Using population census data, this paper examines the causal effect of a second child on married women's labour force participation in urban China. To ameliorate the endogeneity of fertility, we exploited twin births as the source of variation in fertility. Although the ordinary least squares estimates indicate that having one more child significantly reduces female labour force participation by around 6 and 9 percentage points in 1990 and 2000, respectively, our causal analyses suggest very small negative effects in 1990 (around 2 percentage points) and an insignificant effect in 2000.