Housing Subsidies, Incentives, and the Phantasm of Efficiency Analysis

-- Posted by Neil H. Buchanan

In my Dorf on Law post last Friday, I previewed a debate that was to be held later that day at the ABA Tax Section's Fall meeting in Boston.  There, in a modified high school/college debate format, we discussed the current U.S. tax subsidies that are quite consciously designed to increase the number of people in this country who buy the residences in which they live, rather than renting them.  The Affirmative team argued for a tax code that would have no incentives for home ownership, while the negative argued for a continuation (with possible modifications) of the current system of subsidies.

The specific resolution for debate was that "the United States should make the Internal Revenue Code neutral with respect to home ownership."  Due to an odd turn of fate, I ended up debating for both teams, which means that I was guaranteed to both win and lose the debate.  An audience vote determined the winner, with the Affirmative team getting 6 votes and the Negative team 3 votes, which meant that the majority in the room (of about 25 people) was either unconvinced by either side or had fallen asleep.  (I honestly doubt that the latter was true of anyone.)

The format of the debate included a five-minute "cross examination" period after each constructive speech, which amounted to a short Q&A, putting the speaker briefly in the cross-hairs of the (very supportive and constructively non-confrontational) audience.  Not surprisingly, some of the most productive moments of the debate occurred during those cross-examination periods.  I want to devote the remainder of this post to the first question that I received, after my First Affirmative Constructive speech.  That question highlighted what is, for me, one of the most interesting and ubiquitous questions not only in all of tax law, but in modern policy scholarship in general: What role should the concept of economic efficiency play in the design of public policies?

At the beginning of my speech, I (like any good First Affirmative debater) defined the key term in the resolution, which in this case was the word "neutral."  I defined a tax code as being neutral when it, "neither by commission nor omission, tends to give tax-planning-related reasons to alter one's decision" about owning a residence rather than renting it.  I then emphasized that I was NOT using the economic concept of "Pareto efficiency" as our standard of neutrality.

I explained that the reason to define neutrality without reference to Pareto efficiency is that it is not only unnecessary but misleading to do so.  Pareto efficiency, as I have written many times before (e.g., here and here), relies on an unfixed baseline that can be manipulated to make any subjectively preferred policy appear to be objectively efficient (and subjectively unappealing policies to appear, correspondingly, to be objectively inefficient).  Therefore, it ends up being unhelpful -- both analytically, and as a policy matter -- to try to categorize policies as efficient or inefficient.

This is why I use the word "phantasm" as a (clearly disapproving) descriptor of efficiency analysis in the title of today's post.  On a website called thesaurus.com, the entry for phantasm includes the following helpful note: "fantasy  is imagination unrestricted by reality; phantasm  is a ghostly appearing figure or something existing in perception only."  Existing in perception only, indeed!  Thousands of us have been trained to see efficiency, being able to point on graphs to the supposed embodiment of the decrease in social welfare that exists when the government imposes inefficient policies.  When one actually looks under the sheet, however, the ghost does not exist.  What we want to label an objective fact -- "policy X is inefficient, while policy Y is efficient" -- dissolves into the night when we begin to think clearly about what underlies such a claim.

The Affirmative side's position, therefore was that Congress should not enact policies (or to refuse to enact other policies) that would tend to cause people to buy residences, rather than renting them.  Our arguments for that proposition, which I provocatively labeled "Favoring Housing Harms People," were entirely familiar to anyone who has followed the home ownership debate.  The "tax expenditures" that subsidize mortgage interest payments and other aspects of home ownership have the effect of redirecting money that the government could have used to promote other types of spending (or to create tax subsidies for other categories of appealing policies, such as expanding the Earned Income Tax Credit for the working poor).  More broadly, policies that favor ownership of residences change the way economic resources are deployed by private parties as well as by governments, causing the average size of houses and condos in this country to increase by something like 50% over the past generation or so.

As I mentioned above, a question during the cross-examination period captured an important point.  Professor Jonathan Forman, who teaches that the University of Oklahoma College of Law, asked me why I had said that I was not relying on a definition of "neutrality" that tracked the notion of Pareto efficiency, given that my arguments sounded very much like the usual Pareto-style complaints about the "waste" that results when the government shifts a demand or supply curve in a way that "distorts" the (otherwise presumptively efficient) market outcome.  Was this not, he asked, a tension in my argument?

The key point is that one can decry as "over-investment" the growth in the size of private residences without relying on a falsely-objective notion like Pareto-efficiency to do so.  Or, to put the point more specifically in this context, it is just as easy to describe a set of assumptions under which the growth in the size of houses was an improvement in efficiency as it is to describe a set of assumptions under which that expansion in the size of houses resulted in economic inefficiency.  By contrast, if one does what the Affirmative team did in this debate, which is simply to say as a normative matter that the world would be a better place with smaller houses in the U.S. (and, crucially, more resources devoted to normatively superior uses, such as environmental remediation or educational subsidies), then one need not rely on a phantasm to inaccurately describe as objective a subjective statement of priorities.

One might be tempted to say: "Well, you don't need Pareto-efficiency to justify your preferred policy, but I can use Pareto-efficiency to justify that same policy.  So what's the difference?"  (I am not -- repeat NOT -- saying that Professor Forman asked that additional question.  He asked only why my example sounded so much like an efficiency-based argument, which it arguably did.)  The difference is, in fact, quite profound.  Claiming that one is using a Pareto-efficiency standard to objectively condemn housing subsidies relies on a framework that is completely manipulable, whereas relying on a policy judgment that lays out the normative basis of the conclusion makes the terms of the debate clear, and it therefore makes clear why ownership subsidies should (or should not) be repealed.

This distinction caused me to recall a moment in graduate school, during a meeting of the "Political Economy Seminar."  Political Economy was the catch-all term for, roughly speaking, "the small group of marginalized people who reject Pareto-efficiency-based economic analysis."  (Pareto-efficiency-based analysis is known unaffectionately as Neoclassical Economics.)  One graduate student, however, was taking the seminar simply to take every opportunity to argue that the rest of us were deluded and wrong, and that Neoclassical Economics is obviously true and right.

At one point, the professor made a point that was based on an econometric analysis that she had recently published.  The angry grad student scoffed: "Well, that's Neoclassical Economics!"  The professor said that, no, it was simply an empirical conclusion.  The student replied, "Yes, but you used econometrics to reach that conclusion, and econometrics is neoclassical."  This is like saying that Catholics use literary analysis of historical documents to try to prove that Jesus existed, so anyone else who uses literary analysis of historical documents is a Catholic (and, therefore, must believe that Jesus existed).

All of which leads me back to an old, but oft-forgotten point.  Describing how people respond to incentives -- changes in tax rules, and all that -- is not Pareto-efficiency-based analysis.  Believers in the power of Pareto-efficiency analysis talk a lot about incentives, but noting that people respond to incentives does not validate Pareto-efficiency as the basis for sound policy analysis.  Pareto-efficiency-based analyses draw purportedly objective conclusions that, in fact, "exist in perception only."  Policy analysis is enhanced when we ignore phantasms and admit that we are struggling in a world in which objectivity is much to be desired (and applauded in the limited places in which it can be found), but in which subjectivity is always present.