Postscript: A Further Look at ProMarket’s Economics

ProMarket’s new “Addendum to Retraction,” written it appears in response to our recent INET post, doubles down on its critique of our piece which showed that it is feasible for increased output to lead to reduced welfare. The ProMarket addendum is notable for its economic errors.*

ProMarket Economic Error 1: Misunderstanding what is an economic “Equilibrium”

ProMarket: The published version was unclear on whether the theorem contained in the article was a statement about an equilibrium outcome or a mere existence claim.…

An “equilibrium” claim would require the specification of an equilibrium condition, such as “supply equals demand” or “Nash equilibrium.” No such condition appears or is even hinted at in our piece. Therefore, no careful Ph.D. economist should have ever thought that our piece was about “equilibrium.”

More plausibly, some readers might have thought our article was about a social “optimum.” At ProMarket’s behest, we sought to correct that misunderstanding but were blocked by ProMarket from publishing it (an excerpt is now in The Sling). Instead of being about an “optimal” condition, our claim is merely a feasibility claim (namely that it is feasible for increases in output to decrease welfare).[1]

Confusing a “debate about optimality” with a “debate about equilibrium” is a serious mistake. Social “optimum” points are not generally “equilibrium” points: an obvious example is that the social optimum point of a Prisoners’ Dilemma game is not the Nash Equilibrium of that game. Granted, the Second Fundamental Theorem of Welfare Economics does provide conditions under which a Pareto Optimum (which is necessarily a social optimum only for some ethical beliefs, not all ethical beliefs) is a competitive equilibrium, but that theorem has completely unrealistic assumptions, including one—all firms are perfectly competitive—which makes it inappropriate for ProMarket, whose raison d’être is “the exchange of research and ideas regarding the subversion of competition….”

ProMarket Economic Error 2: ProMarket claims Characterizing the Feasible Set is not Relevant in Economics

ProMarket: “For this result to be relevant from an economic point of view—and important for publication in ProMarket— one needs to specify how a point along the PPF is chosen in equilibrium and show that, maintaining this rule constant, the level of utility decreases when the PPF moves to the right.”[2]

Our result is a characterization of the feasible set. Characterizations of feasible sets are certainly “relevant from an economic point of view”—after all, the most famous definition of economics, due to Robbins, is that economics studies the “relationship between ends and scarce means which have alternative uses,” and so understanding the scope of the “scarce means” is a key part of defining any economic problem.

ProMarket Economic Error 3: You Authors Don’t Maximize Utility, You Hacks!

ProMarket: Neoclassical economics assumes agents maximize utility. In working with the authors to revise the article, we were open to considering other behavioral rules as long as the rule was kept the same before and after the rightward shift of the PPF.

In economics, the specification of a feasible set must be independent of behavioral rules such as utility maximization. If it were not, the distinction between means (the feasible set) and ends (utility) would be lost. In keeping with this distinction, characterizing the feasible set, which is what we do in our paper, ought not to involve utility, contrary to ProMarket. ProMarket does not understand that our claims, since they are only about feasibility, are weak—we say output increases may or may not increase welfare—and therefore require little mathematical apparatus. Or perhaps ProMarket simply sees no value in weak claims; but strong claims come with the shortcoming that they require strong assumptions which often fail in the real world, a pitfall not to be underestimated, as illustrated in the paragraph after next.

ProMarket is not consistent about its concern for utility maximization. It applies to us, but not to Prof. Herbert Hovenkamp and ProMarket Advisory Board member Prof. Fiona Scott Morton (“HSM”), who wrote in ProMarket:

We try to limit these strategies with regulation, consumer protection, and product safety or information standards. When those tools do a good job, output returns to its role as a good proxy for consumer welfare. This is the setting antitrust law takes as given…

HSM do not prove that their central claim, the second sentence of this quotation, is the outcome as the equilibrium of a model of utility-maximizing agents. Then why did ProMarket publish their paper?

Furthermore, HSM’s claim, because it is considerably stronger than our claim, raises doubts about its academic rigor which our claim (which ProMarket calls “self-evident”) does not. HSM do not define what it means for “regulation” to “do a good job.” This is a serious omission. If “doing a good job” means regulators should use output as a proxy for welfare, then their statement is not a claim but merely a tautology (“when output is a good proxy for welfare, output is a good proxy for welfare”). But if “doing a good job” means regulators should not use output as a proxy for welfare, then HSM’s claim contradicts itself.

Setting that problem aside, suppose regulators follow HSM’s advice and allow larger and larger firms. Those firms, maximizing their shareholders’ value, will endeavor to set up a self-perpetuating ecosystem of regulation, legislation, and (perhaps well-intentioned) scholarship. Suppose they succeed (due to politicians maximizing campaign contributions, regulators maximizing “revolving door” income, consultants maximizing billing hours, academics maximizing grants, universities maximizing donations, judges maximizing junkets, and news media maximizing viewership). Is that kind of regulatory, legislative, and intellectual environment even consistent with “regulation” doing “a good job”? HSM’s argument requires the answer to be “yes,” but their understanding of where regulation comes from is so (disingenuously?) naïve that they do not exhibit any awareness of the question, let alone answer it using the equilibrium of a model of utility-maximizing agents. HSM’s argument actually needs a model to back it up, because intuition strongly suggests that they are wrong in thinking that “(socially) good regulation” is an equilibrium outcome of “each firm producing more and more output.”

