First-best v Second-best

Wow, this is great: Vulgar Second Best Theory, bonanza edition.

First we have Dani Rodrik claiming that simple supply and demand thinking doesn't always work (granted), so state interventions are justified. Rodrik doesn't bother to justify this: it's self-evident, to him and his fellows, because they see imperfections and have many models from which to choose. The rest of us just aren't very sophisticated, I suppose. Then we have Tyler Cowen asking (worth reading) how the state interventions got promoted to second best and asks who the real utopian is. Then Rodrik responding that since "almost no-one questions whether [health, education, social insurance, macroeconomic stabilization] properly belong in the government's arsenal", then the state's role in any other activity industrial policy is justified. Alex Tabarrok rightly calls him on defending this idea on the basis that the heart of his argument is that the intervention in both cases is "targeted on a loosely-defined set of market imperfections that are rarely observed directly, implemented by bureaucrats who have little capacity to identify where the imperfections are or how large they may be, and overseen by politicians who are prone to corruption and rent-seeking by powerful groups and lobbies."

Again, let's review the basics of Vulgar Second Best Theory:

1) Identify market failure (any will do, no need to quantify, ignore counteracting multiple failures, ignore private responses, ignore existing state interventions' effect on "market")
2) Identify intervention
3) Declare victory (ignore unintended consequences, ignore government failure, no need to verify outcomes, everything can be fixed with refined intervention)

Once upon a time (August 2006), I had an exchange with David Anderson, a guest poster at Environmental Economics blog. Among the points I felt worth discussing:

• David asserted that petroleum is artificially underpriced with no explanation -- why? How does he know? By how much? That would be an interesting post. On the one hand, we know that gasoline taxes have generally exceeded profits in the past 20 years. On the other hand, the justification for those taxes is to build roads. On the other hand, only 1/3 to 2/3 of those road taxes actually make it to the roads. On the other hand (sheesh, 4 hands?), we know that the costs of security and the externalities listed as well as other costs are not included in the price. However, the latter would be at least partially offset by the former and some discussion of magnitude would seem to be consistent with alleviating imperfect information. Incidentally, I found that we had spent about $727 B total on the war between 2003 and 2006, and we consumed approximately 146 B gallons of gasoline per year, so (727/3)/146 =~ $0.61 / gallon for Iraq. How much has the price gone up since the start of hostilities?
  
• We learned from an article in The Economist in August 2006 that the world's largest private reserve held by Exxon is only the 14th largest reserve when including state-owned oil companies. Isn't that the opposite of a free market? David asserted the existence of "market power" without any backing discussion -- what does he mean by it? How do you measure it? Why didn't it exist in 1998 when oil prices collapsed? Wouldn't market power imply an artificially high price? Which dominates - the artificially lowness or the artificial scarcity-driven overpricing? And wouldn’t the fact that two of the larger petroleum companies -- Shell and BP -- have substantial investments in alternative energy suggest that any exercise of "market power" to keep solar down indicate a lack of self interest? That would be an interesting claim if he could substantiate it.

I went on ask some questions that needed to be answered before I was ready to concede that market failure and not technical problems or production cost was the reason we couldn't drive cars running on sustainable fuels. Then, David graciously thanked me for the exchange and asked a question similar to the one posed by Rodrik in his industrial policy post:

The crux of our differences may lie in your acceptance of the market failure (your items 2 and 3, which you say are "true") and yet your characterization of the conclusion that markets fail to yield efficient outcomes (your item 6) as controversial.

I responded this way:

• If you want to say, "Market failures exist in this market, therefore this market is inefficient," I can agree with that. If you want to say, "Market failures exist in this market, therefore markets have and will always fail," or "Market failures exist in this market, therefore markets cannot deliver solutions", I think that is a compositional fallacy and an overstatement.
• If you want to say, "I spy a market failure, therefore there can be no free market solution in this market," I am going to have a problem with that, too. First, I doubt whether there are very many markets that don't have at least one type of failure present, so intervention can almost always be justified on the slightest pretext. Unfortunately, most seem to stop their analysis at that point without looking at the inefficiencies introduced by the recommended intervention. The market may not provide a perfect solution, but the best intervention may actually be a third-best solution to a secondary market solution. Radio, which seems to fit the definition of a public good, seems to work pretty well with the inclusion of commercials.
• Second, I think knowledge failures are at least questionable for several reasons. I've read that imperfect information is a cornerstone of Austrian theory and that Vernon Smith's work shows that markets perform amazingly well despite knowledge gaps. Buyers and sellers almost always possess information the other does not. Akerloff would have us believe that a market in used cars can't exist; and yet it does. Markets are also amazingly adept at providing solutions to market failures: Carfax, OU, and CU, for example.
• Third, when a market suffers from multiple failures, they don't add only in one direction; they frequently cancel one another. The state ownership of oil and the corresponding ease with which OPEC should be able to cartelize should raise the price, while externalities imply an artificially low price - which dominates? We know that the cost of our interventions in oil-producing regions is not accounted for in the price, but the risk premium brought about by the unstable regimes and regions that happen to possess the oil and our interventions in them is. Now add Hotelling into the calculus, and figure that the cartel members are going to cheat to drive prices downward, while federal taxes and regulations (not all of which are rational or efficient) raise the price.
• Fourth, "the free market" is not a monolithic organization with a common goal. Toyota is killing GM and Ford with its high efficiency lineup - does anyone truly believe they won't respond? If they don't and fail, are we worse off (less competition) or better off (the better product and organization takes market share)? BlueSun, Piedmont Fuels, Yokayo and other dealers are trying to make a go of biodiesel, and the current high price environment makes them the low cost producer. Incidentally, I can't resist pointing out that you can't buy a new diesel in any state that has adopted CARB, so a grey market in used diesels has sprung up (a market response to a failure caused by intervention!).

Second Best is a simplistic meme that will not die, but needs to. Because its advocates believe their additional analysis is so much more sofisticomated than First Best Analysis, they fail to realize that an even more sophisticated analysis skewers intervention in the same way as their analysis skewers First Best analysis. Their failure leads them to believe that anyone who does not accept their view must be a First Best analyst; it's the left-right, one-dimensional point of view moved from political science into economics, with no more satisfactory results.

Tyler has an interesting approach: he believes that many successful markets have positive externalities and are therefore better than First Best outcomes. In the Rodrikian taxonomy, my approach would be called nth-Best. But I'm not an economist.