Matching on Baseline Characteristics

[Anupam Ghosh]

I want to estimate the causal effect of a time-varying heterogeneous treatment on the US county crime rate. In my setting, about 30% of counties are treated and the rest are pure controls. As one of my robustness checks, I want to do some form of nearest-neighborhood matching, wherein only those counties that are closest in terms of a vector of baseline (1980) characteristics will be retained in the sample. Can anyone suggest if there is a way to do this? I want to implement the following regression later. I guess I only want to estimate the PScores.

qui reghdfe LProperty_3034_p Minor_Hurr0_4 Minor_Hurr5_10 Major_Hurr0_4 Major_Hurr5_10, absorb(i.fips i.year#c.Share_Police_1980 i.year#c.Share_IndexCrime_1980 i.year#c.Share_Black i.year#c.Share_Male i.year#c.Share_YoungNoHS_1980 i.year#c.Share_Unemp_1980 i.year#c.Share_BPL_1980 i.year#c.Log_Pop_Density_1980 i.year#c.Log_CoastDist_1980 i.year#c.Log_AvgWages_1980 ) vce (cluster fips)

[George Ford]

kmatch eb (or whatever option you want).

it creates a weight.