The Lowest Deep on Hoxby-Rothstein
I don't have a dog in this fight. I'm not qualified to judge this fight:
The Lowest Deep: Does Competition Among Economists Benefit Junior Faculty?: There is a ridiculous battle in ivory tower economics: Jesse Rothstein, assistant professor at Princeton, wrote two comments over the last couple years criticising Caroline Hoxby's well-known and influential paper 'Does Competition Among Public Schools Benefit Students and Taxpayers?' (AER, 2000; NBER WP Version), suggesting that her results are overstated.... Hoxby finally had enough and posted a vitriolic and bruising response on the NBER website.... Hoxby found that students in cities with many competing school districts perform better on standardized tests than those where there were fewer school districts to choose from; competition improves outcomes. Rothstein, as part of a grad student assignment, attempted to replicate the results but couldn't and asked Hoxby for help. What happened next is fuzzy.... Rothstein embarked on a multi-year mission to disprove Hoxby's results....
My personal opinion is that 90 percent of Rothstein's points are technically correct but irrelevant in practice; they just don't affect the results much. [Who cares if Hoxby is using the 'cluster' command instead of Moulton's standard errors?] On the more substantive points the results are initially startling but hinge on judgment calls where I think Hoxby has it right (e.g. the 1st stage should be run at the MSA level, not the individual level). After reading both, it's hard for me not to conclude that Rothstein's 'preferred' estimates are 'preferred' largely because they differ from Hoxby's and to wonder why the editors of the AER thought it fit to crowd their scarce pages with their back-and-forth.
While the merits of the debate are of some academic interest, the papers are worth checking out just because they're so antagonistic. It's like a nerdy Celebrity Death Match....
I do remember Richard Freeman laying down the three laws for econometrics you can trust:
- It had better be there in the ordinary-least-squares regression.
- It had better still be there in the econometrically-sophisticated high-tech instrument procedures.
- It had better still be there for small technical tweaks to the econometrically-sophisticated procedures.
Rothstein makes a convincing case that Hoxby doesn't satisfy (3), if his definition of "small tweaks" is correct--and I'm not qualified to judge whether he is right. And the connection between voting-with-your-feet within metropolitan areas and average school quality was never there in the OLS at all.
On the other hand, Caroline Hoxby has, in my view, the advantage of theory on her side: it would be very surprising if competition for students did not have an effect on school quality, and if school quality did not have some effect on outcomes.
The public middle school my children go to--Stanley School, in Lafayette, California--is truly excellent (even if massively underfunded by any objective standard, as all California public schools these days are). It's principal, Fred Brill, is extremely highly regarded. Nevertheless, if there had been another equivalent school that we and other parents could have voted-with-our-feet to send our children to, there is no way in the world that Fred Brill would have dared put the teacher he did in charge of my son's sixth-grade core class. No way at all.
