Efficiency? I Do Not Think That Word Means What You Think It Means

by Neil H. Buchanan

I need to take a short break from following Donald Trump's ongoing attempted coup.  This is turning out to be every bit the existential crisis about which I and others have long warned; and even though people are currently (understandably) expressing confidence that Trump will ultimately fail, I still have serious doubts.  But that is for another day (probably tomorrow).  How about a palate cleanser?

As it happens, I have again been thinking about the incoherence of the concept of economic efficiency.  As Professor Dorf announced back in May, a new Buchanan-Dorf article is forthcoming in Cornell Law Review under the title: A Tale of Two Formalisms: How Law and Economics Mirrors Originalism and Textualism.  Because of the long time frames involved in law review publishing, we are only now being asked to respond to the final round of editorial queries for the article.  (Our submitted draft can still be found on SSRN).
As Professor Dorf explained, and as I explored in two followup columns (here and here), our article describes the state of play in "originalism and textualism" (O&T) and "law and economics" (L&E), with the real oomph of the piece being our observation that conservative academics in those two fields make opportunistic and arbitrary choices that just so happen to lead to right-wing outcomes nearly every time, even though O&T and L&E should not on their own terms be so predictably parallel.  Worse, adherents of both categories of conservative theory loudly purport to be objective and non-ideological, even as they rely on plainly biased assumptions.  (I would say that the theories are "rigged," but Trump and the Republicans have taken that word away from us, at least for now.)

O&T (as the scandalously dishonest recent Supreme Court confirmation hearings in the Senate reminded us) claims that courts can be constrained in their interpretive efforts by being forced to adhere to the words of documents and their objectively determinable meanings, which supposedly means that judges cannot play fast and loose with statutes and constitutional provisions.  Meanwhile, L&E is based on an economic theory that says that the best approach to all legal and policy questions is to determine what the "efficient" choice would be.

Here, I want to return to the economic efficiency concept and explain why the L&E move (which is simply an adaptation of the framework invented by the group of economists known as neoclassicals) is truly an empty exercise within which anyone can embed their policy preferences.  I do so both to respond to an interesting objection that we have heard and because it turns out that even people like Chief Justice John Roberts understand (perhaps without realizing it) that the efficiency concept is incoherent and manipulable.

To be as brief as possible, the neoclassical concept of efficiency (sometimes called "Pareto efficiency" or "economic efficiency") is ultimately a con game,  Because this concept of efficiency supposedly maximizes "net social welfare," it requires aggregating and then subtracting the costs from the benefits of any situation (or possible change in a situation).  If there is an alternative where the difference between benefits and costs is greater, you are not yet efficient.  If you move in a way that increases that difference, you have done something "Pareto-improving," that is, you have increased net social welfare.

What could go wrong?  Well, the answer to any efficiency inquiry depends on which costs and which benefits one measures.  One foundational problem is that neoclassical analysis requires assuming that there are no "interpersonal" effects to measuring benefits -- that is, even if I feel much better about a world with less income inequality, that does not count as a benefit.  Why not?  Because it "makes the math intractable" (an especially ironic claim from people who proudly proclaim that their field is a Science because of all the math that they do).

More broadly, any attempt to measure or compare net social welfare is necessarily going to depend on what we measure and what we ignore.  In one example that Professor Dorf and I explore in our article, an economist could say that Finland's income-based traffic fines are inefficient because the harms from speeding are correlated with things like property damage and personal injuries, which are not correlated with income; but another economist could say that Finland's system is efficient because Finns want to prevent people from buying their way out of the social contract, and income-based fines give everyone a meaningful incentive to obey the law.  In other words, in the latter case "social welfare" includes a shared value of equality, but in the latter case that is ignored.  Neither approach is right in some cosmic sense, because it simply depends on what one says matters to the people in question.

