lukestein’s avatarlukestein’s Twitter Archive—№ 10,557

  1. …in reply to @causalinf
    @causalinf @RevOfFinStudies Thanks, Scott! We decided to mainly focus on the effect of having an *account* in the household (endogenous, plus subject to massive measurement error) rather than the effect of simply being proximate to a bank. The former seemed to benefit from the IV implementation…
    On twitter.com Twitter logo lukestein’s avatar ♻️ 2 Retweets ❤️ 1 Favorite Mood +2 🙂
    1. …in reply to @lukestein
      @causalinf @RevOfFinStudies …but of course given the instrument, the underlying reduced form is basically a diff-in-diff. I l♥️ve the figures that show those specifications, but alas they’re in the appendix. I always show them when I present, though!
      oh my god twitter doesn’t include alt text from images in their APIoh my god twitter doesn’t include alt text from images in their API
      Permalink On twitter.com Twitter logo lukestein’s avatar ♻️ 2 Retweets ❤️ 1 Favorite Mood -1 🙁
      1. …in reply to @lukestein
        @causalinf @RevOfFinStudies I’m also realizing now Scott that you may—very reasonably!—have been asking about a traditional diff-in-diff over time rather than over distance as in my last tweet. Not doing that has a simpler and sadder explanation: …
        Permalink On twitter.com Twitter logo lukestein’s avatar ♻️ 2 Retweets ❤️ 1 Favorite Mood 0
        1. …in reply to @lukestein
          @causalinf @RevOfFinStudies The decennial census didn’t include slaves. (A separate Slave Census didn’t even include names, and seemed largely focused on enumerating slaves as if they were capital😢⛓). So there’s basically no pre-period outcome data for our study.
          Permalink On twitter.com Twitter logo lukestein’s avatar ♻️ 2 Retweets ❤️ 2 Favorites Mood -8 🙁