,
11 tweets,
4 min read
Read on Twitter
Today’s reading was @jmb112485’s paper, which concludes that Chicago’s public housing demolitions increased crime in the short run, the opposite direction from the effect found in my JMP papers.ssrn.com/sol3/papers.cf…
@jmb112485 If someone is going to debunk your paper, this is the way you want it done. A thoughtful and well-conducted study to advance science, rather than pick fights
@jmb112485 Jesse uses a similar estimation method to my paper as well as the Aliprantis and Hartley paper on the same topic and gets similar estimates (yay, the results replicate!) /3
@jmb112485 But then makes a compelling case that the estimates are biased by spillovers into both the surrounding area (which I tried to address) and other public housing (which I didn’t) /4
@jmb112485 He conducts a time-series analysis on Chicago crime by month and finds that demolitions are followed by moderate short run increases in crime 
@jmb112485 One spillover that he points out may be really important is gangs forced into other gang’s territory by the demolition, increasing gang violence. That affect would be correlated with demolition timing and potentially part of my control group /6
He has some awesome data on gang territory in Chicago that I think could have some really interesting additional uses, and shows that crime increases along the territory borders after a demolition
So, the paper convinced me that my finding are potentially biased due to spillovers and that spillovers into gang territory is especially important to consider
I’m still wrapping my head around the time series analysis, since it’s a different method than is usually used in the crime context, which is usually analyzed at a micro (sometimes very micro) scale
But it’s a thoughtful paper and extends and gives context to both my and Aliprantis and Hartley’s papers. I applaud the effort /end
Unroll @threadreaderapp
