Most people in rural China have no plans for retirement other than the ingrained
Chinese tradition that children care for old parents. Actually there are also no sources of social support
such as social old-age insurance to rely on in rural people’ old age for a long time in China. In 1992, a
social old-age insurance program, rural pension program, was initiated by the Chinese government to
firstly establish a social security system in China’s rural area. The rural pension program experienced
rapid development in the beginning years but grounded to halt after 1998. Since either children or
pension program provides support for elderly, we expected that these two can be viewed as substitutes
to some extent. Using data from China’s 2005 mini-census, we find that rural people who have at least
one son are less likely to participate in pension program and each additional son and daughter both
decreases their participation rate. Moreover, the effect of an additional son is much larger than that of
an additional daughter. In addition, both evidence from mini-census and China Health and Retirement
Longitudinal Study show that peasants accessing to pension are less likely to rely on their children for
support in old age. These findings suggest that demand for children, especially for sons are partly
driven by concerns relating to care in old age; children and formal social old-age insurance are
substitutes for support in old age. We then expect that implementation of social old-age insurance may
mitigate rural people’ demand for children, especially sons and thus correct China’s severe sex ratio bias
to some extent. We test this hypothesis using the difference-in-differences strategy, and find that
increase of sex ratio at the region level slowed down after the implementation of the rural pension
program. Overall, our empirical analysis in this paper implies that sex ratio bias is partly due to
demanding for sons for support in old age and carrying out social old-age insurance in rural China are
helpful in mitigating demand for children and correcting sex ratio bias.