On Chicago Economics and Samuelson's Technocratic Conception of Economics

Samuelson's exposition is consistently mathematical, and presupposes a knowledge of advanced calculus, higher algebra, and differential equations...I consider this failure to provide translations for the "literary" economist a serious shortcoming of his work. He dismisses translations into words as "mental gymnastics of a peculiarly depraved type." I disagree. There is no depravity, nor is there virtue, in telling other competent economists things in a language they all can understand-there is simply responsibility to the canons of scholarship.--G.J. Stigler (1948) reviewing Samuelson. [HT David M. Levy]

Even extremely well informed readers do not always recognize the significance of the books they review; or if they do, they may wish to suppress it in the hope to prevent a certain outcome. George J. Stigler, one of the leading Chicago-Economists of the 20th century (Nobel 1982), insists at the start of  his 1948 review of Paul Samuelson's Foundations of Economic Analysis that "much of [the] subject matter is currently of much diminished interest." Samuelson's (1947) Foundations is the book that (together with Samuelson's textbooks that followed) brought together the formal languages and techniques that created the mathematical 'revolution'* and its accompanying technocratic ideal which has been its self-image of economics since. (This is not to deny later contributions by Arrow, Debreu, McKenzie, etc.) After Samuelson (Nobel 1970) -- by his own lights "the last 'generalist' in economics," -- economics became an esoteric field for specialists and specializations. 

Stigler's review was not very influentual (most citations to it are by recent scholars) and is less known than, say, Boulding's [see this essay by Khan]. Even so, Stigler review is instructive because it articulates a number of themes that become distinctive of 'Chicago' (and, thus, a crucial feature of the intellectual edifice what is often known as 'neo-Liberalism.') Before I elaborate on that, Stigler's review is at its best when it explains Samuelson's methodological strategy. One often hears that mathematical economics is not empirical (as well as not realistic). But as Stigler notes (correctly), "The leitmotiv of Samuelson's treatise is the search for meaningful theorems, that is, theorems which are in principle capable of empirical contradiction" (either introspectively** or by way of social scientific data). In the context of commenting on Samuelson's empirical methodology, Stigler asks, 

what if the theorem is contradicted by observation? Samuelson says it would not matter much in the case of utility theory (p. 117); I would say that it would not make the slightest difference. For there is a free variable in his system: the tastes of consumers. These tastes are not observationally (or operationally) defined, so any contradiction of a theorem derived from utility theory can always be attributed to a change of tastes, rather than to an error in the postulates or logic of the theory.

 

Here Stigler puts his finger on a key issue. One might think this is a devestating objection because it allows economists to explain away any inconvenient, empirical data. And, indeed, appealing to changing tastes explains little (when saying that an agent maximizes some utility (function) all other things being equal and confronted with recalciltrant data). This is the core criticism explored in a paper by Robert Michael and Gary Becker (Nobel 1992), which forms the basis of a reformulation of decision-making: "if behavioral responses are attributed to differences in tastes, not much more can be said since there is no useful theory of the formation of tastes. If, however, they are attributed to differences in production processes, these in turn imply differences in prices and income, and some guidance about these effects can be obtained." As Stigler put it in writing with Gary Becker in a very famous paper, three decades after his review of Samuelson: "the economist continues to search for differences in prices or incomes to explain any differences or changes in behavior." Once (stable) values can be assumed, one can treat agents -- as Milton Friedman had insisted in his famous 1953 methodology paper -- as utility maximizers. [Friedman and Stigler were already close friends; they had been colleagues at Minnesota in 45/46, and corresponding regularly in this period.]

 

Now, as I have discussed before, Stigler and Samuelson had a very polemical exchange over the so-called new welfare economics in 1943. Although it is widely assumed that Stigler 'lost' it, Stigler has not forgotten; he alludes to it in his review:

I persist in my belief that this is one of the less fertile areas that modern economists till; it is symptomatic that we have elaborate instructions on how to form welfare judgments that do not depend on value judgments, but we have no illustration of the application of this technique to a real problem of contemporary policy. Samuelson indeed offers much support for this skeptical view, by this enumeration of the assumptions of the new welfare economics (pp. 222ff.), most of which are held to be partly invalid. But he fails to examine other facets of the problem, one of which, I think, is especially significant. When economists are writing freely on desirable policy, that is, when they are not writing on methodology, the disputes are almost always over how the economic system works, and not over the goals that should be sought

Stigler here anticipates a central commitment of the technocratic conception -- consensus over fundamental values -- made by Friedman in his methodology essay: "I venture the judgment, however, that currently in the Western world, and especially in the United States, differences about economic policy among disinterested citizens derive predominant from different predictions about the economic consequences of taking action - differences that in principle can be eliminated by the progress of positive economics." (Friedman (1953) "The Methodology of Positive economics; it has a notorious afterlife in Friedman's 1976 Nobel acceptance speech (recall).

Note, however, that in his review of Samuelson, Stigler assumes consensus over values among the experts ("economists…writing freely on desirable policy") not necessarily the public. In fact, in the concluding paragraph of the review (quoted at the top of this post), Stigler advocates that Samuelson makes his work intelligible to fellow experts, other economists, not the public at large. This is a matter of "responsibility" to 'scholarly canons' (not society). So, economists have duties to the guild or, perhaps, the ethics of inquiry, not necessarily to society. (If economists are also treated as ordinary agents, then one can say about them what Stigler notes about "particular entrepreneurs;" they are "in the industry because it is the most profitable place to be.") Of course, once economists are trained in or recruited from the ranks that understand "advanced calculus, higher algebra, and differential equations," and it is assumed that they agree on fundamental values (which they are taught is not part of their subject matter), then the door is opened to economists-as-social-engineers.+

 

*The language of 'revolution' was not imposed post-facto by Kuhnian inspired historiographers, but as Stigler explained to Kuhn in the early 1960s, a way economists understood their own development (from Adam Smith onward--see this post on Menger/J.N. Keynes).

**The defense of the relevance of introspection is Stigler's not Samuelson's.

+That is to say, as early as 1948, Stigler had already made key moves away from the views of his teacher, Frank Knight, who resisted these trends.

Comments

[Alan Nelson]

Interesting stuff! Some random comments:

It is easy to imagine analogous complaints about the use of modern logical techniques in the early 20th century. ["Why can't Russell explain the difference between internal/external negation without all that technical apparatus?"] And the natural reply is the same one Samuelson must have made to Stigler--Why not learn some logic/math?
And that is exactly what happened. Virtually every grad program in Phil has an intro logic requirement and in Econ most grad programs have some math techniques courses for first year students.

The Becker "reformulation" of consumer theory is clever and interesting. But isn't the stated motivation quite wrong? Just as "tastes" can be adjusted to cover unexpected outcomes, internal utility production functions (or whatever) can also be adjusted post hoc. (Or so I suggested in an old paper).

[Eric Schliesser]

Yes, by the time Stigler reviewed the book that process was already in full swing. Today I happen to be reading "How Foundations came to be" by Samuelson, Journal of Economic Literature, 1998. It's autobiographical so can't be fully trusted, but he recounts the people and geographic distribution involved in early reading groups on Foundations. (The early and quick uptake is really astounding.) These, in turn, started to demand a curriculum that could prepare students.

Yes, you were right in that paper, Alan. I keep hoping that I can tempt yo 'back' into philosophy of economics!