This study evaluates the off-farm labor response of rural households participating in the Grain for Green program in China, the largest conservation set-aside program in the developing world. Using a panel data set that we designed and implemented, we examine the impact of the program on changes in off-farm labor participation between 1999 (pre-program) and 2004 (post-program) using a difference-in-differences approach and several extensions that account for program intensity. We also test whether the program impact is diverse depending on level of physical and human capital of participants. We find that on average the Grain for Green program has a positive effect on off-farm labor participation. Importantly, however, we find that program effects vary across groups of individuals in the sample. For example, we find that lower initial levels of wealth enhance the impact of the program on the off-farm employment activity. This result supports our view that the Grain for Green program may be relaxing liquidity constraints for the participating households and that is one reason why participants are more likely to find off-farm employment compared to non-participants. The positive impact of the conservation payments on off-farm labor is in stark contrast with the findings in the US where most studies have found that government payments to farmers decrease off-farm labor participation. One reason for the difference in findings between China and US may be because there are more impediments to participating in off-farm labor market in the poor areas of rural China (the areas in which the programs are being implemented) compared to the US and Grain for Green helps overcome these constraints. It could also be that there are differences in the age structure of the farming population between China (which is generally younger) and the US (which is generally older). This interpretation is reinforced by the finding that, while the average impact is positive, there is an even larger measured positive effect for the younger cohort. The measured effect of Grain for Green is negative for the older cohorts. We also find no impact on off-farm labor participation for individuals with low educational attainment (and positive for those with higher levels of education), suggesting that human capital is necessary when trying to achieve a structural change to earning activities. If policymakers want to achieve a win-win outcome through Grain for Green by meeting both the program's environmental and development goals, they may need to provide extra support (for example, through greater assistance to education) to the vulnerable sub-populations in the program areas.