Piketty and the Dangers (Again) of Arguendo Reasoning

-- Posted by Neil H. Buchanan

There has been an explosion of interest in Thomas Piketty's new masterpiece, Capital in the Twenty-First Century.  My most recent post (last Friday, here) was written after I attended an event at a DC think tank, where I witnessed one extraordinary example (out of many, many examples) of the red-baiting that Piketty's book has inspired among American right-wingers.  I also plan to write my annual scholarship review for the Jotwell website about the book, and the discussion among economists and other scholars promises to continue for months, if not years.

Today, due to limited time, I will make only one point about the response thus far to Piketty's book.  In a blog post yesterday (here), Paul Krugman argued against lefty economists who object to Piketty's methodological choice to use orthodox economic tools to explicate his arguments.  In particular, Piketty uses "aggregate production functions" as a starting point, and he makes the further assumption that the owners of Capital and Labor (the two inputs into the aggregate production function) are paid their "marginal products."

Although this is in some ways a mind-numbingly technical point, it is actually quite easy to understand.  The basic idea is that economic activity generates incomes, which are generally categorized into two boxes: wages/salaries or profits.  The standard economic approach acts as if the total amounts of capital and labor can be coherently measured by one number each (even though that is not true), and it then uses a mathematical shortcut that purports to "prove" that labor is paid nothing more nor less than its appropriate total reward, based on its productivity, and the same is true for capital.  That is, there are no "extra returns" for capitalists to expropriate from workers.  The mathematical shortcut, however, cannot help but "prove" that result, because using that form merely assumes what one purports to prove.

Krugman makes two very important points: (1) It is not fatal to the analysis that production functions are a convenient analytical fiction, and (2) One can believe that capital and labor are paid their "just" returns without giving away the inequality debate, because we can still argue that the ways in which people come to own capital are unjust.  That is, we can conditionally allow that GM is not underpaying labor without conceding that the owners of GM deserve to own GM (and thus deserve their share of GM's profits) in the first place.

This is an important and elegant point, but I think that it is an extraordinarily powerful example of the dangers of arguendo reasoning that I have discussed here on Dorf on Law in the past (especially here and here).  That is, I see how one can believe that we could still win the debate over inequality, even after strategically setting aside our very real objections to the two points that Krugman is willing to hold in abeyance.  But holding them in abeyance is not likely to be static or reversible.  The more we concede, or even fail to argue, important points on which the debate could be waged, the more likely it is that those points will be viewed (even by people on the better side of the debate) as having been fully conceded, not just provisionally set aside.

And this is going to be a fierce debate.  Giving any ground, even for clever debating purposes, carries extraordinarily high consequences.  I see Krugman's point, but I think it is naive to imagine that we can hang everything on the idea that the public at large will be open-minded on the question of who should own capital.  The stakes are too high to let everything come down to that one question.