Why M-PESA should offer savings accounts

by Olga Morawczynski : Monday, September 28, 2009

A recent CGAP brief on M-PESA, which I co-authored with Mark Pickens, revealed an interesting finding on savings. Nearly one in three of the informants in Kibera (an informal settlement in Nairobi) kept a balance in their M-PESA accounts. A survey conducted in 3,000 homes in Kenya by FSD found this number to be even higher. A majority of users stored money with M-PESA. Some used the stored value for daily consumption. Others made frequent deposits of “small money” to accumulate a larger amount.

The fact that M-PESA is being used for savings illustrates the latent demand for appropriate savings products in Kenya. By appropriate, I mean accessible, affordable and safe. M-PESA scores very well on two of the three. In regards to accessibility, there are about 11 times more M-PESA agents than there are bank branches in Kenya. With no monthly or maintenance fees, and free deposits, M-PESA is much more affordable than most other savings services. However, it was not designed to be, nor is it regulated as, a savings mechanism. As a result, there are risks involved with this type of savings; liquidity being the most pertinent. If the network goes down, or the agent is out of float, customers cannot access their cash. Despite these inconveniences, many continue to use M-PESA to store their “small money”. My informants explained that the benefits outweigh the costs. Some also made clear that M-PESA was actually safer than the banks. Given the history of bank crashes in the country, this is no surprise.

Users did have one complaint—M-PESA does not pay interest. Since Safaricom is not registered as a financial service provider, providing interest on savings is not an option. This is a shame. The electronic value stored in M-PESA is backed by cash in the Central Bank of Africa (CBA). This account is generating a significant amount of interest for Safaricom. What the mobile operator plans to do with this money is yet to be seen. But one thing is clear—it should be poor Kenyans that benefit from the interest on their savings and not Safaricom. For this to occur, M-PESA wallets should be linked to accounts managed by formal financial institutions. This is something that Zap, Zain’s mobile money service, has already done. Poor users could then use their M-PESA wallet to accumulate cash, as deposits are free. The accumulated amount could then be transferred to the formal account, which provides interest.

Why Safaricom has not made the move towards the mobilization of savings is not clear. But it should. This would make it easier for Kenyans to reap the benefits of formalized savings by having an account that is affordable, accessible and pays interest. It is not just Safaricom that has shied away from savings. This year’s CGAP/GSMA Mobile Money Market Sizing Study claimed that by the end of this year there will be 120 mobile money deployments (live or in pilot) around the world.

Mobile operators are not launching savings products due to business model and regulatory constraints. Most will focus on developing their core mobile money services before supporting cross-domain business models with micro-deposit institutions including microfinance institutions and savings and credit cooperatives (thanks to Mike Catalano, of OpenRevolution, for making this point).

If this is the case, then it could be a while before we see the mobilization of savings via the mobile. This is unfortunate. There are currently billions of dollars stored under mattresses in the developing world. This means that the market for savings could be just as lucrative as the one for remittances. Perhaps the mobile money industry needs to shift its focus.

Comments: Comments and trackbacks are open.

15 Comments RSS 2.0

  1. September 29th, 2009 at 3:38 am, Kenya1 ()

    I was also thinking it is time the mobile operators start giving loans against the deposits. The “hot loans” could be tied to the Savings on the mobile accounts. For example you won’t be able to make withdrawals when you are servicing the loan. Tell me if the poor woman who sells vegetables in the market could get small loans from the mobile operators, won’t that be real nice?

    Kenya1 is a mobile content specialist and social changemaker in Nairobi, Kenya

  • September 29th, 2009 at 7:55 am, Ravi Rajagopalan ()

    Hi all, It is entirely possible that telcos will stay away from becoming financial services providers in the fullest sense. My work with regulators in India and elsewhere, as well as with some big telcos inform me that regulators could get very serious about the systemic risk posed by a non-financial player in payment systems. Especially since there are huge gaps in KYC data availability to formal financial systems, that in themselves present threats to countries faced with domestic or international terrorism threats. Look at the approach taken by Bank Indonesia in this regard. For a telco to become a financial services provider taking deposits and paying interest requires partnerships with banks AND a serious commitment away from the core business. I am not very sure whether in the current climate, this is a commitment telcos are willing to make. I am willing to be corrected.

