The empirical analysis of the determinants of migration and remittances in Kenya and the impact on household expenditure patterns

This thesis conducts empirical analysis on the determinants of migration and remittance sending decisions in Kenya and the impact on the expenditure patterns of households using cross-sectional household survey data. The first empirical chapter explores the factors that influence the subsequent migration decisions of Kenyan siblings using binary logit models. The findings reveal that preceding sibling migrants have a strong negative effect on the probability of migration for other siblings. Evidence in support of migration as a joint household level decision is obtained as preceding sibling and non-sibling migrants are found to exhibit similar effects. Conditional on migrating, siblings are shown to utilize existing sibling networks by moving to the same internal or external destination as preceding migrants. Discrete failure time models are also employed so as to account for any neglected heterogeneity at the household level. Controlling for neglected heterogeneity, the overall effect of preceding sibling migrants is found to be statistically insignificant. However, non-sibling migrants are found to decrease the probability of migrating. The second empirical chapter examines the remittance behaviour of multiple compared to sole sibling migrants, and the motivations of Kenyan siblings in sending remittances to their household of origin. No evidence of selection bias in the decision to remit is detected when a Heckman selection model is estimated. Using probit and OLS models, the presence of other siblings is found to decrease the probability of remitting but to have no effect on the amount of remittances sent. The amount of remittances sent by other siblings is also found to have no statistically significant effect on the remittances sent by a sibling using IV regression methods. In the third empirical chapter, the expenditure patterns of Kenyan households are investigated according to whether the household is a migrant or non-migrant household, and whether a migrant household is in receipt of remittances or not using an Almost Ideal Demand System (AIDS) approach. The analysis reveals that remitters who are spouses and siblings of the household have higher bargaining power towards the allocation of remittances to physical investments and durable goods, respectively. The expenditure patterns also show that remittances are not pooled together with general income when allocating the household budget towards durable goods and physical investments. In addition, the findings reveal that the reported uses of remittances by Kenyan households contrast with their actual uses. In the fourth chapter, the uses of remittances for the acquisition of physical investments and durable goods are analysed in more detail using IV and bivariate probit models. Remittances are found to be exogenous for the durable goods category but endogenous for physical investments. The evidence obtained is supportive of remittances being used by households to purchase these categories of commodities