,
30 tweets,
7 min read
Read on Twitter
Historically, econ has thought employers as more rational than women and minorities. It translated in how economists define and measure discrimination.
Here is a short thread on my PhD thesis bc it’s Sunday and it’s raining 1/28
Here is a short thread on my PhD thesis bc it’s Sunday and it’s raining 1/28
My PhD was on the history of a specific (now standard) approach to discrimination (*the economics of*) I studied the origin and reception of theories and models, as well as empirical analyses and their uses outside academia, from the 1950s until the 2000s, in the USA. 2/28
While you can conceptually separate approaches that seek to explain discriminations, and those that wish to measure and produce evidence, it’s intertwined in the production and reception of the analyses - here Becker and Arrow's models 3/28
Becker produced an analysis of the effect of competition on racial discrimination, his « taste-based Trade model » is not an explanation of the origin of discrimination. His "influence" on the field can better be captured by the reception of his later work with Mincer 4/28
The *idea* of statistical discrimination (in econ) goes back to the 19th century, but was formalized by Arrow when he was asked to reflect on what kind of data was needed to measure discrimination properly (at RAND, part of the story here: bit.ly/2vAWP8c) 5/28
To emerge as *the standard* (and not the first) econ approach to discrimination, it meant that what was done before was rejected as outside economics & that a separate *economic* way to look at discriminations, compared to other social sciences, be accepted #boundarywork 6/28
This "separate way" can be sum up by the type of question econ were asking themselves: « What if discrimination is efficient? » and is determined by how labor economics get redefined after WWII by Chicago’s price theory diffusion (among other things) 7/28
Second part is on the debate between direct (a range of experiments) and indirect measures (essentially Oaxaca-Blinder and derivatives methods) of discriminations (of course hist of these methods is not specific to discrimination, but to applied micro since the 1970s) 6/28
Wage decomposition methods operationalized human capital theory in a particular way and look at the supply side of the market. It imperfectly measures how personal characteristics get valued on/by the labor market, it does not say *how* employers set wages 7/28
This aspect is especially important in the rhetoric developed by economists who did experiments on discriminations and reject this indirect way to look at behaviors (historically studying discrimination has been first to study the "discriminated against", not only in econ) 8/28
Lab and field experiments on discrimination were first done outside econ, and get absorbed in it with modifications of major settings. Audit studies started in the 1970s in the « activist tradition » (bit.ly/300GSWG) and lab experiments in the 1980s 9/28
Here again, the question of the moving frontier of the disciplines is central, it was also the time (1970s and 1980s) when some econ proudly call themselves "imperialist" 10/28 bit.ly/2gb2P0e
In the 1980s, when Judy Rich send her paper using audit studies to an economic journal, she was told this was not economics. In the 2000s, Bertrand & Mullainathan get published in the AER. The difference? randomization, and the credibility revolution happened. 11/28
Third part of my phd was on the public discourse of the economics of discrimination and how it was contested within and outside the field 13/28
A chapter is on Friedman’s use Becker’s model to go as far as writing anti-discrimination laws (e.g. the 1964 Act) are like the Nuremberg laws. I also draw a comparison with debates around Hanna Arendt's position on segregation and forced desegregation 14/28
Then, I look at how economics try to empirically answer « Are markets/laws efficient to cure discriminations? » in the 1970s and 1980s, ending with Heckman et al. paper on the effect of the Civil Rights Act (positive effect). 15/28
And finally how judges confronted the behavioral hypothesis made by economists that employer took rational decisions. (some elements on that history here: cleocz.com/2019/01/16/som… 16/28
This last part was a way to question three figures of the economist in the history of expertise: the public intellectual (Friedman), the policy evaluation guy, and the expert witness in courts. 17/28
So results 1) path dependency to tools that encapsulate a (narrow) way to observe and define discrimination. Economists want to isolate discrimination, and because of their tools, get caught in experimentator’s regress (always something else could explain discrimination) 18/28
2) you can’t separate measurement issues from theoretical questions, you can’t separate how a theory/method became prominent in academic field, from the public discourse it generates, and social demand it get inserted in #internalexternalhistoryisbullshit 19/28
3) Large part of the debates were driven from "outside" (social demand for policy matters) but large discussion happen in isolation (from other social sciences, and other parts of the field, and, of course, in isolation of radical and feminist econ) 20/28
4) the place of study of discrimination says a lot on the trajectories of industrial relations and labor econ: it reveals majors changes in the raod to standard labour econ, esp how labor market used to be outside the application of price theory and then at its core 21/28
5) Intersectionality is not (only) a fancy postmodern concept, discriminations has always been thought in relation to other power relationship or different groups 22/28
(not in a totally in a satisfying way if you think of the eugenics/sexist convergence of the interwar period or to the general/generic way to apply a model to any type of discrimination however very different the phenomena and their history differs) 23/28
5) The different theories (from the two standard one to implicit discrimination and other behavioral econ one) are not integrated in what would be a general theory of discrimination, nor the basis of an *economic* theory are very explicit in recent works #awayfromstructure 24/28
6) *What to do if discrimination is efficient?* (with their own def of efficiency) is a specific question asked by econ, and I think a majority still answer: nothing. But I may be wrong and would love to see empirical studies on the beliefs of econ today 25/28
Especially regarding teh difference between labor economists and others economists, and the perception of discrimination as well as occupational segregation. 26/28
To conclude, If I had the possibility to do my PhD all over again, I would concentrate only on 1971-1985 period and include debates with alternatives approaches to discrimination within econ (even if it is not *a dialogue*)(just because I want to write on Bergmann's works) 26/28
And maybe, write it in English ... 27/28
(I wrote this because the recent debates on economists' beliefs on twitter
made me think of the uselessness of PhDs in history of economics)
And now it's hailing (!!!) in Switzerland, I have to find another activity.
End. 28/28
made me think of the uselessness of PhDs in history of economics)
And now it's hailing (!!!) in Switzerland, I have to find another activity.
End. 28/28
