Tom Cui @tom-cui.com · Dec 12 What do we get as a result? An inferred set of borders where we believe zoning districts change *within cities.* But instead of retrieving entire borders, we can filter on their "straightest" segments, which are least likely to be drawn to protect high-value land or development.
0 likes 1 replies
?
Replies
Collecting these border segments together lets us use a cleaner border discontinuity design, but also lets us identify a more policy relevant effect: *Context-specific effects,* or "how do regulations' causal effects vary based on their marginal distortions to surrounding market-supplied density?"