Comparing Branchless Banking in Bangladesh and Pakistan
Bangladesh is a recent entrant into branchless banking – deployments only began in earnest in the middle of 2011. CGAP reviewed the first year of branchless banking (referred to as “mobile financial services” in Bangladesh) together with Bangladesh Bank up to March 2012. At that time, Bangladesh had only a year’s worth of experience, but in its second year up through March 2013 Bangladesh has made an 8-fold jump in accounts and agents. What can be learned from benchmarking Bangladesh to the regional leader Pakistan which had begun two years earlier in 2009? Both have similar populations, mobile penetration and income levels making the comparison even more interesting.
First, consider supply numbers. We use the latest figures by the State Bank of Pakistan for the fourth quarter of 2012 (quarterly reports on Pakistan) to compare with figures provided by Bangladesh Bank for the first quarter of 2013. Many thanks to the State Bank of Pakistan and Bangladesh Bank for providing detailed explanations of these figures.
The best measures of total supply come from the volume of client transactions. Total transaction value in Pakistan is $1.5 billion but this includes $646 million of back-end agent liquidity management transactions. Removing these leaves the total client transaction volume at $879 million. This is a more comparable figure to the $1,030 million client transaction volume in Bangladesh.
Latest Data from Central Banks (*includes only client accounts)
(Jan – Mar 2013)
(Oct -Dec 2012)
Transaction Value per quarter ($US millions)*
Transactions per quarter (millions)*
Average Transaction Size ($US)*
Agents end of quarter (000s)
Accounts end of quarter (millions)*
Transaction numbers and average size are also similar. Bangladesh’s 4.6 million wallets is more than double Pakistan’s figure, but this is a bit misleading. EasyPaisa, the largest mobile money service in Pakistan, has 5 million users, but of these 4 million do not have mobile accounts. The gross figures also do not capture that Pakistan provides a wider range of utility bill pay, G2P and other services (airtime top-ups are not a large proportion for either country).
However, on sheer volume Bangladesh’s infrastructure of agents and accounts has surged. The quarterly figures reveal that Bangladesh appears to be growing faster; transaction values quarter-over-quarter grew by 7% for Pakistan versus 57% for Bangladesh.
The Bangladesh growth is driven primarily by bKash, a subsidiary of BRAC Bank, which accounts for 80% of market share by several measures. But Dutch Bangla Mobile is also on a steep growth trajectory. Bangladesh’s largest private bank - Islami Bank - just launched MCash which is poised to make a large contribution.
This growth has put Bangladesh onto a scale and trajectory close to that of Pakistan. It is less important at any point in time which is bigger or growing faster, as both markets today are among the global leaders. It is more insightful to learn from a comparison between the two.
What common factors contribute to the rapid scaling in Bangladesh and Pakistan?
One commonality is that providers are large companies expecting to make large investments. In Bangladesh, BRAC Bank (majority owner of bKash), Dutch Bangla Bank and Islami Bank are among the leading retail banks. In Pakistan Tameer/Telenor’s EasyPaisa, United Bank’s Omni, Mobilink’s MobiCash, Askari Bank’s TimePay and Habib Bank’s HBL Express have all large institutions behind them. CGAP estimates aggregate investment in Pakistan of $104 million so far. Another commonality is that there is one larger early provider generating most of the supply: EasyPaisa in Pakistan and bKash in Bangladesh.
Another commonality is that both Pakistan and Bangladesh leverage existing agent networks to gain scale. This has required adding on to existing ecosystems through local distributors. This leverages “middle men” who already distribute airtime or fast-moving consumer goods.
In both countries agents are non-exclusive. Early data from a Gates Foundation agent-mapping survey indicate that one-fifth of agents work for multiple providers in Bangladesh (though data collection is still in process). Estimates from Pakistan suggest the proportion may be a bit lower. Agent exclusivity has not been necessary to promote early scale. However, some in Pakistan hold that non-exclusivity of agents is slowing further expansion. In Bangladesh this has not been a topic of much discussion yet.
