Yesterday, I asserted that often what we argue about are "incentive problems", rather than "calculation problems", and that pretending they are disagreements over the classic calculation problem just muddies the waters.Greg Mankiw has a post up on Pigovian taxes today that help reinforce this logic. Quoting Holman Jenkins, he writes:
"Even if you believe saving gasoline is a holy cause, subsidizing electric cars simply is not a substitute for politicians finding the courage to jack up gas prices. Think about it this way: You can double the fuel efficiency of any car by putting a second person in it. You can increase its fuel efficiency to infinity by refraining from frivolous trips.
These are the incentives that flow from a higher gas price. Exactly the opposite incentives flow from mandatory investment in higher-mileage vehicles. You paid a lot for a car that costs very little to operate—so why not operate it? Why bother to car pool? Why not drive across town for a jar of mayonnaise?"
Let's start with the second paragraph first. This illustrates perfectly my point that we're not talking about a calculation problem. Mankiw, Jenkins, and I are all assuming that Lange and Lerner were wrong on the calculation problem, and that individuals acting in the market on their decentralized knowledge is the only way to efficiently determine when and when not to drive and buy gas. The concern is to use a gas tax to try to fix distortions in the incentive structure (ie - the property rights regime), but not to calculate, optimize, or determine any sort of solution to the question of consumer choice or firm production. Why not? Because the government is really bad at answering those questions.
I take more issue with his first paragraph. It's a general equilibrium world, Prof. Mankiw. Certain types of energy are underutilized for precisely the same reason that other types of energy are over-utilized, and a subsidy is ultimately just a negative tax rate. So I take major issue with his first paragraph, but the logic of the second paragraph is sound and consistent with what I was saying yesterday.
That's not to say I think subsidizing electric cars is necessarily a good thing. Gasoline fueld cars run on gasoline and electric cars run on coal. From a climate/externality perspective they're not all that different. All the subsidy to electric cars really does is help American energy companies, and maybe positions us well for the day when the power grid is cleaner. But for right now, in terms of negative externalities, one is no better than the other.
.
UPDATE: Jonathan Finegold Catalán blogs about private calculation and public regulation with respect to the oil spill at Economic Thought. I have a thought in the comment section there. I think he raises some good points, but I specifically criticize it from an externality-based and from a welfare-economics perspective. However, I would add here that Jonathan's point can also be criticized from an Austrian perspective. He seems to completely ignore Israel Kirzner's insights about the market as discovery process where the failure of poor choices directs the economy towards equilibrium positions as surely as the success of good choices. Jonathan tries to argue that self-regulation will lead to less oil spills. The market discovery process perspective would adamently disagree. More oil spills would be predicted to occur as entreprenuers discover the rightness or wrongness of their choices and calculations. This may be an market equilibrating process, but it is not a process for guaranteeing a reduction in oil spills, as Jonathan suggests it is.
The incentive and calculation problems are somewhat intertwined. The consequences of a change of institutional rules are often not those that were predicted. Even though proponents recognise that markets better allocate resources, they are trying to guess what the market would do with the new rules. If such consequences were so easy to "calculate," why cannot they also predict markets without the new rule?
ReplyDeletePerhaps they're intertwined. I think there's probably a spectrum.
ReplyDeleteThe easiest "incentive problem" I'm thinking about is an externality. This has absolutely nothing to do with calculation. And perhaps "optimization problem" is a clearer word than "calculation problem", because it's a question of optimization that Mises, Hayek, Lange, and Lerner were concerned with. An externality doesn't present an obstacle to optimization. What it does is distort the institutional infrastructure that the optimization technology is brought to bear on.
"The consequences of a change of institutional rules are often not those that were predicted."
True, but in the purest sense a good policy maker's job isn't to steer the outcome - it's to provide an undistorted institutional framework to work with. Jonathan Catalan, in an earlier post, suggests that the appropriate institutional fix may be to assign property rights to the Gulf of Mexico. I think there are reasons why, while in theory that might make sense, that's not the best way to go. A second best solution might be to raise the costs of screwing up the Gulf of Mexico (which is really why Jonathan wants to introduce property rights anyway - to have producers and consumers internalize all costs and all benefits).
But let's just stick with Jonathan's idealized solution of simply assigning property rights. As long as we remain in the abstract, that works for me. When we do that, we don't do that because there are specific consequences we want to affect. We do that without any expectation of the consequences, because the understanding is that once those institutional imperfections are in place, the market will optimize. In that sense, it is very much the ordoliberal approach of Ropke and Eucken.
Of course, not everything is as cut and dry as adjustments to the property rights regime. Other "incentive problem" solutions" will blur the line between an "incentive problem" and a "calculation problem". I think the point is we still disagree with Lange and Lerner. We don't go into these decisions assuming the government can produce an optimal result - we go into it assuming that they can't, and weigh that against the benefits of adjusting the institutional framework (if there are any benefits).