| Dick Langlois |
I have been wanting for some time to write about an interesting paper by Gilles St. Paul called “Endogenous Indoctrination.” (I wasn’t familiar with his work, but he seems to do interesting things, including this.) Here’s the abstract:
Much of the political economy analysis of reform focuses on the conflict of interest between groups that stand to gain or lose from the competing policy proposals. In reality, there is also a lot of disagreement about the working of the policy: in addition to conflicting interests, conflicting views play an important role. Those views are shaped in part by an educational bureaucracy. It is documented that the beliefs of that bureaucracy differ substantially from those of the broader constituency. I analyse a model where this effect originates in the self-selection of workers in the educational occupation, and is partly reinforced by the insulation of the educational profession from the real economy (an effect which had been discussed by Hayek). The bias makes it harder for the population to learn the true parameters of the economy if these are favourable to the market economy. Two parameters that govern this capacity to learn are social entropy and heritability. Social entropy defines how predictable one’s occupation is as a function of one’s beliefs. Heritability is the weight of the family’s beliefs in the determination of the priors of a new generation. Both heritability and social entropy reduce the bias and makes it easier to learn that the market economy is “good,” under the assumption that it is. Finally I argue that the capacity to learn from experience is itself affected by economic institutions. A society which does not trust markets is more likely to favour labour market rigidities that in turn reduces the exposure of individuals to the market economy, and thus their ability to learn from experience. This in turn reinforces the weight of the educational system in the formation of beliefs, thus validating the initial presumption against the market economy. This sustains an equilibrium where beliefs and institutions reinforce each other in slowing or preventing people from learning the correct underlying parameters.
I was catalyzed to write today because of a related article I recently saw in the Times, which enthuses giddily about a paper called “Why Are Professors Liberal?” by two sociologists called Fosse [N. B. not Foss] and Gross. The Times lauds the paper for its sophistication and use of the quantitative. Here is the abstract:
The political liberalism of professors — an important occupational group and anomaly according to traditional theories of class politics — has long puzzled sociologists. To shed new light on the subject, we review research on professorial politics over the past half-century, identifying the main hypotheses that have been proposed to account for professorial liberalism. Using regression decomposition, we examine hypothesized predictors of the political gap between professors and other Americans using General Social Survey data pooled from 1974-2008. Results indicate that professors are more liberal than other Americans because a higher proportion possess advanced educational credentials, exhibit a disparity between their levels of education and income, identify as Jewish, non-religious, or non-theologically conservative Protestant, and express greater tolerance for controversial ideas. Together, the variables linked to our hypotheses account for about 43 percent of the political gap between professors and other Americans. We conclude by outlining a new theory of professorial politics that integrates these findings, moves beyond existing approaches, and sets an agenda for future research.
What struck me, of course, is how less sophisticated this is than the way economists (including St. Paul above) have thought about it. By this I don’t mean just the econometrics but rather the causality assumptions. In this paper as in most of the literature, the implicit assumption is that there exist exogenous ideological positions and that people sort into these positions according to various dependent variables. The “new theory of professorial politics” in the Fosse and Gross paper turns out to be typecasting: people with certain characteristics tend to sort into professions known to harbor people who possess those characteristics. The professoriate is typecast as left-wing in much the same way that nursing is typecast as a female profession. A clever twist here is that, if this is true, it means than non-leftists should shut up about the bias of the academy, since screaming about it just makes the typecasting worse and makes the academy even more left-wing. (Does that mean that sociologists should shut up about, say, the typecasting of engineering as a male profession?)
The Times article sees the paper as an argument against indoctrination as an explanation for the bias in academe; but of course it’s not, since it is formally agnostic on causality. (The paper’s eagerness to dismiss an indoctrination explanation is amusing, of course, since sociologists are normally quite eager to endorse the indoctrination explanation when the issue is business or advertising.) “Indoctrination” in the sense of the brainwashing of students by teachers is not the only kind of reverse-causal mechanism we could think of. In the spirit of Hayek as invoked by St. Paul above, perhaps we should think of left-wing doctrines not as pre-existing ideologies into which academics self select but rather as the name we give to those ideologies created by the academy. Fosse and Grosse do talk about this idea a bit in surveying the literature, especially “new class” theories. But, again, it’s far from clear how a regression decomposition rules out reverse (or mutual) causality of any kind.
This might be an interesting big-think area to write about — if I had the time or comparative advantage.