02

Employing province-level annual data covering the period 1988-2004,4 we find – by OLS
and instrumenting using the one-child policy programs – that sex ratios raised crime rates,
and estimate an elasticity of about 3.7 percent. Since the 16-25 sex ratio have risen by 4%
(from 1.053 in 1988 to 1.095 in 2004) and crime rates by 82.4% over the same period, this
suggests that the rise in sex ratios can account for up to one-sixth (14.8/82.4) of the overall
rise in criminality during the study period. Our point estimate is robust to the inclusion of
a number of province level covariates that may conceivably influence crime, such as income,
employment rate, education, inequality, and urbanization rate, and a battery of robustness
checks briefly outlined below.
Our empirical strategy is to regress (province-year) arrest rates5 on the sex ratio