There’s a limited (if steadily growing) list of pioneers among mobile banking implementations reaching lower income clients…GCash, M-PESA, Wizzit, etc. Most of the buzz focuses on how they differ from traditional banking. But we realized we knew very little about how these pioneers rate against each other. Pricing seemed like a good place to start, since we’ve been itching to better understand that.

Sarah Rotman and Mark Pickens of CGAP’s Technology Program pulled together a pricing table – perhaps the first of its kind publicly available – comparing the prices of 6 branchless banking pioneers: GCash and Smart Money in the Philippines, M-PESA in Kenya, WIZZIT and MTN Banking in South Africa, and Tameer Microfinance Bank’s pilot with POS terminals in Pakistan. We also put them head to head with the “big four” banks in South Africa, where good pricing data was available. We eliminate differences across countries and currencies using the World Bank’s latest purchasing power parity figures.

This is a work in progress, and comments are welcome. What we found was unexpected: at least one freemium mobile banking pricing scheme (read on to find out who). And while our research confirms the common view of these pioneers as cheaper than banks, we discovered some may not be affordable to the poor. This (a) raises some key questions about uptake among lower-income, mass market clients, (b) shows industry is still very early in understanding how to price m-banking, and (c) ought to give regulators pause if they’re considering imposing a price cap.

We tend to think of the early mobile banking implementations as cut from the same cloth. But their pricing schemes are very different. MTN Banking charges 10 times more (USD 6.15) than Smart Money (USD 0.61) for a basket of 12 transactions that meet a household’s financial needs for a month.

Most intriguing was GCash, which looks like a pure freemium pricing model. A client doing a USD 25 money transfer and a USD 5 airtime transfer (consistent with remittance and airtime spends for low-income clients) will find they actually pay nothing. The 10% discount Globe Telecom gives to clients who buy airtime via GCash covers the full cost of the transfer. Mobile banking for free? Not for everyone: true to a freemium model, prices ramp up quickly for more intensive usage of the GCash wallet. But even at a fee of USD 1 for the bigger basket of a dozen transactions, GCash joins Smart in making the Philippines consistently the cheapest among their peers.

(Affordability varies by number, size and type of transactions. See the end of this post for an explanation of the 3 transaction profiles we constructed: (1) a typical user doing just several transactions per month, (2) the basket we found with WIZZIT clients in 2006, and (3) a user doing 12 transactions per month, which we refer to as the “full service basket” since it attempts to reflect a usage level if an unbanked person used a branchless banking account to meet their family’s full financial needs of checking account balances, saving, sending remittances, paying bills, making store purchases.)

While a range of prices prevails among them, the pioneers are still cheaper than traditional banks (if our look at South African is any guide). At the widest margin, Smart Money from the Philippines was 15 times cheaper than Standard Bank, the most expensive bank we looked at in South Africa (since good data is available about prices charged by the country’s Big 4 banks – Absa, FNB, NedBank and Standard). At the full service basket, all of the branchless banking pioneers were cheaper than the Big 4, with the exception of MTN Banking. MTN comes in higher than its peers due to the high fee it charges to use Standard Bank branches. This may very well be a legitimate business decision to steer clients clear of higher cost channels. But if low-income customers really do value face-to-face transactions, as we concluded from CGAP’s 2006 South African research, they will certainly have to pay for it at MTN. South Africa has pretty high bank fees. We would love to get more data about bank prices to run this comparison in other markets.

