Stupid article. Sure, some of Levitt's work is no giod p, but much it is very good, and much more generally the notion that instrumental variables estimation is all about "cleverness" is silly. See e.g. Imbens' response to much, much more sophisticated and nuanced versions of Scheiber's crap:
http://www.econ.uiuc.edu/~roger/courses/574/readings/Imbens_RespDeaton_Heckman.pdf
Levitt is too stupid to do IV correctly. He cannot do weighted least squares correctly, or fixed effects correctly: something any reg monkey with access to Stata can do. He is a freakonomics huckster and a major reason why University of Chicago's econ dept has gone down the tubes. That is why Chicago is having a very difficult time hiring good people any more: it is no longer regarded as a serious econ dept with all the joksters like Levitt, Derek Neal, Casey Mulligan, John List, etc around that dept
Since Levitt arrived to Chicago the dept has gone to the toilet and student morale has reached new lows.
Explanation: students take math camp and then heavily quantitative courses to end up having trouble getting jobs while a skilled salesman selling irrelevant but cute regressions with no quantitative skills nor training is having great academic success within the same department. That's why so many Chicago students leave academia.
Hilarious comment.
Levitt is the reason econ salaries are so much higher than sociology/poli sci. Before Freakonomics, the gap was much smaller. Afterwards, there was a huge human capital boost to the field in that people who would have gone into medicine, law, other top PhDs chose economics instead and that higher demand drove up prices for professors of economics.
Levitt is the reason econ salaries are so much higher than sociology/poli sci. Before Freakonomics, the gap was much smaller. Afterwards, there was a huge human capital boost to the field in that people who would have gone into medicine, law, other top PhDs chose economics instead and that higher demand drove up prices for professors of economics.
What a stu/pid comment.
Levitt is the reason econ salaries are so much higher than sociology/poli sci. Before Freakonomics, the gap was much smaller. Afterwards, there was a huge human capital boost to the field in that people who would have gone into medicine, law, other top PhDs chose economics instead and that higher demand drove up prices for professors of economics.
Lmao good troll
28 out of 43 economists are low iQ regression donkeys according to our experiment here.
I use a linear probability model instrumented with proximity to a donkey to predict that most commentators have not run a regression, with the angriest commentators predicted to have run as few as negative five regressions. I include fixed effects to control for poster id, which controls for any concerns you are about to raise. My findings imply that there are 212 percent welfare gains if the government shut down EJMR.
Tell someone about some variable, thanks let me think of what a good instrument for this might be.
Of course IV is useful. The question is whether its being used ALL THE GODDAMN TIME. There is no doubt that it is discouraging researchers (some who are brilliant but mentally lazy) from working toward building a real theoretical basis for understanding data.