Professor Gutting needs to take a statistics course. Chetty et. al. are well aware that we do not know what is the value of having a great 3rd grade teacher. This is a random variable whose mean and variance may vary across the population. A monkey will still be a monkey even it is taught by a great teacher while my son may gain greatly from having a great teacher. Chetty et. al. have used unique longitudinal data (following students from age 8 until they are young adults and merging in data from IRS tax records) to have an outcome variable to link later life outcomes to the treatment effect of certain teachers. Their statistical model yields an estimate of the average effect on earnings for a certain demographic group (such as white kids) from being exposed to a high quality teacher. In a diverse world, Gutting is correct that this statistical exercise does not recover the full distribution of treatment effects of being exposed to an excellent teacher (the monkey would learn less and earn less later than the average kid exposed to the excellent teacher).
But, the average effect is still an interesting parameter to recover. If parents are risk neutral and don't know their child's type (i.e monkey or median kid or Einstein) then they will value having the information about how the average child responds when exposed to an excellent teacher because their best guess of their child's ability is that the kid is average. This is the Heckman "essential heterogeneity" agenda. Under the "veil of ignorance" (if I can borrow a philosopher's term from Rawls) voters will cast their votes for high taxes to pay for teachers if they believe Chetty's results and are risk neutral and know that they don't know their child's ability. In this sense, Chetty's paper is important in terms of informing public policy. Of course Gutting is right that not all children (think of the monkey) need an excellent 3rd grade teacher (especially if the cost of hiring her is very high), but knowing the average treatment effect is a good start for learning about the entire distribution of returns.
Jim Heckman has written an entire AER paper on this topic. So, Professor Gutting should nudge Dr. Heckman to team up with the Chetty team.
Carneiro, Pedro, James J. Heckman, and Edward J. Vytlacil. 2011. "Estimating Marginal Returns to Education." American Economic Review, 101(6): 2754–81. | |
DOI:10.1257/aer.101.6.2754 | |
Abstract | |
This paper estimates marginal returns to college for individuals induced to enroll in college by different marginal policy changes. The recent instrumental variables literature seeks to estimate this parameter, but in general it does so only under strong assumptions that are tested and found wanting. We show how to utilize economic theory and local instrumental variables estimators to estimate the effect of marginal policy changes. Our empirical analysis shows that returns are higher for individuals with values of unobservables that make them more likely to attend college. We contrast our estimates with IV estimates of the return to schooling. (JEL I23, J24, J31) |
In this case the marginal policy change would be to increase the supply of excellent teachers and have more schools hire these people. As these people join different schools, at the margin, would a future Chetty research team recover the same average treatment effect? Or would the marginal school who hires such a teacher have a lower treatment effect ? Where treatment effect is the increase in lifetime earnings from having a better 3rd grade teacher teaching at a specific school