between the first and the last available years. Tboung Khmum province is not included in the CSES survey. 4 In Panel B, coefficient on the time trend is negative and, given that the average length of the post-treatment period is about four years, the total size of the SEZ effect on female employment is attenuated to about 2 percentage points. Negative growth effect, however, is not robust to other specifications in Table 2 and Section 7. 5 Specification in Panel B shows that entry of SEZs is associated with about 4.6 percent decline in income inequality