Because of that open-endedness in neoclassical theory (which is imported without modification into L&E analysis), I have argued over the years that liberals should start being as promiscuous in their claims about efficiency as conservatives have always been.  When a conservative says that, say, "the minimum wage is inefficient," I want to stop myself from saying that efficiency is an empty concept and instead say that, no, the minimum wage is efficient and that not having a minimum wage is inefficient.  Under the arbitrary assumptions on which conservatives rely, I am wrong.  Under the equally arbitrary assumptions that I can make, it is they who are wrong.
As another example, consider the claim that housing subsidies are inefficient.  Why are they inefficient?  Because the government is imposing itself into the economic system and changing the costs and benefits to which people would otherwise respond to maximize net social welfare.  But the word "otherwise" is doing all of the work in that claim, because it assumes that there is a non-governmental baseline against which government intervention can be negatively compared.  But what are zoning and every other land use law (to say nothing of redlining) if not government intervention?  And if we try to say that the government cannot be involved at all in setting property rights, then we have reached true incoherence in thinking that we can have property ownership -- a legal concept enforced by government (even if the "government" is a warlord) -- without government.  Setting those baseline assumptions and calling everything else "government intrusion" is simply a circular exercise.

Another way of saying this is that efficiency is meaningful only in the way that a map is meaningful, that is, only when we have metrics and directions that -- while arbitrary -- people agree to use.  If someone asked me where a celestial body is, and I said, "It's just north of Mercury," people would think I was addled.  (The B-52's song "53 Miles West of Venus" plays on exactly this absurdity.)  But in Gainesville, Florida, a civic group set up a Solar Walk in the mid-2000's with obelisks representing the Sun and the planets (still including Pluto, by the way) spaced in proportion to the solar system's distances.  And Gainesville's Westside Park is, in fact, across the street and just north of Mercury.  When people here ask me about locations, I often respond, "So you know the Solar Walk?  Good, then this will be easy.")

Once one has specified the terms, the rest falls into place.  If the terms are open and in dispute, however, then the universe has no up or down, there is no common orientation, and in economics there is no way to say what is efficient or inefficient.

During one of my conference presentations of the Buchanan-Dorf paper, a sympathetic colleague raised an interesting but incorrect objection.  It was interesting because he said that it sounded as though our argument is essentially a post-modern deconstruction of the word efficiency.  That is, if anyone pushes hard enough, it is not at all difficult to make it seem that all language is meaningless.  Efficiency would become meaningless only because all words can be made meaningless.
Not being post-modernists, we would have been surprised (and not at all happy, to be honest) if this were a fair characterization of our argument.  As it happens, however, the objection is incorrect because we are not saying that words ultimately have no meaning.  We are saying something much more specific, which is that the concept that economists use (and it does not matter to us whether economists label this concept "efficiency" or, for that matter, "Fredness") does not specify a coherent and consistent concept.  "I can make arbitrary distinctions that turn my preferred policy into something that I'll label efficient" only works if there are reasons to accept those arbitrary distinctions and to reject all others.

I noted above that even conservative jurists seem to understand the arbitrariness that I am describing.  In an op-ed published by The Washington Post this past summer, Professor Laurence Tribe (while focusing on the Supreme Court's most recent abortion decision) noted that Chief Justice Roberts (and the late Associate Justice Antonin Scalia) disparaged constitutional balancing tests.  Tribe summarized their views:
"To the chief justice, applying a balancing test is a futile and improper exercise for courts, one that masks 'an exercise of judicial will' in the guise of a 'neutral utilitarian calculus.' In Roberts’s view, there is 'no plausible sense in which anyone, let alone this Court, could objectively assign weight to such imponderable values and no meaningful way to compare them if there were.' Echoing the late justice Antonin Scalia, Roberts said that seeking to do so would be akin to 'judging whether a particular line is longer than a particular rock is heavy.'"
But economic efficiency itself falsely purports to be a "neutral utilitarian calculus."  A great deal of effort has been put into claiming that incommensurable values can be consistently measured and thus compared (through moves that are beyond my scope here), but even if those efforts were incorrectly deemed to have succeeded, that does not change the more fundamental problem.  Ultimately, every attempt to say what is and is not efficient depends on setting an arbitrary baseline.  Everything is both efficient and inefficient, not because words have no meaning, but because that specific concept has no non-arbitrary stopping point.

Having written the above, however, I fully expect to see Roberts and the Court's other conservatives lapse into lazy reliance on conservative efficiency claims on a regular basis.  It sounds objective and scientific, so who cares that it is not?