In other words, ProMarket asks HSM for no rigorous utility-maximizing model, though HSM need one to confirm their claims; and ProMarket asks us for a rigorous utility-maximizing model, though we do not need one to confirm our claims. By Occam’s Razor, we ought not provide one.[3]

ProMarket Error 4: ProMarket Behaves Inconsistently in Its Editorial Decisions

ProMarket: “Most importantly, the article only raised a technical possibility, not an economic one, and thus it was self-evident and uninteresting from an economic point of view.”

We are pleased it was “self-evident” to ProMarket—once we were the first to show it. Particularly our second graph, showing feasibility of both agents being worse off with an Edgeworth Box expansion, is not completely trivial to conceptualize.

Our “economic” point is that an increase in economic output can be wasted, by being mostly directed to producing things of little social value (private airplanes?), with little additional production of things of great social value (homeless services?). In other words, our “economic” point is that output fetishism is an economic error. If this is “self-evident,” then why is most macroeconomic and growth theory and practice single-mindedly devoted to maximizing output?[4] In our forcibly and wrongly retracted paper, we discuss the fetishism of GNP.

Output fetishism is an extremely easy error to fall into in one-output frameworks such as the ubiquitous “Consumer Welfare Standard” graphs of consumer surplus. Our graphs are much better than those graphs because, although our graphs are simple, they are not that simple. By using a two-output framework, our graphs demonstrate possibilities that authors using one-output frameworks can never give consideration to. Witness HSM:

A useful definition of “consumer welfare” is that antitrust should be driven by concerns for trading partners, including intermediate and final purchasers, and also sellers, including sellers of their labor. These all benefit from high output, high quality, competitive prices, and unrestrained innovation.

If this market is for toasters,[5] then it is true that the toaster manufacturer and its trading partners may benefit from increased production of toasters; but humankind does not live by toasters alone (we do not live in a one-output world), and why would an increased production of toasters not be accompanied by a fall in the production of something else?

Our paper merely points out the range of feasible outcomes and hence need not specify behavior, but HSM’s paper requires the unambiguous (though unstated) condition that a rise in “toaster” production definitely does not require a fall in the production of another output. Where is HSM’s proof that that “free lunch” holds in the equilibrium of a model of utility-maximizing behavior? What assumptions underlie that proof? (Increasing returns to scale, i.e. “natural monopolies”?) How realistic are those assumptions?

Alternatively, instead of a claim that production is happening strictly inside the production possibility set, perhaps HSM is thinking that a rise in toaster production would cause a fall in the production of another output but that such a state of affairs would be good. Why? It is contradictory to claim that when studying the “toaster” market, “more toasters will be better” even if that means “wok” production must fall, and to simultaneously claim that, when studying the wok market, “more woks will be better” even if toaster production must fall. Those are inconsistent preferences. HSM really need to write down a model with more than one output to clarify for the rest of us, and maybe for themselves, what kind of world they have in mind. Of course, ProMarket makes no such demand of them, only of us.


ProMarket’s decision to apply inappropriately weak editorial standards to HSM’s piece arguing that “the test for mergers should be whether a merger will lead to reduced output,” and inappropriately strong editorial standards to our piece arguing the contrary, reveals that its aim “to bring the rigor of academic discussion […and…] follow the highest standards for its publications” depends very much on who is making the argument and in what direction. But ProMarket is not going to end up rescuing output fetishism. Nor will the arguments of the other devotees of this cult.[6] Welfare economists, who work in the subspeciality of economics that address these issues, are in essentially unanimous agreement that output is not a proxy for welfare—not only because they know how to use multiple-output models and take seriously the idea of no free lunches, but also because they take into account rigorous empirical studies that show what things actually enhance the lives of real people.


[*] The original language of the retraction was the result of a negotiated settlement between us and ProMarket. In publishing the Addendum, ProMarket has violated the spirit of that settlement.

[1] As for the way we drew the initial condition, as we show in The Sling, our claim has nothing to do with whether the initial condition is optimal or not. This is because the (to us irrelevant) optimality of the initial position is determined by the juxtaposition of the initial indifference curve and the initial output production possibility frontier (“PPF”) in our first graph, and the juxtaposition of the initial indifference curves in the initial Edgeworth Box in our second graph; whereas our claim only involves the juxtaposition of the initial indifference curve and the final PPF in our first graph, and the juxtaposition of the initial indifference curves in the final Edgeworth Box in our second graph.

[2] ProMarket again incorrectly makes the equilibrium argument here.

[3] Also, unlike ProMarket, when we criticize the writings of other authors, we do not hide behind anonymity.

[4] See Onur Özgöde (noting that growth is the dominant policy paradigm).

[5] The toaster example is from Hovenkamp, “Antitrust’s Borderline,”, p. 4.

[6] John Newman points out on pages 577–581 of his article that there are very many people supporting the idea that increasing output increases welfare: Phillip Areeda, the ABA Section of Antitrust Law, Frank Easterbrook, Mark Janis, Mark Lemley, Christopher Leslie, Michael Carrier, Justices Gorsuch and Kennedy and Thomas, and “Professors Boliek, Cooper, Epstein, Haber, Hazlett, Hurwitz, Lambert, Lipsky, Manne, Semeraro, Teece, Wright, Yoo, and Yun.” See also Cristina Caffarra (“The standard response of the antitrust church to this debate is that protecting low prices and greater output is always good….”). Some of those individuals have published in ProMarket, either about that topic or other matters.

Share your perspective