In mid December, I went to an interesting rural mobile banking conference in Chennai, organized by Ashok Jhunjhunwala, attended by about 35 people, primarily  various mobile banking start-ups, banks, a few banking regulators and a couple MFIs. About 2/3s of the way through the conference, Ashok posed a question to the attendees “Do we all agree that ideally, any mobile phone user should be able to cheaply, reliably and easily send 500R (~$12) or less to anyone else, simply by knowing their phone number?”  Everyone collectively nodded their heads. “Then what is the regulatory, legal, technical and business model issues that need to be solved to make that happen?”

In many ways, the first question really is a great vision – primarily because it’s simple and virtually every major player – tech companies such MSFT and Google, the government, the telcos and the banks – could get behind it and publicly support it in principle – an important and necessary step for the cooperation and legal framework wrangling that will lie ahead if this vision is to be made reality.

First, let me state why I think the vision is important (because I’ll admit on the face of it, it seems a little like an obscure wish of telcos and banks hoisted on to unsuspecting phone users).

1. Why the telcos should care: India, like many emerging market countries today, is in the midst of mobile phone explosion in growth. Through some ingenious regulation that set the network interconnection fees lower between competing telcos (more on that later), a rising economy and vast potential market, India has about 110m mobile subscribers and is adding 5mil a month. It also has some of lowest per user monthly bills in the world – about 335R in November 2006 – meaning that the technology really is achieving the twin goals of reaching many, many people while becoming more and more affordable. It’s a wonderful network and self-enforcing effect that leads to high returns for providers and high quality, cheap technology access for many people. Now compare the number of mobile subscribers in India (110mil) with the number of bank accounts holders with access to digital payments systems such as debit and credit cards – 10 mil. What’s odd of course is that mobile phone users constantly convert cash into a digital form – aka talk minutes – that is meticulously tracked (and in places like Africa traded) just like a currency. The telcos in effect have already setup thousands of nationwide “bank branches” in the form of phone-wallas that collect cash from customers and then “charge-up” their prepaid software accounts.
2. Why Google, MSFT and other web companies in India should care:  If you look at many of the most interesting web innovations in the last 5 years – youtube.com, myspace.com, 2GB mailboxes, free VoIP calls, etc. – the business model behind many of these was online advertising, whereby new services and content is given free to end users, in exchange for a fraction of them clicking on ads. Clicking on ads in itself is not actually that valuable to the companies who buy ads though (a common myth I’d argue) – the most expensive ads are in fact the ones that result in likely sales when the user clicks on the ads (e.g. google charges about $7/click to online flower delivery companies in the weeks before Valentine’s Day and flower delivery companies pay it because a significant number of people who click on the ad actually buy $50-$100 goods). These purchases occur almost exclusively through digital payment schemes – mostly commonly credit cards. What’s interesting then is that outside of the 7 most developed countries – which also happen to the places where the most number of people have easy access to digital payments systems – online advertising is virtually non-existent (MSN made about $4 mil in ads in India last year on millions of users). In short, one cannot really have an abundance of free, online content and services unless there are willing advertisers, and a country will likely not have advertisers willing to pay enough unless their target customers and actually turn those ad interactions into purchases. Those purchases are effectively impossible – or simply too expensive – unless there are ubiquitous digital cash systems. Ubiquitous mobile banking services could provide this. 
3. Why the government should care. Cash ultimately is dangerous (it’s easy to steal), decays over time, impossible to trace. Helps overall economic development, enabling the creation of online+delivery only businesses such as ebay, dell and amazon.
4. Why MFI’s and banks should care: Credit is important because it enables people to bet on their future earnings to make investments in themselves. In developed countries, credit is relatively easy to secure – a combination of enforced contract law, strong unique identifiers such as social security numbers – and via these mechanisms, the actuarial science of determining credit rates is well established. In many developing nations, no such enforcement or user tracking systems exist and thus, the trust systems have tended to be based on local social pressures. A villager borrows money from the local loan shark at high rates and if they don’t pay, their family could be in danger. More benignly, MFIs such as Grameen and Unitus have figured out how to rely on the social pressure of neighbors to keep each other honest and pay previous debts on time. The problem with these existing systems is they are not particularly scalable and still requires lots of handling and transport of cash. Obviously, cashless systems enabled through mobile banking could significant help these issues.

(At some point later, I'll flesh out the Philippines as the example.)

Towards solutions that might work:

Likely step 1: Enable cheap intra-operator transactions. 

Keeping the money within the system of one telco (or more precisely inside one bank who has partnered with the telco in a rev share arrangement) allows the telco to earn interest on the float regardless of how customers move money around within the system.

From an end users perspective, it’s somewhat annoying but not entirely un-useful. If my family or co-workers are on the same network – either because we’ve bought a family plan or my company has partnered with a particular telco to provide phones/accounts to all their employees, respectively, I can easily transfer money between these folks. From a business perspective, it’s a little annoying because they must maintain phone accounts on several networks but this may be acceptable.

Hallmarks of a successful strategy here:

• Technical: SMS + GSM SIM card based: taking the Philippines as a model, there are advantages in using GSM – with its interchangeable SIMs that work on any GSM enabled handsets, whereby the banking software is embedded into the SIM and then distributed to every phone on a given telco’s network. Messages are then invisible exchanged via telco’s SMS side channel. This has the advantage that the system works on virtually every GSM handset made in the last 5 years, without requiring GPRS data plans.
• Regulatory needs...
• End user experience needs to be simple...

Step 2 (where we’d like to be): Enable cheap inter-operator transactions.

I really do think that the model of cheap call interconnection fees provides a model here. Some history…

Enable a system where given a mobile number, a bank, telco or company can look-up the associated bank routing information associated with the mobile number. Thus, if user A (Adam) wants to send user B (Barry) 500R, Adam’s phone company – looks up Barry’s mobile number and receives back Barry’s associated bank routing number. Adam’s phone company – likely with an associated banking account – then deposits money from Adam’s account into Barry’s account.

Such a system purposely does not specify any of the technology companies may implement to enable this, nor even which banking systems that could be used (e.g. a given mobile number might have associated VISA/MC and  bank routing accounts associated with it).  It does not need to specify the charges that customers should pay for any part of the chain – it simply specifies that the data look-up to gain associated banking deposit information for a given mobile should be low. There may be significant costs to collect cash and convert it into digital form, just as there may be costs to convert digital cash back into real cash. Each of these are opportunities for new technology solutions and business models – this system just seeks to ensure that the required data to interop between these systems is kept low, so as not create an unneeded price floor on transactions.

Such a system could be a centralized DB, maintained by the government or a mobile consortium. It also could simply be a set of standardized interconnection APIs between banks and telcos.

What else is needed:

• Regulatory concerns
• Everything else...:)

Hopefully, I'll more to say about this in the coming months and weeks - I find it to be pretty interesting space with lots of potential.