Why do migrants send money back home? Distinguishing the different motives helps us understand the role these transfers play in influencing the behavior of households, and the policy implications of alternative motives can be very different.
I tried answering this question using micro survey data from Bangladesh on possible motivations, using a multivariate regression model.
The results were a little unexpected. Overall, the evidence contradicts the argument that remittance-receiving countries have little scope for policy intervention. The analysis shows that remittances are not driven exclusively by the need for family support but also by the migrants’ skill and education level and motivation to transfer their savings as investment in their home country. Thus, contrary to conventional wisdom, remittances play a vital role in not only supporting consumption but also in serving as an important source of investment funding. The extent to which remittances contribute to investment depends on the supportiveness of government policies and whether the economic environment is conducive to investment activity.
Surprisingly, none of the demand side variables—the existence of a surviving parent or spouse—seem to matter. Among the supply side variables, education and skill matter most.
A migrant with secondary education (8 to 10 years of schooling) is likely to remit Tk 30,000 (US$ 375) more on average per annum than a migrant without secondary education and a migrant with even higher education is likely to remit on average Tk 40,000 (US$ 500) per annum more than a migrant without higher education. The destination country does not seem to matter i.e. migrants’ remittance behavior does not depend on their country of employment.
There is also fairly strong and robust evidence that motivations other than altruism are important determinants of remittance behavior. The more land owned by the recipient household, the higher the remittances it receives, indicating that remittances are motivated by continued maintenance of land assets at home or perhaps the aspiration to inherit the land. More powerful evidence is the positive and highly significant coefficient on the pre-remittance income of the receiving households. This is completely unsupportive of the altruism hypothesis and quite consistent, though not the only possibility, under the self-interest hypothesis. The coefficient itself is also economically significant--one taka increase in the pre-remittance income of the household increases remittances by Tk 0.16.
There is no evidence supporting remittance decay with the length of stay abroad. If remittance decay were present, the remittances would be expected to decline the longer the migrant’s stay abroad. The results however show that the coefficient is positive and highly significant while the coefficient on square of length is negative and significant. This suggests that the amount remitted by a migrant worker actually increases, but at a decreasing rate, with the migrants’ length of stay, controlling for other variables.
If migrant’s remittance levels are positively related to the length of stay, as in this case, aggregate remittance levels may not decline, even if the rate of new migration is insufficient to offset the losses due to death or return migration. From the migrant sending country’s perspective, the extent to which remittance is responsive to variables other than the needs of dependents left behind determines the space for government policy interventions to induce higher remittance levels.
The bottom line for Bangladesh appears to be investing in enhancing the capacity to remit by increasing the education and skills of its potential migrants and providing a hospitable environment for investment of migrants’ savings at home.
Photo credit: Arne Hoel/World Bank