• There needs to be reasonable wireless infrastructure and use for mobile banking to succeed. Turns out that while wide and high penetration of mobile is certainly helpful, it’s unnecessary for launching successful mobile financial services (MFS). Mobile banking can grow with, and not just because of, mobile usage. M-PESA, the world’s most successful mobile banking service, launched when mobile penetration in Kenya was roughly 20 percent. Tanzania and the Philippines were at similar levels of penetration when their MFS systems launched.

• Mobile banking succeeds where there is low penetration of formal banking. Penetration of traditional banking varies in countries with mobile banking, and widely available formal banking doesn’t seem to diminish the value of mobile alternatives. In fact, the volume of potential customers (those unbanked or excluded from commercial banking) doesn’t appear to be a driver of mobile banking. In countries with high penetration of traditional banking, MFS can be additive; but in countries with low penetration, mobile offerings don’t necessarily lead to transformational outcomes.

• There must be strong regulatory frameworks to limit risks and ensure secure service. The study found no consistent pattern in the strength of financial systems across countries where mobile banking has succeeded. Legal protection ranged from relatively strong (for an emerging market) in Kenya to relatively weak in the Philippines. (There is no easy way to quantify and holistically assess these factors; here, the researchers used the World Bank’s strength of legal rights index—“the degree to which collateral and bankruptcy laws protect the rights of borrowers and lenders and thus facilitate lending”—as a proxy, catch-all measure.)

• The high cost of formal banking leads to greater enthusiasm for lower cost, mobile alternatives. This makes intuitive sense: if Option A is pricey, people will want a cheaper Option B. But with regards to MFS, the cost of the banking system does not seem to be a driver. In mapping costs for money transfer, a common transaction, the researchers found that for both domestic and international remittances, costs in countries with mobile banking varied widely. (It’s not clear whether indigenous alternatives such as padala, fei chien, or hawala, or services such as Western Union.)

• Insecure areas tend to see greater uptake of mobile banking, as people seek safer methods for carrying, storing, and transferring money. The study found that security (as measured by the homicide rate) does not necessarily explain enthusiasm for mobile banking. Compared to Colombia, South Africa, Brazil, and Mexico, Kenya is a relatively safe country and thus the rapid adoption of M-PESA cannot be attributed to the need for security in dangerous environments. While risk mitigation may be a contributing factor, it doesn’t seem to be important as it’s often made out.

Mariscal and Flores-Roux’s conclusion? With the exception of basic wireless penetration, the enabling factors we so often speak of neither explain the existence nor the success of mobile money systems. Market variables, they conclude, thus need not be enabling, they only need to be non-hindering. While the researchers only examined three mature deployments (Kenya, the Philippines, and Brazil) at depth, they pepper their investigation with evidence from Tanzania, Afghanistan, Rwanda, Colombia, South Africa, and Mexico.

Mobile money systems, Mariscal noted, are simple in concept, but highly complex in execution. Take infrastructure development. Beyond physical locations, operators must also have a network of people to serve as cash-in, cash-out agents. And while it may seem redundant to emphasize that mobile agents are indeed human beings, I’m consistently surprised by how often this fact is overlooked in the design of services and business models.

Effective systems respect and play to the unique characteristics/constraints of the human ‘nodes’ that operate within them — mobile finance is no different. An established network of cash-in and cash-out points does not equate a productive network of cash-in and cash-out agents; the latter must be nurtured through shrewd recruitment, training, incentives, management, and support. This takes careful research, planning, and design. Further, though many speak of the need for agent ubiquity (both wide and dense penetration), I rarely hear talk of the need for reasonable standardization in agent processes. The former leads to usage—agent ubiquity is easy enough, it just takes time and money. The latter encourages loyalty—but ensuring your users can get their money each and every time is not so easy, and it requires a system design that understands human (agent) motivations and limitations.

But I digress. In short, it appears that some of the more sympathetic narratives around mobile banking have been liberally generalized and even conveniently hyperbolized. We need to reexamine industry adages using actual data, not opportune anecdotes. Mariscal and Flores-Roux show that mobile banking outcomes, like most everything else, depend on unique factors for each context. Thus, rather than trying to identify easy wins and replicable formulae, service providers would be wise to roll up their sleeves and get down to work. (Hint: Good design starts with good research.) Clone deployments have a long shot at success, but custom solutions informed by rich contextual understanding stand a fair chance.

Image: Ken Banks, kiwanja.net