Tuesday, August 27, 2013

Positivist Poseurs

-- Posted by Neil H. Buchanan

In my Dorf on Law post this past Friday, I again discussed the open-endedness of the supposedly rigorous and "normatively neutral" economistic approach to analyzing social issues.  This is hardly new ground for me, of course.  I again noted that it is possible to flip any economistic conclusion on its head, simply by changing what one counts among the costs and benefits, especially when one takes into account things that are inherently difficult to measure (but that are conceptually central to a complete analysis).

At the end of the post, I explored an objection to a recent NYT op-ed by Peter Singer.  Singer had claimed to be able to prove, via a cost-benefit analysis, that any moral philanthropist would want to give money to fight trachoma, a blinding eye disease that afflicts people in the poorest countries, rather than giving that money to build a new wing on an art museum.  In a letter to the editor, a reader pointed out that the fabric of a society in which one expects people even to think about such moral concerns is woven from educational and cultural investments that combine to "produce moral people," as I characterized it in my post.

My point was that one can always set up a perfectly plausible cost-benefit analysis, in which the outcome relies on some difficult to measure -- but essential -- variables.  That is not to say, of course, that the more complicated example is right, in any meaningful sense.  In his post yesterday, Professor Dorf offers a fundamental objection: "I'm somewhat skeptical of the particular claim because it is possible to value art but not morals."  That is both right and wrong, in exactly the way that economistic reasoning makes everything right and wrong.

That is, the broader (but still economistic, cost-benefit oriented) approach that the letter writer implied does not require that we put a value on morals directly, but rather that we acknowledge that moral societies do different things than immoral societies do.  We might not be able to measure morals (although I would bet that some economists claim that we can), but we can certainly note that many societies that value art and culture tend also to exhibit more public-spiritedness.  Even before we reach the question of supporting public cultural institutions, such societies tend to be populated by people who would think to give money to prevent another person's blindness (even when that other person is not of the donor's clan), rather than saying, "I've got mine, Jack.  Sucks to be you."

Again, however, Professor Dorf's point is also right, because we should be skeptical of any particular claim that is based on a cost-benefit analysis -- especially one that presumes contestable cause-and-effect relationships.  (Skepticism need not lead to rejecting all claims, of course.)  Professor Dorf mentions the Nazis, who put a strong emphasis on art, but who were quite willing actually to kill innocents, not just to allow them to die of preventable diseases.  Surely, that example would lead even a committed arts lover to doubt that money spent on the display of public art inexorably produces people with kinder hearts.

In other words, neither Professor Dorf nor I would say that the letter writer's point is "true" in the sense that it proves that the money should definitely go to build the new wing on the art museum.  And neither of us would say that it is "false" in the sense that it could not be true under any plausible assumptions about cause and effect.  But Professor Dorf's objection is deeper, as I will describe momentarily.

If I were in the exact position of Singer's philanthropist, I would give the money to fight trachoma in an instant.  Why?  Because, if I want to use a cost-benefit approach, I can easily be convinced that the marginal difference I could make to the trachoma victims is much larger than the marginal contribution that I could make to the world's level of civilized attitudes.  Yet cost-benefit analysis is almost certainly not really what is driving my moral reasoning.

What we end up with, therefore, is another situation in which the economistic approach proves too much, and we are left trying to use other approaches to find acceptable answers to our questions.  Professor Dorf, attacking utilitarianism from what he describes as "the opposite direction" of my attack, says that utilitarianism's real problem is that it is too demanding, rather than insufficiently demanding.  By this, I think he means that utilitarianism, if taken seriously, would require people to take truly radical actions (giving almost all of their money away, as noted below), whereas my description of the problem seems to suggest that utilitarianism is totally open-ended, not requiring a person to do anything that he does not already want to do.

Again, I think this is both right and wrong, in ways that leave me in agreement with Professor Dorf.  The letter in the Times was not, after all, really objecting to the conclusion that Singer draws from his specific, heavily loaded example.  The objection is, as Professor Dorf describes it, that Singer's approach is too demanding.  If we use Singer to justify donating to fight trachoma rather than giving to build a wing on an art museum, how do we not end up concluding that we should never fund any of the things that we believe make society civilized, so long as there are sick people in the world?

Interestingly, however, one of the examples that Professor Dorf mentions is itself subject to the "insufficiently demanding" objection.  A cost-benefit justification for progressive taxation, relied upon by liberals, is that $1 in the hands of a billionaire does almost nothing to increase his happiness, whereas the same $1 in the hands of a starving child can make an enormous difference in that child's life.  Professor Dorf accurately describes it thus: "Given the diminishing marginal utility of wealth, in a wide range of circumstances one can do more good with one's money by giving nearly all of it away than by spending it on oneself and one's family, even once one takes account of transaction costs." Therefore, simple utilitarianism calls for us to redistribute from rich to poor.  Professor Dorf's "too demanding" objection is that this would call for complete egalitarianism, because there is no limiting principle to the marginal redistribution that utilitarianism seems to require.

The problem is that the premise of this example might not be true, as a positive matter.  My former colleague Sarah Lawsky published an article in 2011, in which she looked at the available evidence that could tell us whether the marginal utility of a dollar in the hands of richer people really is lower than its marginal utility in the hands of poorer people.  She found, at the very least, that there are too many anomalies in the statistical evidence to be especially confident about the standard liberal assumption.

Professor Lawsky's point, however, was not that we can therefore simply stop worrying and love regressive taxation.  Instead, she argued that people can (and should) favor progressive taxation, but that they should admit that they do so either because they are making assumptions unsupported by the available evidence, or because they are ultimately not basing their moral decisions on cost-benefit analysis.

Which is how we all end up in the same place.  The economistic approach -- in which self-styled amoral analysts purport simply to measure costs and benefits, reporting to the population at large that some policies will result in a net increase in social utility, while others will decrease it -- is a sham.  Every time someone like Singer comes along and says that there is an objective answer, based on positive (not normative) analysis, he says too much and too little.  It is a pose, and it ultimately relies on something else to reach its conclusions.  What looks like a lack of evidence is actually a lack of clarity.