Under the assumption that markets are complete, the simultaneous production and consumption decisions made by farm households are substantially simplified into a recursive system with production choices preceding consumption decisions. This is a powerful assumption that lies at the heart of many empirical models of farm behavior in the literature. The majority of studies that have assessed the validity of the recursion assumption determine whether there is a link between labor demand on the farm and the demographic structure of the farm household. If markets are complete, there should be no links. Empirical implementation of these tests is complicated by endogenous behavioral responses of farm households and complex measurement challenges. Using extremely rich data that were designed for this research, we develop and implement a new test for market completeness that exploits the fact that, under recursion, farm profits only affect consumption through an income effect. Exploiting plausibly exogenous variation in the local market prices of farm inputs, we test the implication using longitudinal survey data collected over six years from a large sample of farm households in rural Java, Indonesia, by estimating a flexible demand system and taking into account time invariant farmhousehold heterogeneity. Overall, the assumption that markets are complete is rejected but there is an important sub-group of better off households who behave as if markets are complete.