This paper explores a general rule of the one-child policy: the ethnic minorities
were allowed to give birth to at least two children in the entire 1980s, as a natural
experiment to identify the e§ect of the one-child policy on the increase of sex ratios
in China. Using the Chinese population census in 1990, we Önd that the one-child
policy has indeed played a dominant, if not the sole, role in the increase of sex
ratios in China. The estimated e§ect on the probability of being a boy is as large
as 1.01 percentage points for the post-treatment birth cohorts, which implies that
the strict enforcement of the one-child policy has causally increased the sex ratio at
birth by 4.4. Further exploration reveals that the policy e§ect is mainly driven by
rural residents, and the second and higher birth parities. In addition, the robust
analysis indicates that the estimated treatment e§ect is not likely to be confounded
with the rural household responsibility system reform, the availability and spread of
modern gender selection technologies, and the epidemic of hepatitis B