Mobile Wallets: Is a Transition Underway in Bangladesh?

28 October 2013
12 comments

Let me begin by introducing Mushtaque (not his real name), a rickshaw driver in Dhaka, Bangladesh’s capital, who I’ve known and ridden with for a few years. When I saw Mushtaque last month he had been operating a mobile phone wallet for 5 months. Over that span he had saved up $75; equivalent to half the price of his rickshaw. This is an encouraging story, but I will return to Mushtaque and show why his wallet use is not as simple as it might initially appear.

A man fixes a wheel on a rickshaw
Photo Credit: Aj Ghani

In Bangladesh, over-the-counter (OTC) transactions usually happen as a person-to-person transfer arranged between a sender and receiver. The sender and receiver each use an agent to complete the electronic transaction. The two agents involved can make these transactions happen in a variety of technical combinations (most often as cash-in transfer of value to an account or a cash-out from an account). Such an approach requires an agreement between four parties: sender, sending agent, receiving agent and the receiver. We do not understand well why this appears to be quite common in Bangladesh.

Nadeem Hussain recently blogged about the unexpected use of OTC transactions in Pakistan. Bangladesh too has its unexpected share of OTC transactions, but It is difficult to estimate the precise proportion. After research conducted this spring, Pi Strategy estimated that at least half of Bangladesh’s mobile transactions volume was via OTC. Indeed, use of OTC transactions is significant enough that Bangladesh’s central bank issued a circular in September discouraging mobile payments by individuals without a mobile account. But it is likely that market dynamics will shift in Bangladesh, now in its third year of mobile banking deployments, especially as the use of e-wallets continues at a high pace - maybe even reaching 10 million by the end of the year, or nearly 1 for every 15 inhabitants.

Pial Islam of Pi Strategy in Bangladesh argues that the time required to open an e-wallet account is a big part of what drives people towards OTC transactions. Another factor is that the OTC transaction via mobile is similar to the already existing pattern of payments that occur through couriers or post offices: clients likely see mobile as another way to do the same thing. Finally, agents just seem to like doing OTC transactions. But we do not fully understand the drivers.

The central bank of Bangladesh and its mobile financial service providers in Bangladesh are pushing for a transition where existing OTC users shift over to pure e-wallet use. The aim is to do this with a variety of mechanisms: explicitly clamping down on agents offering OTC, opening new e-wallet accounts, and continuing a customer awareness campaign on maintaining the privacy of transaction. CGAP, along with with Pi Strategy, is undertaking research to understand why such transactions are happening and who uses which kind of service.

But let’s come back to Mushtaque the rickshaw driver to see how his experience may explain some of the challenges of e-wallet use. He is illiterate and lives on a near subsistence livelihood from his physical labors. Yet he has an e-wallet and uses it to accumulate savings. But his use of that e-wallet is complicated. Mushtaque had to be persuaded by friends to open the acount (admittedly I also encouraged and cajoled him). He did not immediately have the necessary ID nor could he complete the account opening information. It took several days of effort to come up with his ID and a friend willing to help him fill out the paper work.

Transacting also requires a serious effort. Since Mushtaque can’t read, he lacks the confidence to manage his own account. When he wants to deposit money, Mushtaque visits the agent across from his residence and hands the agent the cash and his phone number. Upon cash-in Mushtqaue does receive an SMS confirmation, but my impression that his trust in the agent is more important. After all, he can’t read the SMS. Cashing out is still more complicated for Mushtaque because he must then enter his PIN and the agent’s mobile number amid a string of instructions. He may not even remember his PIN. For these reasons Mushtaque won’t cash out on his own. Instead, he calls a friend to come and help. Mushtaque hands his phone to the friend who deducts the electronic amount from Musthaque’s account and hands the cash to Musthaque. He could also simply hand his phone to the agent, which I suspect happens as well.

The reality is that many poor like Mushtaque are likely early on to use their phone in this way if they are to have a wallet account at all. Mushtaque is not only illiterate but he is also physically and mentally taxed by his daily work life, and he has little patience to adopt new technologies. He could do it, but it is hard work. Many like Mushtaque may simply use mobile phone payments service via OTC transactions, which removes the sheer effort associated with learning a new service and OTC happens fast. How many other Mushtaques are there who do not have the support mechanisms necessary to establish an e-wallet? What technology changes or support might persuade them to try out a wallet out? Critically, Pial Islam observes, “how do we make e-wallets attractive that the poor and illiterate will find having one critical enough to overcome these hurdles?”

For mobile to achieve its full potential, markets like Bangladesh are a critical testing ground.

I heard some encouraging news from Mushtaque the other day. He had saved enough to buy a cow for his brother in the village. To get the cash to his brother, however, he again had to involve a trusted friend to deduct the amount from his e-wallet and send it (via OTC) to an agent near his brother’s home.

