endogeneity concerns associated with including the number of SEZs in a given year within a district as a dependent variable, I construct an indicator variable for SEZintensity, following Wang (2013) . The indicator variable, M d , is equal to one for the treated district, if, at the end of the sample period, more that one SEZ is located within a district, and is equal to zero when only one SEZ is present. I augment the baseline specification by interacting the indicator variable for multiple SEZs with the dummy for SEZ presence. The resulting specification for testing the
This study examines the socio-economic impact of special economic zones (SEZs) in Cambodia---a prominent place-based policy established in 2005. The paper employs a database on existing and future SEZs in Cambodia with matched household surveys at the district level and documents stylized facts on SEZs in a low-income country setting. To identify causal effects of the SEZ program, the paper (i) constructs an alternative control group including future SEZ program participants and districts adjacent to SEZ hosts; and (ii) employs a propensity score weighting technique. The study finds that entry of SEZs disproportionately benefits female workers and leads to a decline of income inequality at a district level. However, the findings also suggest that land values in SEZ districts tend to rise while wage levels remain largely unchanged relative to other districts. In addition, the paper tests for socio-economic spillovers to surrounding areas and for agglomeration effects associated with clusters of multiple SEZs.