This exchange between Heckman and Chetty from last month is interesting, including on data quality, linkage, use of rank-rank regressions, interpretation of neighborhood effects, using mover designs here vs in MTO study (Heckman’s comments begin at 1:27) http://www.kaltura.com/index.php/extwidget/preview/partner_id/1449362/uiconf_id/14292362/entry_id/1_kb8z3hvs/embed/auto?&flashvars%5BstreamerType%5D=auto

Very revealing of different approaches: when dealing with a criticism Chetty mentioned digitizing new earlier admin data while Heckman suggested using more structure and method of simulated moments with existing survey data like PSID.

Heckman critiqued the neighborhood exposure interpretation saying experimental evidence from MTO suggested effects accrue earlier in childhood than suggested here.

Chetty had a compelling response: MTO was comparing movers with stayers while this is comparing movers to “good” vs movers to “less good” areas. Here the disruption effect of moving cancels out between treatment and control allowing + location effects even later in life.

This type of exchange would be useful for journalists and other non-economists to watch to get a better sense of how empirical economics actually works. Sometimes contentious, rarely uncritical acceptance of findings.

Something that struck me: when Heckman questioned the causal interpretation of location on racial disparities (as opp to sorting), example of unobserved factor usually "ability" or "parenting practices." Too often this is the frame. But what if sorting instead on discrimination?

If people of color have different amount of "cultural capital" that is partially protective against discrimination and also allows entry into "good neighborhood," the critique of causal effect of location is not due to "ability bias" but "discrimination bias."

FWIW, I think Chetty's argument that place matters is fairly compelling, so I am not critiquing that. Rather, I think it's useful to avoid a frame where the default alternative explanation always necessarily goes to racial disparities in ability or parenting practices.