Remembrances of things past: evidence from a twenty-year Kerala panel

ABSTRACT We construct a panel of 242 households from five consecutive Kerala Migration Surveys that span 20 years at five-year intervals to study the fundamental determinants of the decision to migrate abroad as well as the decision to remit. Accounting for time-invariant unobservables and allowing migration and remittance behavior to depend upon previous choices clarifies our understanding of both decisions. Migration and remittance behavior display positive serial correlation over a five-year time horizon and the presence of a return migrant in the household increases the likelihood of migration by 13% and remittances by 4%. Migration is 1% more likely in female-headed households, 4% less likely when the household head is employed, increases by 0.4% for each additional year of the household head's age and is 6% more likely in households that are asset-poor. Remittances are between 20% and 70% more likely to obtain when the migrant was married at the time of migration and 3% less likely when the household head is employed, the latter suggesting either an old-age security or a co-insurance motive. Evidence in favor of a very strong inheritance competition motive is found in that each additional male heir increases the likelihood of remittances by between 8% and 31%. Based on our econometric evidence, and in particular our findings pertaining to serial correlation and the presence of a return migrant in the household, it is likely that both migration from and remittances to Kerala will quickly rebound to their pre-pandemic levels.