Preventing the Digital Trail from Going Cold: Lessons from Mexico

In recent years, digitizing and delivering Government to Person (G2P) social payments into individual recipients’ bank accounts has been considered a potential gateway to financially include significant numbers of poor people. Two key questions to consider in these efforts are: How far should the digital trail reach and how should an effective G2P agent network be built and managed?

An ambitious project supported by the Bill & Melinda Gates Foundation between 2008-2013 to digitize G2P payments of Mexico’s largest conditional cash transfer program—Oportunidades, which has more than 6.5 million recipients —and build a network of more than 7,000 banking correspondents through a public rural retail network—DICONSA—and a state savings bank—BANSEFI, provides a number of useful insights (read the complete public summary report).

Photo Credit: Yolanda Luna, 2012 CGAP Photo Contest

Electronic tracking of Oportunidades payment delivery through account-linked and biometrically-enabled cards increased program efficiency and transparency and improved targeting, timeliness and reliability of delivery to recipients. For their part, recipients benefited from a 77% reduction in transaction and opportunity costs, as G2P pay-points were set up no further than 4km from their homes. BANSEFI and DICONSA also benefitted from a juicy commission for each payment made.

But success did not extend to the subsequent financial inclusion initiative. Availing recipients of financial services beyond G2P payment delivery through a digital channel turned out to be easier planned than accomplished. The banking correspondent footprint was neither significantly nor sustainably expanded—the number of agents stood at 261 in July 2013 and only 115 were disbursing G2P payments to a mere 14,000 recipients. And the promise to offer a range of relevant financial services to previously unbanked populations, especially in rural areas, went unfulfilled.

Fingers have been pointed at an Oportunidades-mandated requirement that compels rural recipients—more than 5 million people—to cash-out their benefit in full, at a designated pay-point during a specified period of time, then use the card solely for identification purposes. This restriction effectively renders rural recipients’ accounts into bimonthly, temporary repositories of G2P funds, and nothing more. It also cuts the reach of the digital trail short.

But what would happen if rural recipients’ accounts were fully functional?

As things stand today, depositing rural recipients’ G2P funds in a transactional account they can use wherever, however and whenever they choose would have two main drawbacks: (i) Oportunidades would have to relinquish obtaining proof of life through biometric identification as the country’s financial infrastructure does not support biometrics; and (ii) recipients would have to forego the cost-reduction benefits they have enjoyed, since capillarity of ATMs or POS-enabled cash-in/-out/electronic payment-receiving points in rural areas is practically non-existent.

The agent network established under this project achieved very limited outreach and was also plagued by execution challenges, ineffective supervision arrangements, liquidity management problems, poor connectivity, security concerns, and lack of relevant and affordable financial products, among others. Understandably, this is not the kind of service experience to which Oportunidades wants to subject its recipients.

What are some of the practicalities to consider when building and managing a built-for-purpose agent network to improve its chances of success? Here are five key insights from the project:

1. A channel designed and custom-built for G2P payments is not easily converted to a financial services distribution arm. To have a chance to succeed it requires: (i) sufficient agent capillarity; (ii) a dependable agent network; (iii) a robust technology platform and reliable connectivity; (iv) a stable flow of account-deposited G2Ps; (v) solid liquidity management; (vi) a relevant and affordable product offering and marketing; (vii) trusted, well-trained and adequately incentivized agent operators.
2. A built-for-purpose agent network for G2P payment delivery will incur high initial set-up costs and high on-going liquidity management costs. Sufficient funding must be made available to cover planned and unforeseen expenses that a network will incur while it is able to scale its outreach and transaction volumes.
3. Whatever the structure of the agent network, it must be dependable. There is no best structure for an agent network, but having a business-oriented, experienced and capable agent administrator(s)—whether provider-operated or outsourced—is a must. The administrator must be capable of establishing, growing, managing and training the agent network sustainably.
4. Incentives need to be aligned for all participants in the correspondent banking chain, but particularly for correspondent shopkeepers. While all participants need appealing-enough incentives to participate in this effort, client-facing banking correspondents need to be particularly motivated through training, technical support and sufficiently attractive commissions.
5. Security risks will increase costs and may deter potential agents from participating in G2P payments. Given that most G2Ps are paid in cash in remote rural areas, there is bound to be a risk of robberies and crime. It is essential to identify and manage security risks adequately and to integrate this cost into the equation. If crime is severe and requires that a larger proportion of the commission be spent on security, risk and cost may outweigh the benefits to payers, including banks and agents.

As this case study shows, setting up a successful network of agents in rural areas to disburse G2Ps electronically and offer a suite of adequate and affordable financial products with the necessary depth and outreach to be relevant in poor people’s lives is a very challenging undertaking that will continue to require careful planning and deep pockets.