Thursday, June 06, 2019

Everything Is Both Efficient and Inefficient as a Matter of Economics

by Neil H. Buchanan

In policy debates, and especially in tax policy debates, the ultimate ace in the hole (note that I avoided saying "trump card") is an efficiency argument.  My side favors an efficient policy, but your side foolishly stands for inefficiency!  Everyone wants to be on the right side of that divide -- because no one could affirmatively defend being inefficient!! -- which means that the outcome of the efficiency debate matters greatly to everyone.

That is not to say that everyone is willing to favor efficiency over everything else.  One of the classic questions in policy analysis, after all, is the supposed "equity/efficiency tradeoff," wherein the people who favor a policy because it helps the poor specifically or mitigates inequality more generally say that any inefficiency caused by their policy is more than made up for by the moral value of reducing inequity.

But why does that debate always find its combatants in the same positions, with conservatives sure that their policies are efficient and liberals conceding the point?  The answer is that what the world knows simply as "economic theory" -- as if there could only be one such theory -- is actually a very particular economic theory that is (consciously or otherwise) built to lead to the conclusion that redistributive policies are bad.

I use the normative term "bad" rather than "inefficient" because, even though economics students are taught to say that the theory presented in their textbooks is objective and morally neutral, it takes studied ignorance to believe that being guilty of inefficiency is not presumptively a bad thing.  "It's inefficient" generally means "We shouldn't do it," or at least that we ought to have a very good reason to deviate from the ideal.

But what if the very concept of efficiency is inescapably empty?  What if any policy proposal, or any situation at all, can honestly be described both as efficient and inefficient, depending on the way we frame the analysis?  We need not ask "what if," because it is actually true that efficiency ultimately means nothing.  Understanding why that is true is an interesting journey.

In my Dorf on Law column earlier this week, I described the dominant (or orthodox) economic theory as "undead."  By that I meant that it was possible for true believers of orthodoxy to claim that their theory is flexible enough to defeat all objections by opportunistically adapting their assumptions to cover every situation, miraculously bringing the orthodoxy back to life every time it seems to have been killed.  For example, even when it looks like people are not maximizing their utility, we are told that they are maximizing a utility function that just happens to include an input variable that explains the behavior.

The problem with this infinite flexibility is that it requires people to sneak in ad hoc limitations to prevent the theory from leading to conclusions that they do not like.  "People's utility includes a desire for reduced income inequality," for one pointed example, is perfectly consistent with the mechanics of utility maximization, but somehow it is not permitted when making efficiency assessments.

Consider a few additional examples.  At least through the 1980's (and, so far as I know, continuing through today), the Japanese government heavily subsidized what clearly seemed to be inefficient rice production on land within the Tokyo city limits, justifying the policy as a matter of respecting "culture."  During contested U.S.-Japan trade negotiations at that time, my dissertation advisor commented that he understood that rice was an important part of Japanese culture, but as far as he could tell, auto factories were an important part of the Midwest's culture.  If one is going to excuse supposed inefficiency in subsidizing one, why not the other?

And the point is not that both are inefficient (and thus that both should be abandoned).  It is that both can easily be described as efficient by people who take the relevant preferences seriously.  Similarly, one can argue that the very existence of Central Park in New York City is both inefficient (Think of all the expensive buildings that could be built there!) and efficient (Who would want to live in a city without green space?)

Even more interestingly, I recall reading an op-ed in the 1990's that argued that the U.S. military academies are jaw-droppingly inefficient.  Summarizing an impressive statistical analysis, the author described the costs of training future officers in West Point, Annapolis, and Colorado Springs, compared with the costs of lesser-known officer training programs (like the one depicted in "An Officer and a Gentleman").  He then described standard measures of career success (promotions, longevity, etc.) and found that non-academy officers not only performed better on a per-dollar-invested basis but actually outdid the academy graduates in an absolute sense.

In response, an academy graduate wrote a harrumphy letter to the editor complaining that the author of the op-ed simply did not understand all of the hard-to-measure ways in which the academies were oh so terribly important to the military and the future of humanity.  Like rice fields in Tokyo and auto glass plants in Toledo, the orthodox theory could be flipped by deciding which other factors ought to matter in assessing efficiency and inefficiency.

And these are hardly special exceptions.  In my column, I claimed that orthodox economics inherently and relentlessly leads to an anti-government stance by economists.  Because the Invisible Hand is said to guide us reliably to the best of all possible outcomes in this best of all possible worlds, the visible hand of government can only make things worse.  This ignores the simple fact that no economy can exist without government(s) setting and enforcing contestable rules, but why let that get in the way of a good libertarian rant?

Some commenters responded to that column by asking about the post-WWII consensus that government should at least intervene in the economy enough to control inflation (and maybe even maximize employment and growth) through either monetary policy alone (Milton Friedman) or monetary and fiscal policies together (Paul Samuelson and a cast of thousands).

