其他摘要：Off-farm labour participation is an important way in which farm households adapt their labour resources to farm labour needs, and is often viewed as an income integration and an insurance against risk. Nevertheless, it has also been questioned as a step for exiting agriculture. It is therefore important to assess whether or not it is a permanent status and which are its determinants. Most papers on this issue are based on cross-sectional analyses and thus disregard the problem of persistence in the state. Using a 5-wave panel of Italian family farms we estimate different dynamic nonlinear panel data models of the determinants of off-farm labour participation. We allow for two sources of persistence: unobserved heterogeneity and state dependence, and in addition we control for the initial conditions problem. We find a strong persistence in the state and our findings show that, when taking all these features into account, the present work state is almost totally explained by the previous state and by idiosyncratic characteristics. The variables concerning the farm and the farmer’s characteristics, typically found to be relevant in cross-sectional analyses, are not significant in the dynamic setting. The reasons for the inconsistency between our results and those of cross-sectional studies are discussed, and an interpretation of how the determinants influence the off-farm labour participation is presented. The distinction between true state dependence and individual heterogeneity has important policy implications that are discussed.
