Abstract
Using a unique data set on the financial sector, this article assesses the impact that financial sector development has on international remittance flows for a sample of 64 countries. The results show that greater financial sector development – as measured by bank branches per 1000 km2 – results in greater remittance flows to a country. However, this study also documents that transaction costs have no impact on remittance flows. This latter finding has important policy implications as reductions in transaction costs are often cited as an important approach to increase remittance flows.
Notes
1See ‘Global Poverty Action Plan approved by G8 Leaders’. Bureau of International Information Programs, US Department of States, Washington, DC (2004).
2Previous studies, Freund and Spatafora (Citation2005) and Niimi and Ozden (Citation2006) in this genre have typically used a more macro-oriented proxy for financial development such as the ratio of bank credit to GDP. However, people's access to remittance flows is better measured by a more micro-oriented proxy such as the one employed in this article.
3Because the data on the bank variable are available for only 1 year, a panel regression was not possible.
4This measure has been used by several other studies such as Agarwal et al. (2006) and Guiliano and Arranz (Citation2005). Also see Chami et al. (Citation2008) for a discussion of other measures of remittance flows and their use.
5It should be noted that in addition to the variables used in the final regressions, remittances as a percent of GDP, HGDP growth and GDP per capita growth were included in the separate regressions as dependent and independent variables, but yielded no significant results. As a result these results are not reported.
6Other variables were also used in the regressions, such as size of country and real exchange rates. But none of these variables enhanced explanatory power nor were the summary statistics improved. As a result and to conserve space we do not report these results.
7Regressions were also run after excluding India, China and Mexico from the sample. Although the significance levels and explanatory power of the model declined, all the same variables in the earlier reported findings remained statistically significant at the 5% level. These results are not reported.