Remittances represent an important source of income for millions of households around the world. The size of remittance flows, as compared to a country’s own domestic output, can reach numbers as high as 30% (that’s in Tajikistan, if you’re wondering.) This has led economists and policymakers alike to ask whether remittances can be relied upon to spur development. One way this might occur is if remittances are more likely to be spent on productive investments, increasing the domestic income-earning potential of households.
A review of the literature suggests that remittances are spent differently, but none of the studies I found perfectly identifies a causal relationship between remittances and spending patterns. Some find that remittances are associated with an increased share of education in expenditure (see Edwards and Ureta 2003; Adams 2005; Adams and Cuecuecha 2010; Yang 2006; and Cardona Sosa and Medina 2005). Others find that remittances are associated with increases in spending on various investment and consumption goods (Taylor and Mora 2006; Airola 2007; Castaldo and Reilly 2007).
What none of these studies have the chance to do, however, is ask whether the remittance sender would have preferred that the transfer were spent in a particular way. When it appears that remittances tend to be spent on education, for example, is that because the remittance sender insisted on it or simply because the receiver felt that was the appropriate destination?
This is where a recent paper by Nava Ashraf, Diego Aycinena, Claudia Martínez, and Dean Yang, entitled “Remittances and the Problem of Control: A Field Experiment Among Migrants from El Salvador” comes in. They partnered with a bank in El Salvador to offer Salvadoran migrant workers the opportunity to open an account to which they could send their remittances. It’s worth noting that El Salvador is among the countries with the largest remittance flow measured in comparison to GDP, at 15%. They also randomly offered some of the participants the opportunity to have greater monitoring and control over their remittances. Specifically, they could send remittances to a secure bank account that only they could access (although it could be used to transfer to the account of another receiver as well.) Migrants who were offered accounts over which they had the most control sent more remittances in addition to saving more at US-based banks. This is only one study with a specific population, but it suggests that in at least some areas, migrants would prefer that more of their remittances were saved. When they are given the opportunity to ensure that this happens, they send more money home.
Considering the implications of this study brought me back to the mobile payments-remittances link. (Remittances could even be credited with spurring the adoption of mobile money—“Send money home”was the original M-PESA slogan, targeted at migrant workers.) Mobile payments can decrease transaction costs significantly (in Kenya, for example, the fee structure of M-PESA means it can be significantly cheaper than using Western Union or other methods (Jack, Suri 2010). But with all this focus on greasing the wheels to send remittances, we can’t let other issues slide. Even if the cost of sending remittances is driven down to almost zero, migrants may hesitate to send remittances if they are concerned about how (or whether) they are going to be spent.
Mobile payment platforms could even be used to address this potential issue. I would be interested to see what would happen if remittance senders using mobile platforms were given more options on how to control the use of their transfers. Does any of what we have learned about commitment savings products apply when the saving commitment is for someone else? Perhaps e-float could be digitally tagged with a “no cash out before” date set to coincide with the start of the school year, an important holiday, or until a “goal” amount had been accumulated. This would serve as a commitment device not only for the remittance receiver, who could not be tempted to immediately spend transfers upon receipt, but maybe even the remittance sender as well. An e-float tagging system could encourage remittance senders to transfer smaller amounts, without the fear that these small sums will be subject to consumption temptations on the receiving end. The system also means that the sender would not need to keep the accumulating sum in his or her own account abroad, and so would be less tempted to spend the cash that was meant to be sent home.
The regulatory challenges of implementing such a system through M-PESA (or a similar service) would be considerable—M-PESA, like many other mobile money platforms, is not a deposit-taking service. Safaricom’s deposit-taking product M-KESHO has had lower takeup than hoped, but perhaps adding some of these “control” features could give it a boost. There may also be downsides to tying up the funds too tightly; access to cash in an emergency could be crucial. Still, it is worth thinking both big and small about designing services to harness the joint potential of remittances and mobile payments.