Informal health insurance mechanisms: Do Pakistani migrants protect parent households against consumption losses after illness

The study tests the hypothesis that migrants offer insurance against illness. The analysis uses panel data on about 732 households spanning 44 villages over 3 provinces in rural Pakistan over three years with a 20-year series of district level rainfall data. The probability of sending out a migrant was treated as an endogenous variable and instrumented in a first stage Probit equation using as instruments the characteristics of household adults and leads of the number of disability days that might have been expected by the household. Consumption variability and remittance receipts by households who had sent migrants were then studied to see if remittances were more likely in the period following an illness shock and whether the remittances lowered consumption variability.; The first model examines the receipt of remittance following an illness shock in the presence of migrants. The data suggests that household level illness shock positively influence remittance from international migrants, but not from internal migrants. District level shock, measured as a deviation in rainfall, positively affects the amount of remittance from all households if they are likely to receive transfers. However, the likelihood of remittance is not affected by rainfall shocks.; The second model examines consumption smoothing in the presence of illness shocks and migrants. The data suggest that household level shocks, such as illness shocks, and income shocks adversely affect household consumption smoothing. Households that are poorer and have a long illness shock or an income shock are unprotected against consumption variation in the presence of migrants. The results however show that the presence of especially internal migrants, and not international migrants, may have some protective effect on smoothing consumption among poor households.; Finally, although this study shows that consumption smoothing is an important motivation for gifts and informal loans; nevertheless, Inter-household transfers, alone do not appear to be used primarily to smooth consumption for all migrant households, which means that the efficiency of risk sharing through all migrant transfers is rejected. Risk sharing against idiosyncratic shocks does not take place at the village level but rather within networks of friends and family. The study shows that not all categories of shocks are equally insured even in the presence of networks.