[This is an email written by Vernon Smith on January 29, 2020, republished with permission. Yesterday's post, prompted some discussion between us.--ES]
Eric:
Here is a 500-word summary of my position; I copy Sabiou,* as we have been working it out together for markets, since July 2019:
Beginning with Cournot, market theorists began to model agents in terms of the outcomes of their actions, turning away from the classical focus, first, on the origins of action, then on outcomes. Cournot wrote D = f(P) for the first time [so did Dupuit, independently, six years later. The time was ripe!] Cournot could not model price formation, as in Smith (1776, ch VII), a text known well to the French followers of Smith. So Cournot got price by summing the quantities offered by sellers, inverting f (P) and computing P. I now understand this move as a well-meaning cop-out from the perspective of classical style thinking and articulating.
This outcome focus describes how I had come (through my graduate education) to think about economics, but all that changed beginning with the first experiment in 1956. Somewhere in my early papers I referred to the Cournot price as emerging from a black box after you fed it D, but I did not yet understand the true significance of that reference: A metaphor for economic theory.
For the utilitarian youth, 1862-1870s, especially in the Jevons-Walras formalization of Ui (x1, x2,...), all action was concerned with the i (individual). Now, because i chooses the xi's to Max U, there is no distinction between origins and outcomes, and the whole economics profession--one paper at a time, each beginning with an immediately preceding paper--cannot even think in terms of origins. All are deeply committed to the proposition:
Action implies outcome, if and only if max U.
The pervading error is in the "only if." The behavioral economists, for example, cannot imagine any form of other-regarding action that is not utility function motivated, so U becomes U (Own payoff, Other payoff) and a whole sub-discipline emerges from the "social preferences" slogan, and they imagine they have rescued economics from its foundation in self-interested action. The BEs do good experiments, however; the problem is with the straight-jacket motivation, in which every newly discovered experimental "effect" is put into U, ex post hoc, and always confirming their program. The psychologists knew and know better, but they do less-interesting experiments, and have infected the sampling population with lies (deception) and coercive (as a condition for a major) participation.
In HUMANOMICS, Bart and I show that the "only if" is dead wrong by simply providing one counter example, Smith’s TMS in which all are strictly “and rightly” self-interested. If there is one counter example there must be many. TMS enlightens the out-sized failure of theorist to predict observed outcomes in the trust (and ultimatum) games of the 1980s-90s
But the same issues apply to WON, where Smith’s analysis was the complete converse of the utilitarian’s: He went from quantities, xj, that the j's wanted and revealed a willingness to pay for, to the collective of buyers and sellers who found prices, as in my first experiment—trivial, anyone can do it.
My position is that we are only returning to that tradition. If that be radical, so be it.
Vernon
CC: Bart Wilson, Sabiou Inoua
[*This is: Sabiou Inoua--ES]
Comments