| The first chapter of this dissertation investigates the dynamics of health. We estimate a simple model describing health dynamics that allows for heterogeneous first order state dependence. We find evidence of state dependence in health. This finding suggests that health is inertial at the individual level. We find that the occupancy of a health state today affects future health for approximately five years. In addition, we find that health dynamics are strongly tied to socioeconomic status; we find a gradient in the persistence of pre-existing illness and we find a steeper gradient in the onset of new illness. Our findings are informative of the nature of the gradient in individual health dynamics.; The second chapter considers the use of migration between the US and El Salvador as a means of self-insurance for rural Salvadoran households. We show that households responded to adverse agricultural conditions in El Salvador by allocating more members to the US. In addition, we show that households responded to damage sustained during the 2001 earthquakes by retaining household members in El Salvador. These findings are consistent with a model in which agricultural shocks lower returns in El Salvador, thereby, inducing northward migration and housing damage raises the marginal utility of housing in El Salvador, thereby, creating exigencies at home which stunt northward migration. Finally, the effects of the earthquakes were independent of household wealth suggesting that the earthquakes affected migration through a channel other than liquidity constraints.; The final chapter of this dissertation tests the proposition that migrants are healthier than non-migrants. Our findings are supportive of this proposition and, thus, suggest that health induces selection in migration. In addition, we find evidence of a graded relationship between health and mobility so that good health is more closely associated with higher degrees of mobility than it is with lower degrees of mobility. Our results suggest that healthier people are better able to relocate to areas with higher wages and, thus, suggest a mechanism that would generate a causal effect of health on earnings. Finally, we show how non-random durations in the panel can create a selection bias. |