Comments

Submitted by Mosiuzzaman on
Another interesting data found in p2p transfer. We have analyzed 10k transactions of 20 agent points, 500 each and found that 78.6% were P2P transfers and rest were Cash Out. However, when we talked to those agents, they told us, real cash out transactions were double, but it is done by P2P. Meaning that, agents maintain multiple personal/customer wallet to provide cash out service. When a customer comes for cash out, agents tell them to transfer the money to their personal wallet,[not to their agent wallet] which costs only 5 taka but they are taking service charge from customer in full. They transfer this money again to another customer for Cash In. ex. For cash out of 5k, customer is paying 90 taka(considering 1.8% as cash out charge) but agent costs is 5+5=10 taka(for two times p2p cost), a total 80 taka profit. Since BB has set some Cap to personal wallet, agents are maintaining 20-30 wallets even to do this job. This is nothing but a loss for banks which should be taken care immediately.

Submitted by Mosiuzzaman on
Interest on Wallet balance will decrease OTC transactions by encouraging unbanked people to open mobile wallet and keeping balance in it. No. of transactions and volume can also be encouraged by giving reward bonuses. May be, these facilities are coming in near future for end customers.....

Submitted by Abul Kashem Md.... on
Dear Mr. Greg, thank you for interesting article. You have mentioned that Mushtaque had saved up $75 in his wallet; equivalent to half the price of his rickshaw. Although it was a difficult task for him to open a wallet meeting all the compliance, due to his wallet he could saved money and bought a cow for his brother. And I am sure, one day, he would be able to buy a motor-rikshaw from his wallet savings! If he used OTC services only, in which case he had no wallet, how he would be able to save money? That means OTC transaction can’t help him to materialize his dream! OTC can be used for sending money and as you know, in Bangladesh, there are many means and ways to send money – either formal or informal. Then why banks being a sophisticated business entity with hill-size trust of the customers should enter into this narrow business? Banks should always help customers grow savings.

Submitted by Peter Goldstein on
Greg: Very interesting post. I am wondering about comparison to other countries where issues of illiteracy and education deficits are also prevalent, such as in African, but where wallet usage rates tend to be higher. Might is not be worth a comparison to see how the dynamics differ from country to country? Peter

Submitted by Louise on
Illiteracy levels hinder the adoptability of some technological innovations but I believe making e-wallet as simple as possible can encourage adaptability. With few requirements upon subscribing and making the menus simple and easy to understand.

Submitted by Mosiuzzaman on
Thanks Mr. Louise. Yes, Illiteracy is hampering in adoption of mobile banking service. But, I think IVR base solution & easy to use mobile applications based on symbols/signs can be helpful for uneducated people to enjoy mobile banking service.

Submitted by Chowdhury Mohid... on
I would like to convey my thanks to Greg for writing such a good piece of Article based on personal contact and experience. OTC transactions are very common in P2P transfers. I have recently tried to verify this phenomenon and found that some Agents are unwilling to make OTC transfers and insist for beneficiary's Wallet ID. It is a good sign. Agents have started to understand the rules of the business and enforcing this. But still, I agree with Greg, lot of Agents are resorting to OTC transactions. Service Providers should now start monitoring the activities of the Agents.

Submitted by Raihan Alhusain on
Thanks for sharing the great article. These barriers/ Hurdles are very common in a country where illiteracy rate is much higher. However; through proper customer education, awareness and also bring technological ease in the transaction we can overcome this. The Key agenda of Mobile Banking is Banking the Un-banked and OTC transactions are creating a glitch in the execution of the agenda.

Submitted by Greg Chen on
Dear Raihan, thanks for your comments. While I agree the goal is to help the Un-Banked I wonder if forcing each and every one of them to use a mobile wallet is the right approach. Instead of thinking only about being "banked" we might think instead about how we make the services relevant to customers lives; some may want to be banked in very different ways than we imagine. Greg

Submitted by Delwar Hossain Azad on
Greg, very well written in very easy to read and understand. One big missing point is that most of the OTC transactions are driven by the Agents maintaining multiple personal wallets (15-20 in some cases) from where they do transaction between agents on behalf of the customers. The Agents publicly display or will tell you to which wallet one should bring or send money to. Some service provider doesn't charge for Cash-in which makes the matter encouraging as well. For longer term business interest and service portfolio increase will depend how well service provider is dealing this unscrupulous practice rather than looking at short-term gain.

Submitted by Greg Chen on
Azad, while I share your view that wallets should be promoted with agents and clients, I also wonder if there are clients for whom wallets may never be attractive. Some portions may prefer the OTC option. Even Rickshaw driver Mushtaque has a wallet, but still uses OTC for some payments. Should we prohibit this? Greg

Submitted by Azad on
Greg, your observation is so right and that's why we have been talking on "proportionality" in KYC & AMLA bla bla if we really mean to attack financial inclusion. It's so difficult to understand or even accept any rational behind a blanket KYC & AMLA for all without any consideration that most of the population is either illiterate or semi-literate. The need of that segment is "small ticket" and in very "few in number" which can easily be catered through appropriate regulation. We all have to understand and accept that there is NO PERFECT WORLD and calculated & predictable risks taking is a must to accommodate the needs of the market. Hidden/Underground world is more dangerous and harmful for the economy than allowing some imperfection that can be traced and predicted.

Add new comment