The largest uptake in Pakistan is from over-the-counter (OTC) transactions where senders and receivers remit through agent accounts. CGAP research and a snapshot survey by Pi Strategy Consulting indicate significant OTC uptake in Bangladesh too – probably more than half of all transactions are effectively OTC. A non-account OTC “token” initially offered by bKash has been taken out of service. However, transactions have shifted to another OTC approach: senders hand cash to an agent who transacts through a customer and agent account enabling the receiver to obtain cash at another agent location. These payments happen in 5-6 different ways as negotiated between agents and customers, but all transactions go through KYC-checked customer and agent accounts. Therefore, CGAP does not believe there ought to be regulatory concerns with this practice.
That said, OTC is a potential diversion for both Pakistan and Bangladesh. It falls short of customers independently using their own mobile accounts – a behavior that will improve client retention and open up the possibility of additional services. Providers in both Pakistan and Bangladesh recognize this and have their own business reasons to shift to direct customer mobile wallet use. It will be instructive to watch how both markets progress on this transition.
What are the key differences between Bangladesh & Pakistan?
The similarities are striking, but there are underlying differences as well. In Bangladesh the two leading providers have access to the USSD channels that can reach 90% of mobile subscribers. In exchange for USSD access MNOs typically receive 5-20% of revenues, but MNOs are not driving the business. Some hold that, under current regulation, MNOs in Bangladesh could play a more active role through partnerships but so far such partnerships have not been fully developed.
In Pakistan the two MNO deployments are only able to reach off-network subscribers with OTC services – contributing to the smaller mobile wallet figures. This points to another difference in Pakistan with more onerous account opening KYC requirements. Even the lowest-KYC accounts (level 0 and level 1) are less than half in number of what Bangladesh has reached in a shorter time span. Bangladesh’s visually confirmed voter identification card and paper account opening process facilitate faster account growth. Pakistan is also limited in the number of agents who can open accounts (only 27%) – no such limits in Bangladesh.
In Pakistan MNOs lead two services, EasyPaisa and MobiCash. Both naturally prioritize mobile wallets among their subscribers – a quite different flavor than Bangladesh. Besides these two, the other three large deployments in Pakistan are led by banks. The fact that there are 5 large players in Pakistan makes it a more competitive market. This has also led to some early discussions about inter-connecting services though it is now generally considered premature. In Bangladesh inter-connecting services has not been widely discussed, but may arise down the road.
Concerted government action in Pakistan to channel government payments (G2P) to citizens over branchless channels has been a significant motivator for large companies to enter. Bangladesh’s government is testing G2P delivered digitally including the Employment Generation Program for the Poorest (EGPP) with the World Bank. But the Bangladesh government has not committed to a government-wide shift which could generate even more interest from providers.
With only 2 and 4 years experience respectively, it remains early days for both countries. Yet it is not too early to suggest that other countries may do well to emulate much of the early progress we are witnessing in both Pakistan and Bangladesh.
------ The author is CGAP’s regional representative for South Asia.
Greg, great article. I believe that BEFTN could be a good vehicle to inter-connect the services. Also, I am really happy to see significant progress in Bangladesh.
The study finding is interesting. It would be more interesting if the socio economic status of these clients of branchless banking in these two countries where more unbanked poor people live.This has also more relevance in the context of importance of client focus in the demand side being recognized in CGAP
Dr. Rengarajan, you are correct that little is said about who uses these services. In fact, in Bangladesh little systematic data is available on this yet. So we hope that new research will offer more data in the coming months. Greg
With regard to OTC approach being practices in Bangladesh whereby money is transferred with an agreement between the customer an agent. Don't you think this exposes vulnerabilities and possible fraud; as account belongs to an agent and he can choose to give money to the intended recipient or not.