But while they beat the banks, a poor person won’t always find these 6 pioneers affordable - in fact, they can be very expensive. We wanted to see which providers will be cheap enough for the poor. Again, we looked at South Africa, using a FinMark Trust definition of affordability (2% of household income) applied to households at a national poverty line of R430 / month proposed by the National Treasury. 58% of South Africans fall under this poverty line, which equates to USD 55.51 or about USD 2 / day. For a client using the full service basket, none of the South Africans — WIZZIT, MTN, or the countries Big 4 — gets close to meeting the threshold for affordability. This surprised us. Fees ranged from 7.6% to 14% of the poverty line, or about 4 to 7 times greater than what would be affordable. Prices are still 3.5% to 6.4% of the poverty line with the somewhat smaller basket of 9.2 transactions from the 2006 WIZZIT user profile. Employing a higher national poverty line would shift this up somewhat, but we don’t think it will change the outcome. However, the operator models from other countries are more affordable for this basket. M-PESA would cost a South African on the poverty line 2.48% of their income. That’s getting within striking distance of the 2% threshold. But the Filipinos are still the kings in this category: Smart Money would cost just 0.82% of income, and GCash would actually be… free.

What are some lessons suggested by this first look at pricing?

First, if we know m-banking can often be cheaper than traditional options, or even downright free, why aren’t more poor people leaping to sign up? Clearly, pricing is just one element driving customer adoption. You would expect the importance of price to be inversely proportional to client income. So poorer clients should find cheap or even free services very attractive. But while several million Filipinos and Kenyans are using mobile financial services, there are many millions more who aren’t. What else beyond price is preventing adoption curves from getting steeper? We need to look into this.

Second, the range of pricing schemes suggests there’s still a lot of work for providers, consultants and advisors to test what’s optimal. Pricing drives revenue projections and also affects expected uptake, in turn shifting payback periods for a firm’s investment. Our research showed that providers still have very different thinking about what to charge and how to charge it. For example, GCash clients pay a 1% flat fee when they convert cash to or from electronic value. By contrast, M-PESA clients’ largest cost is incurred when they initiate a transfer. As we noted here, we found M-PESA prices are optimized for remittances, but not for smaller transfers like microloan payments. As a result, M-PESA is forgoing much of the business of MFI clients. That’s just one example. This first cut at comparing pricing left us thinking there’s still a lot of experimentation to do with product terms for branchless banking aimed at low-income segments.

Third, the diversity of pricing ought to give regulators pause if they are considering price caps. It’s probably too early to know what prices ought to be. Even if a cap is acceptable to one player in the market (say it’s the first to approach the central bank for permission), others in the next wave may find different pricing is required to make the business case. Take GCash and WIZZIT. Globe is probably willing to offer GCash for essentially free to many of its customers since the 10% discount they offer on airtime saves Globe an even larger amount on commissions they otherwise pay to stores selling Globe airtime. By contrast, mobile banking is the main business for a (relatively) cash-poor start-up like WIZZIT. While WIZZIT explicitly targeted the unbanked from Day 1, they cannot afford too many loss leaders among low usage clients. So they charge more than all the other pioneers we looked at, at least for the basic basket of 4 transactions.

The lesson: A regulator which had set a price cap based on what a mobile operator was willing to charge might find they lock out a specialized company like WIZZIT. Regulators might be advised to steer clear of price caps at this early stage of market development, since we still know little. But if making sure clients are not being gouged is important, ask applicants to benchmark themselves against other services in the market and prove why they ought to be allowed to charge a certain rate.

Finally, a bit more about how we went about the research, for those interested in our methods. We plugged in 3 user profiles: (1) a low-intensity client who looks like what a lot of the anecdotal evidence tells us is typical: 1 cash-in in order to make 1 balance inquiry, 1 remittance and 1 airtime purchase per month, and (2) a high-intensity customer who treats the service more like a full service bank account, doing 12 transactions (2 of the the above + a purchase, bill payment, and treating it like a store of value (cash-out). We also plugged in the transaction profile of the average WIZZIT user from CGAP’s 2006 study (done in partnership with Vodafone Foundation, UN Foundation and FinMark Trust). You can see how the 6 pioneers stack up on price in the first tab. We adjusted all prices for purchasing power parity using the World Bank’s latest figures, so it’s possible to compare between providers in different countries. Then we dived into one country example: South Africa. On the second tab we look at how the 6 pioneers rate against the cheapest, full service bank account offered by the country’s Big 4 banks (using Bank Monitor data). We hope to replicate this comparison for other countries.