54
of consumption is equal to the effective wage, which is the market wage minus the variable
transaction costs, multiplied by marginal utility of consumption, minus the shadow value of
liquidity times the variable transaction costs of off-farm labor. The utility-maximizing
household modeled here will allocate less time to off-farm labor as liquidity is more
constrained, ceteris paribus. In other words, if an agent’s liquidity constraint is relaxed
through program payments, then an agent allocates more time to off-farm labor and less to
on-farm labor (hypothesis 1). It also implies that the more liquidity-constrained the farmer is
prior to the program, the larger the effect of program payments on off-farm labor (hypothesis
2)
Using Implicit Function Theorem, we can also derive that higher compensation rate
I— > ° I
increases off-farm labor ∖ :. / if the marginal product of on-farm labor is larger than
the effective wage , and vice versa. Testing this hypothesis is beyond the scope of this paper,
although previous studies of the Grain for Green program have found that the compensation
rate typically is larger than the value of the crop yielded by the retired plots (i.e., the
opportunity cost of program participation) (Xu et al., 2006; Uchida et al., 2004). For example,
Xu et al. (2006) found that the value of preprogram production for more than 70 percent of
participating households was less valuable than the compensation rate. Furthermore, the level
of compensation is not trivial relative to the earnings of the typical participating household in
the study region. For example, if an average household in Sichuan Province (Yangtze River