Retail not-banking

[Dave Birch] The relationship between payments and banking is, generally speaking, under some pressure from both developments in technology and wider cultural and business changes. Although these pressures exist in the developed world, they are most visible in the developing world where the lack of existing infrastructure means that technology and business change come together unencumbered. The result is called by some “branchless banking”, although in the payments space “bankless banking” might be more appropriate. Anyway, the essence of the concept is access for the financial excluded through non-bank distribution channels. Why? Well, in many (if not most) countries, most people don’t have bank accounts:

Statistics from the Central Bank shows that banks currently hold less than 3 million accounts, both current and savings across the country. The present development which is not healthy for the economy means that less than 15 percent of the about 22 million population in the country saves with the banks. According to the data, while the economy is cash based, 80 percent of cash is outside the universal banks.

[From 18m Ghanaians Shun Banks – ModernGhana.com]

Look at the example of Brazil to see how the technology, business and social factors can co-evolve: In Brazil, payments account for nearly four-fifths of bankless banking transactions. In other words, transactions that are performed over electronic channels (in countries like Brazil, this almost always means mobile, not the internet) are overwhelmingly payments transactions. Again in Brazil, 90% of the people who use some form of branchless banking, use it to make utility payments and other non-interpersonal payments. And of those people, only 5% have a bank account. This market shape is unlikely to change because the cost of providing these people with bank accounts that have the potential to deliver a wide range of financial services, but the cost base to go with that is simply too high. In fact, I would go further and say that it seems to me entirely likely that of the small number of people who do have bank accounts, many of them only really want payments accounts (if we can them that). Therefore if technology makes it cost-effective to deliver payment accounts into mass markets, then the demand for bank accounts at the low end will fall further.

The role of the mobile phone in all of this is critical. Mobile banking providers have now both the products and the implementation experiences needed to open up much wider markets. They are held back or constrained in one rather obvious way, which is that they depend on the handset manufacturers and mobile operators to provide the platform for the functionality that they need, and this often constrains the customer experience, but nevertheless they are taking transactions to the masses.

I saw a presentation by Jim Rosenberg of CGAP. If you’re not familiar with the organisation

The Consultative Group to Assist the Poor (CGAP) is a consortium of 33 public and private development agencies working together to expand access to financial services for the poor in developing countries. CGAP was created in 1995 by these aid agencies and industry leaders to help create permanent financial services for the poor on a large scale (often referred to as microfinance).

[From Consultative Group to Assist the Poor (CGAP) – Wikipedia, the free encyclopedia]

Jim was talking about the CGAP technology program that is co-funded by the Bill and Melinda Gates Foundation and reporting on some of the World Bank’s early experiences looking at Branchless Banking in the developing world. I thought that his prediction of what he called “shared agent networks” was insightful. The concept is that payment and financial service providers would use third party agent networks to reach customers instead of traditional banking networks. In the countries Jim was talking about this generally means retailers. In India, Airtel has more than half a million retailers signed up for top-up services. It doesn’t take much imagination to see that if these 500,000+ retailers were used to deliver financial services, they would become a powerful player. As it happens, in India’s case, regulation can lead to considerable conservatism in this regard which may well hold back the development of the market and deny financial inclusion to large numbers of people. But that will not be true in all regions. How effective this kind of strategy can be has been demonstrated in cases where the mobile operators use their own agent networks to deliver payment services. Case studies such as GCash in the Philippines and M-PESA in Kenya, are well worth examining. M-PESA, for example, had more than 1.6 million customers a year after launch. There is certainly another dynamic at play here, which is that in some developing countries the fact is that people trust internationally-branded mobile phone operators more than they may trust local banks. But nevertheless it is the combination of the mobile phone, as a secure transaction token, and the agent network for distribution that powers a new market.

This view of the future market whereby providers compete through shared agent networks to deliver simple services made cost-effective through the communication processing and authentication capabilities of the mobile handset, is not really that far-fetched. In fact I might go further and say that in a relatively short time it will be mainstream and it will have an interesting impact on competition. Since the agent network is shared (and one might reasonably imagine to be transparently and competitively priced to all providers) then the providers will have to compete on the basis of product design, marketing, branding and operational efficiency. This may well be a novel competitive approach for some players and one can certainly imagine that people more skilled in competing on that basis may well have considerable success even if payments, say, are not their core business.

According to Cap Gemini’s World Retail Banking Report for 2007, the average retail bank account earns only 77 Euros annual revenue, which is barely enough to cover the cost of payments. By comparison, the average mobile operator had a post-paid ARPU nearly ten times as big. Both banks and operators are looking at mobile banking as a way to boost revenue but there’s no obvious advantage to the banks in this space. Operators can be very successful and could use the shared agent model to move beyond their existing distribution networks. Although they don’t do it at the moment, if M-PESA were to extend their pay-in and pay-out channels beyond Safaricom resellers, it might lead to further growth.

I was reflecting on how operators, banks and others might decide to bring payment products to the market and while I was looking over some notes from another presentation in Amsterdam, given by Craig Kilfoil, a former VISA guy, of PayM8, he drew attention to the factors that led to the successful launch of a product in one African country and the failed launch of a similar product in another African country (I won’t mention the countries). The products used exactly the same technology and processes but they differed in the registration procedure. In the first country customers signed up using the phone itself. Whereas in the other country they had to go into a bank. What’s more, in the bank case, the PIN that they had to use to instruct transfers via the handset, was not the same as the PIN they had for the bank debit cards, which led to some annoyance. Naturally, the phone only system was successful whereas the go to a bank system was not. But Craig was wisely asking people to study the less successful project in order to learn more. It does provide a very interesting case study: The figures that Craig shared with us were that the bank had identified customers who would be viable candidates for a mobile payments service and had therefore individually phoned 3,500 of these pre-qualified customers. Of those, approximately 500 had expressed interest in using the product but only a tenth of them (ie, around 50 people) actually got as far as downloading the application into their handset. And of those, only 11 people ever actually used it. I couldn’t help but contrast this with the M-PESA experience in Kenya where customers go to an operator outlet or agent (of which there are about 1,400 in Kenya at the moment compared to the 78 branches of the biggest bank) and rather than download new applications, are simply given new SIMs.

Now this isn’t necessarily bad news for banks because it suggests that allowing operators to provide the payment service through their own network, or in cooperation with an agent network, means that banks can focus on providing actual banking services to a smaller section of the market, perhaps even through the same agent network. So this restructures the relationship between payments and banking.

While this is currently true in countries where banking infrastructure is less well-developed, it may well be that a substantial fraction of the people in developed countries with bank accounts, don’t really need them either and would in fact be much happier with a cheaper, simpler payment oriented account. In other words, a fraction of the customer base is overshot by conventional banking and vulnerable to innovation. But since it costs the banks money to support those customers and they don’t use more sophisticated (ie, more expensive) financial services, this may not necessarily be a bad thing. If I’m the person responsible for allocating resources in a large retail bank, I might well think to myself that if Vodafone can collect 1 euro transactions from vending machines or transfer 12 euros between two individuals or provide a means for kids to buy music online, more cost-effectively than I can, then I might as well let them and use my resources for something more worthwhile supporting another more profitable banking function.

These opinions are my own (I think) and presented solely in my capacity as an interested member of the general public [posted with ecto]