Imran, one important thing to remember about an OTC transaction is that the sender and receiver can still pick the agent they want to use. Or to negotiate a way of sending together with those agents. Clients can also choose to send money in other ways via a courier or just as cash. Given all the choices agents have an incentive to provide reliable regular service and to keep up their local reputation. Often an agent is someone the client already knows well. Greg
Thank you Greg Chen for your positive response. Yes fresh research is welcome as it has topicality and relevance too in the context of CGAP focus and priorities in the demand side. In this connection following research questions framed out of the contents of various CGPA blog/linked in groups postings, may be useful
That is to probe from sociological perspectives, who ultimately gets benefit in the pyramid in this kind of tech based branchless banking in poverty rich countries like Pakistan and Bangladesh? Does mobile enabled product adequate to serve all the financial needs of community including the poor?
From institutional perspective, Is mobile based branch less mode more efficient in the inclusion of excluded or unbanked in the last mile ?
From economic perspectives, “Does mobile magic create a separate mobile adopters /users group separately and concentrate in business /trading community? Does it deserve priority agenda for the socio economic progress of these nations ?
From ethnographic research perspectives these queries are raised in the context of some negative fallouts of this mobile magic in community welfare in Africa where 34 out of 47 countries with mass market adoption as reported below.
The past experience in M-Pesa and Jipang ku sav (CGAP bog post ) admittedly reveals that
“of course, the poorest of the poor will always be the hardest to reach and we should continue to think about responsible and effective ways to reach them.( Sarah Rotman’s response in the CGAP linked in group (oct 2012). That indicates bottom poor are hardly reached and only small trading/business community is benefited. In that case latter in a community may be viable for mobile business rather than outreaching the vulnerable bottom.
Another posting in this linked in group “ Kenyans find unintended consequences mobile money” has been discussed highlighting the negative points such as rich-poor divide , from the perspectives of excluded bottom .
Again in this linked in group postings, Jamaes Pearse has posted “ Barriers for mobile money in Africa” which reveals among others, that in mass market adoption, “ 44% of non users in Sub Saharan Africa don’t use mobile money because mobile money they think it is not secure” and its limitation “one size does not fit for all” with P2P service alone
All these observations on technology intrusion in the process of financial inclusion drive home a point that the progress need to be more socially and ethically evaluated more in terms of its outreach in the pyramid and welfare outcome in poverty sector without creating inequality. This is not necessarily against mobile money but there are concerns on equity in distribution of tech benefit and tech divide in the society particularly in poverty sector and faster inclusion without productive absorption in economically backward regions.
Thank you for sharing my views
Dr Rengarajan .
Greg, very interesting statistics !
Hi Greg, great article, the MNOs point that you made as "MNOs in Bangladesh could play a more active role through partnerships but so far such partnerships have not been fully developed." I just wanted to add on that, hope you would reconsider.
Airtel Bangladesh limited tied up with Dutch Bangla bank limited and they are not only providing USSD and the retail channel support, moreover airtel has been working hand in hand with Dbbl to conceptualise and execute new products such as m-salary, m-donation, micro credit disbursement, vendor payment etc. Both airtel and Dbbl is working with various new products which will be launched soon to enlarge the ecosystem.
Faisal - Good to hear about the new partnerships under development between Airtel and DBBL. This is a good example of building out the potential of such partnerships and we look forward to hearing more as these roll out in the months ahead. - Greg
Good insights Greg.
OTC transactions have shaped the mobile banking in Bangladesh and Pakistan. But it has side effect too – this slowed down the increase of number of accounts and created issues for the innocent agents. A few days back one extortionist demanded Taka 300,000 from a father against release of his (father’s) son. Father paid money through a few agents’ accounts and after obtaining his son, filed FIR with the police station. Police arrested the agents as virtually they have received the money. The actual receiver of money is nonexistent here. There are lots of similar cases filed with the police stations. If such is the case, should the OTC transactions not to be a concern to the regulators, specially to the Central Bank, BTRC and the Home Ministry?
I agree that if such things happen with ever greater frequency that this should be reason for some attention. And that attention to this might find reasonable ways to hold agents accountable for their participation in any such schemes. The response to this kind of activity needs to be balanced in proportion to how frequent such abuse is and also balanced with the potential benefits some see in using the OTC service as a legitimate means of payment. - Greg
Good Article Greg ~.~