Paper short abstract:

Remittances do not affect spending on human capital or basic goods in Tajikistan. The only category of goods for which the share of expenditures systematically increases with remittances is the conspicuous goods category. Remittances and local revenues cannot be considered as substitutes.

Paper long abstract:

Tajikistan between 2007 and 2011 was the country most dependent on remittances. Using survey data covering this period, I analyse the impact of remittances on household expenditures. I apply a Working-Leser model and I control for both individual heterogeneity using individual fixed effects and endogeneity using a set of instruments. I find that, in general, remittances do not affect spending on human capital or basic goods (including food). The only category of goods for which the share of expenditures systematically increases with remittances is the conspicuous goods category. I propose to use PCA to sort durable goods and separate the basic goods from the secondary goods. Depending on the total expenditures of the household, education and health spending change as well. These changes in consumption habits appear to be temporary: households with long-term migrants or with former migrants have no change in their consumption. Even though remittances are undoubtedly worthwhile in reducing poverty and thwarting emergencies, the paper suggests that they do not necessarily improve the long-term development of Tajikistani households. Therefore, I conclude that remittances and local revenues cannot be considered as substitutes.