This paper exploits a quasi-natural experiment – the U.S. granting of Permanent Normal Trade Relations (PNTR) to China after China's accession to the World Trade Organization – to examine whether trade liberalization affects the incidence of child labor in China. PNTR permanently set U.S. duties on Chinese imports at low Normal Trade Relations (NTR) levels and removed the uncertainty associated with annual renewals of China's NTR status. We find that the PNTR was significantly associated with the rising incidence of child labor in China. A one percentage point decrease in average export tariffs raises the odds of child labor by a 1.3 percentage point. The effects are greater for girls, older children, rural children, and children with less-educated parents. The effect of trade liberalization on the incidence of child labor, however, disappears in the long run, because trade liberalization can induce exporters to upgrade technology and thus have less demand for unskilled workers.