True state dependence and heterogeneity in off-farm labour participation

This paper discusses models of off-farm labour participation estimated with panel data. With this kind of model, unexplained persistence in a labour state may be due to true state dependence (when a previous state modifies attitudes, constraints or parameters) and/or to heterogeneity (unobservable permanent individual characteristics). Probit models assuming either heterogeneity or true state dependence are estimated from a two-period panel survey of Pennsylvania farms, and compared with cross-section or pooled sample models. Statistical tests to infer the source of persistence are proposed. The results suggest that for Pennsylvania farmers and spouses, in addition to other personal, household and farm characteristics found to be important in previous studies, previous off-farm labour state is relevant in off-farm labour participation decisions.