It is important to note that this consensus was seriously attacked for decades by conservative economists, most of whom ended up claiming that government responses to the Great Recession (the stimulus package, Fed intervention, and so on) were foolish and risked hyperinflation.  Beyond that, however, even the liberal economists who favor such interventions make it clear that they merely tolerate such deviations from the no-government ideal (which, again, is an illogical construct).

More to the point, the macro/micro distinction was very much a part of Samuelson's innovation.  The basic idea was that we should (again, note the normative should) always seek to be efficient in our micro policies even as we apply the most minimal macro interventions possible.  This created an opening for the conservative theoretical purists to say that macro should have "micro foundations," which meant that all macro policies should be based on models that assume utility-maximizing individuals acting in markets that are guided by the Invisible Hand.

Even though the micro/macro distinction has not been successfully erased by the reactionaries, the presumption that we should seek economic efficiency at all times (and deviate from the no-government ideal at most rarely) is deeply embedded in policy analysis on all sides.

As I asserted above, however, there is no such thing as economic efficiency.  Or, to put it differently, an assessment of efficiency or inefficiency is hopelessly contingent on one's assumptions, where no set of assumptions is in any way the natural starting point.

In Tuesday's column, for example, I referred to a presentation at a recent conference by Loyola-LA law professor Ted Seto.  (No paper yet available online.)  Seto's argument is that the standard conservative economic argument that the imposition of taxes creates "tax deadweight loss" ignores something that he dubs "anti-tax deadweight loss."  The idea is that, even if one accepts all of the assumptions of orthodox economics that lead one to say that a tax on a good reduces net social utility, there is a parallel analysis that shows that resistance to taxes causes us not to buy public goods that would increase net social utility.

Put more simply, Seto argues that the orthodox deadweight-loss conclusion is merely a statement that taxes come with costs.  As the kids would say (and as Seto would agree, if not in these words): No duh!  Seto then shows that preventing the "market" for public goods to reach equilibrium imposes its own deadweight loss, using an approach that is as theoretically rigorous as is the standard deadweight-loss analysis.  He then notes that it is at least possible that the anti-tax deadweight loss is greater than the tax deadweight loss -- and, importantly, that orthodox theory has no way of telling us generally which one is greater.

It in no way diminishes Seto's important insights to say that they are a subset of a more general phenomenon, which is that it is always possible to take a claim that something is inefficient and turn it upside down.  His analysis is particularly useful both because of the innovation of the anti-tax deadweight loss idea and because it does not require a more general critique of orthodox theory to reach his subversive conclusion.

But the larger context of orthodox efficiency analysis is inescapably filled with such paradoxes.  When a conservative (or, for that matter, many liberals) disparages a tax or any other policy as inefficient, they do so by taking other policies and economic variables (such as income inequality) as given.  But many of those other policies might themselves be inefficient in exactly the same way that the orthodox economist deplores about the policy on which he is currently focused.

Why does that matter?  In orthodox theory, inefficiency exists when we can show that a policy prevents a market from reaching its equilibrium point, based on supply and demand.  But supply and demand are based on outcomes in other markets.  If, for example, the government currently allows (through non-enforcement) wage theft, then employers have more money and workers have less money.  If wage theft were prevented, then the markets for goods that workers prefer would clear with larger quantities.

Therefore, when someone tells me that a policy will prevent a market from reaching equilibrium, my answer is simple: Why do you care about that hypothetical equilibrium?  After all, if we corrected the other "inefficient" imperfections, this market would clear at a different price and quantity.  (And that is to say nothing about where the initial endowments on which supply and demand are based come from.)

This means, in turn, that the equilibrium that would be reached under current laws and policies is both efficient and inefficient.  It is efficient if we take all of the current laws and policies as right and true, but it is inefficient if even one of those laws or policies were to change.  And every economist of every stripe has a whole list of laws and policies that he wants to change.

This is why, some years ago, I started to argue during academic conferences that liberals like me should stop saying that equity is more important than efficiency or that we should stop talking about efficiency entirely.  Instead, we should simply say at all times that what we favor is efficient and what our opponents favor is inefficient.  That the converse is also true means nothing, because we are never going to win a game that is rigged to show that anti-government solutions are always inefficient.

In other words, one can always say with complete honesty that something is efficient or inefficient, and another person can say the opposite (honestly or otherwise).  Since the word efficiency is so rhetorically powerful, why should you not use it?  You will never be wrong -- or truly right, but so what?  Conservatives have been acting as if they are nothing but right all along, with no basis for doing so other than having convinced people that their assumptions are the right starting point.  We should stop playing by their rules.

1 comment:

  1. As Duncan Kennedy put it back in the 80s, "externalities run wild." I often like to reread his work on the seemingly endless malleability of the efficiency concept and related ideas (and how that exceptional malleability seems to operate on several levels of the analysis at once, so you can't even try to fudge your way to a decent answer by cabining the flexibility for one part of the analysis; everything really does run wild).

    Cost-Benefit Analysis of Entitlement Problems: A Critique, 33 Stanford Law Review 387 (1981).

    ReplyDelete