25
participation or on income from off-farm work (Xu, et al., 2005; Uchida, et al., 2005). Furthermore,
since both of those studies used household surveys that collected information on labor-allocation
decisions only for the first three years of the program, it may have been too soon for changes to be
detected. In this study, we use data collected five years after the program began, which may have
allowed sufficient time for participating households to begin to find off-farm employment in numbers
that are statistically detectable.
The positive impact of the program on off-farm labor also is in stark contrast to findings from
studies of the impact of government farm payment programs in the U.S. Previous U.S. studies of
government payments to farmers, including the Conservation Reserve Program, have consistently found
that government payments negatively affect household off-farm employment participation (e.g., Ahearn,
et al., 2005). The results in China may move in the opposite direction for several reasons. The higher
level of income of U.S. farmers compared to what is typical for farmers in the Grain for Green program
in China probably is the most likely reason why farmers in the U.S. do not choose to work off-farm
when offered a government payment (i.e., the wealth effect dominates). In short, the income effect of
leisure may dominate for richer U.S. farmers while the substitute effect may dominate for poor farmers
in China. We also believe that the divergent program effects stem from underlying conditions in the two
labor and credit markets. Although labor and credit markets exist in rural China, transaction costs may
be high enough that households face much larger constraints in accessing them. According to our results,
it appears that Grain for Green is helping to alleviate the liquidity constraints.