How did we arrive at our findings? A brief methodology

UNESCO education statistics show that there are 220 million children out of school in lower and lower-middle income countries (LIC/LMIC). We examined household surveys in Nigeria, Uganda, Tanzania and Pakistan, and calculate that an average of 4% of school-age children[1] in these countries live in families which receive remittances. Applied to all low-income/low middle-income countries, this proportion is equivalent to 9 million children. We use the household surveys to quantify the relationship between receiving remittances and being out of school at the level of individual children and find that children are 40% less likely to be out of school if they are in households which receive remittances. Where possible, we controlled for parents’ education levels and household income, which we also found to be an important factor, but the inclusion of which did not in general affect the coefficient on remittances very much. From this we can estimate that 3.5 million children (40% of 9 million)[2] in LIC/LMIC are kept in school because of remittances.

The World Bank estimates that the total value of international remittances was $689Bn 2018. Using the World Bank’s Remittance Prices Worldwide data, we calculate that the average cost of digital remittances[3] is 5.6%, compared to 7.7% for offline remittances. So digital remittances are 2.1 percentage points or 27.3% cheaper than offline remittances. If all remittances were digital, this would save $14bn (2.1% of $689bn). We calculate that households in our sample spend an average of 5.7% of household resources on education. If this fraction of the remittance savings were spent on education, this would amount to $825 million based on the typical household spend in the 4 countries surveyed (5.7% of 2.1% of $689bn).

Country-level data calculations

Country-level calculations follow the same method and sources. We calculated the fractions of (1) households receiving remittances and (2) children out of school at the country level. These vary a great deal across countries (see table below). The effect of remittances on the chance of being out of school was estimated separately but remarkably consistent (an approximately 40% reduction) across countries.

Our estimates of what the remittances pay for in terms of school books etc, is based on WorldRemit calculation of the costs of these goods at a country level, matched up against the increase of expenditure on education in remittance receiving households. To calculate the latter, we estimated average expenditure on education at the household or child level and then the proportional increase in expenditure in households that receive remittances – the increase in expenditure being the product of the two.


[1] Between 5 and 18 years old inclusive

[2] This is an approximation to the exact calculation, which you can find our replication instructions.

[3] Defined as those datapoints where pay-in methods are either ‘internet’ or ‘mobile phone’.