David Leonhardt has written a very helpful post last week on how to think about the impact of above average kindergarten teachers, in which he answers a question I raised on this blog. Drawing on the work of John A. Friedman and his colleagues, Leonhardt writes:

In other words, students who had a teacher at the 75th percentile of ability — based on how much her students learned over the course of the year, as measured by end-of-year tests — could expect to earn a combined $320,000 more over the course of their careers than students who had a teacher at the 25th percentile. Likewise, students with a teacher at the 51st percentile could expect to earn $320,000 more than students with a teacher at the 1st percentile.
So if it were somehow possible to replace all below-average teachers with above-teachers, the economists calculate that each newly hired teacher — as well as all the holdover above-average teachers — would be worth $320,000.

I raised another question in my post:

How much is a standout parent worth? Some children wil generate a great deal of tax revenue over a lifetime. Should we, in light of this fact, give more generous tax subsidies to parents with the characteristics the best predict high future earnings for their offspring? I imagine the answer is no — this would strike most observers as unacceptably illiberal. But improving incentives for parents, as Robert Stein and Ramesh Ponnuru have recommended, seems like a pretty good idea.

This might strike you as crudely utilitarian, and it is. But it is worth thinking through all the same. Greg Mankiw asked a similar question, and he received an interesting response from Raj Chetty, another one of the paper’s co-authors:

I’m writing in reference to your interesting comment about our Kindergarten paper. I think our results are actually consistent with your perfectly sensible intuition that parents should matter more than teachers, for two reasons:
(1) the Kindergarten class effects are large in aggregate but explain a small share of the variance in earnings (less than 5%) overall. A better class leads to higher average earnings (3% higher earnings for a 1 SD improvement in teacher quality), but there is a lot of variation around the mean.
(2) The best evidence I’ve seen on the long term impacts of parents is this quasi-experimental paper by Bruce Sacerdote published in the QJE. It shows that parental characteristics explain about three times more of the variation in adult outcomes than KG classes, consistent with your intuition. [Emphasis added.]

Again, this raises the question of why the results of the study lead us to conclude that we need more generation compensation for above average teachers rather than more generous tax subsidies — or pension credits for stay-at-home parents who might contribute to a child’s cognitive development, etc.

P.S. John Cassidy of The New Yorker has written a wonderful post on this subject. Here is just one of a number of interesting observations:

Then there is the puzzling finding that the kindergarten effect on wages seems to grow over time. (See Figure 6F.) From 2000 to 2003, those people who attended a below-average kindergarten class earned very similar wages to those who attended an above-average kindergarten class. But between 2003 and 2007, the typical discrepancy in annual wages grew to $508. How does this fit in with the rest of the researchers’ argument? Perhaps it gels, but it also seems to be consistent with a story about innate differences in ability and aptitude that the market gradually uncovers and values.

For all my disagreements with Cassidy, he really is an indispensable economics journalist.