my arguments were always meant to be applied at the government level; otherwise I would have suggested something like giving individuals rights to sue oil companies to get remedy for damages from pollution instead of taxes..."

Taxes can serve a couple of different purposes. One of them is to raise revenue. The other is to change people's behavior in some way. When someone proposes a tax to solve an externality problem, they are generally talking about the latter goal. That is definitely what economists are talking about, and remember that (a) economists invented the term 'externality', and (b) Surowiecki is a financial columnist who assuredly has some economics training.

I admit that I'm reading a little into his statements, because he didn't talk for that long. But when someone invokes the term 'externality' and proposes a tax to solve it, they are bringing in a lot of economic theory as baggage.

"Your trying to separate incentive and fairness makes me feel weird. I never heard anything like it (albeit I am not an economist) and I don’t see how it would be a useful abstraction." Well, the distinction is quite common in economics. The institutional structures that give people good incentives often do not produce results that people consider fair, and institutional structures designed to produce fair results often yield inefficient incentives.

For example, welfare policies are often motivated by a sense of fairness, but they also can give poor people an incentive to work less, have excessive children, etc. That doesn't mean the welfare policies are necessarily bad, because you might consider fairness (or equity or equality or whatever) important enough to justify some inefficiency. But it is indeed possible to distinguish between fairness and efficiency, and to observe occasional conflicts between them.

"A pure _incentive based_ economy means it’s ok to steal as long as the thief pays the Right Price (TM) to _someone_." Well, no. It turns out that laws against theft can be justified on efficiency grounds. When people don't have secure property rights, it reduces their incentive to engage in productive behavior because they don't get to keep everything they earn. In addition, the possibility of successful theft encourages people to spend scarce resources just transferring wealth from one person to another, rather than engaging in activity that increases wealth overall. (I'm not saying this is the *only* justification for laws against theft, though. You could certainly make arguments based on fairness or justice as well.)

Your point about the difficulty of having gov't establish the "Right Price" for things is exactly why, on efficiency grounds, it's usually undesirable to allow involuntary transfers. When something is tranferred involuntarily, we have no assurance that the value of the item to the taker is greater than its value to whoever took it. Requiring consent and payment, though, mean both parties must find the trade advantageous and therefore the buyer's value must exceed the seller's.

I think some confusion can be cleared up by stating what the "right price" is from an efficiency perspective. In general, the right price is one that's exactly equal to the true marginal cost of the good, including both marginal production costs and marginal external costs. This price is desirable because it means consumers will consume all, *and only*, those units whose added value to the consumers is greater than the added cost to everyone involved.

Economists do not usually think government needs to set this "right price," because market forces will do it on their own. But market failure theories point out instances in which the right price allegedly won't emerge on the market.

The standard theory of negative externalities says they are bad because they cause price to fall *below* the true marginal cost, since buyers and sellers don't take the external component of the cost into account.

The standard theory of cartels (and other forms of imperfect competition) says they are bad because they allow firms to charge prices *above* the true marginal cost.

Both of these are efficiency problems. But note that if both occur, they must to some extent offset each other. The price could even, coincidentally, end up being the efficient one. (I'm not saying that's so here. The price could be too high or too low, depending on which effect is larger. My point was that we, and Surowiecki, don't know.)

"What do you mean by 'the value that consumers place on driving'? Gas is not a service provided for free! The value the consumer places on gas is paid for at the pump." Actually, no. For most consumers, the value is higher than the price they paid at the pump. People buy things when the value to them is greater than or equal to the price. The difference between the two is called the consumer surplus. Only the marginal customer -- the customer for whom the product is just barely worth buying -- has a valuation equal to the price at the pump.

"The cost of production and profit margin
+ the cost incurred by the cartel (to the US consumer this *is* a cost)
+ the pollution costs"

You're double-counting, because the cost incurred by the cartel *is* the profit margin. Now, from an efficiency perspective, the existence of a large profit margin means the price is too high (above marginal cost), and therefore consumers will consume too little. But if there are pollution costs that haven't been incorporated into the price, then the effect of the profit margin isn't so bad after all.

Ben said...

By profit margin I meant the non cartel “normal” profit. I think I’m starting to get what you’re saying. I still think that our argument depends on the stance (global vs national) we take.

Anyways, this discussion has turned into too much procrastination for me. I have to return to my thesis work so I’ll let it go.

Anonymous said...

Applying a "use" tax to any item has never worked. It does temporarily transfer money to the state, which the state immediately pisses away on some war or welfare program. Don't even tell me that you really believe that it will go into some fund for the betterment of future generations. Soc Security is a classic example of that kind of failure.

Anyway, "use" tax doesn't work. People still drink liquor and smoke cigarettes, don't they. And if the tax gets too high, a black market will develop. Remember alcohol prohibition. Kind of like the black market in drugs today under drug prohibition.

Okay, gasoline and alcohol isn't exactly comparable. Or is it? We have people already running deisels on bio-deisel made from used fast food oils. There are substitutes and these will enter the market on their own when there is profit to be made.

Tampering with the cost of any items will only damage the economy and cost us all more. It will not do a thing. I mean look at taxes on a gallon of gas now. Taxes make up about 60% of the retail price of a gallon of gas. Is it stopping us from driving? No. Heck, add $10 more in taxes to a gallon of gas. It will impoversh this country so fast it'll make your head spin. But it won't stop us from driving. We just won't have any money left for luxuries and some necessities. It'll just make us poorer. And the money collected in taxes will be wasted on projects (like building a bride in the Mojave desert) that will be created to provide jobs due to the depression that we will be in.

Just let nature takes its' course. We will deplete our oil, and then we will find alternate energy. It'll happen soon enough. Have patience.