  • September 29th, 2009 at 8:26 am, Maja ()

    Hi Olga,
    A great piece – thanks. Being completely new to the M-PESA discussion, please forgive my ignorance; I’m wondering what you think about the comment made in last week’s Economist article (http://is.gd/3MClg) that “some banks [...] have come to see mobile money not as a threat but as an opportunity, and are teaming up with operators.”

    Is there a realistic chance that operators will partner with banks to offer more user-friendly savings services? And also, (how) can users play an active role in product innovation in mobile banking?

  • September 29th, 2009 at 10:16 am, Andre ()

    Hi

    The reason why there are no fees on these accounts are because a back-end account is not needed. If you add in a back-end account on each user to calculate interest, one automatically incurs fees. So it is a sacrifice on one end to help on another. We process Prepaid cards and the same argument goes – it works off one central settlement account and we process the fees off the cards and then settle via Visa or MC at end of day. There is no (affordable) way to keep track of each account balance on a daily basis. But our fees are way below what a bank account would charge yet still giving full functionality to the end user.

  • September 29th, 2009 at 1:20 pm, Mark Smith ()

    Great idea, but how feasible is this really in practice? That is, given all the regulatory barriers.

  • September 29th, 2009 at 6:23 pm, Peter Goldfinch ()

    It seems as if Safaricom and the Kenya regulators will soon each a point where some serious decisions will need to be made about M-Pesa’s future. There is obviously a market demand for banking services but the existing banks for some reason or reasons find it an unattractive business segment.

    For Safaricom to move into offering savings accounts and then loans etc comes across as being a little reckless unless they fully understand the bank’s collective reluctance to meet this market demand. This could be a simple case of the banks inability to service this market cost effectively.

    I can hear people saying that banks are conservative, lazy, unadventurous, have no interest in their customers etc etc. But we have seen many disasters or near disasters when the financial sector has become more entrepreneurial, willing to take risks. The current credit crisis is a long way from Kenya but it is an example of what can happen.

    The answer may be that the Kenya regulators need to define a type of banking institution, which is specifically structured to meet this market segment. Built around an extended M-Pesa service with regulations and mechanisms for prudential supervision that save guard the interests, (money) of the customer base. We have seen in other countries such institutions, as community trusts that service the banking needs of a community. These organizations are normally closer to their customers, more flexible in their thinking and operate off a lower cost base.

    M-Pesa is proving not to be an end in itself, but a starting point on what will be a long journey. I hope somebody has mapped out the journey.

  • September 29th, 2009 at 11:44 pm, Anuj Sharma ()

    There’s lot much discussion about MMT and involvement of M-pasa, MFIs etc who slowly and steadily going towards a formal banking operation.
    Why not think about a scenario wherein a big bank through their hived off subsidiary start full throttle with this idea (it has actually started) -wont it make a bigger and lasting impact? Obviously this is not a debate to be triggered about the mode of collateral in place but this is much safe scenario and good KYC adherence.

    All in all one thing is clear Technology is making lives much easier to all in some or the other way. Appreciate this discussion and enlightening us. thanks

  • September 30th, 2009 at 12:32 am, Ragavan V ()

    Hi all,

    Safaricom has done a splendid job by creating an ecosystem which seems self sustaining. In creation of such access infrastructure there are definitive initial & recurring costs involved.

    At the same time if for a minimal incremental cost ( choice can be given to customers) or if the existing buss model allows interest to be paid to customers then it would be wise on the part of safaricom to look at the long term benefits of paying interest than short term gains.

    Additional Cutomer acquisition numbers and retention will help justify the outflow.

  • September 30th, 2009 at 2:52 am, shaun Campbell ()

    Hi Olga, I read the post and spotted a number of points that i wanted to raise.

    CBA is not the Central Bank of Africa but a retail bank called Commercial Bank of Africa. MPesa don’t get the interest (or treasury income) on funds as the money is not theirs (it is the user’s money) and so the money is held in a Trust account on behalf of the user. In fact the bank keeps the Trust account interest!! There is always a perception that mPesa or Zap make a fortune from interest and treasury – they don’t as this is not allowed. Now it may be allowed in the future is mPesa or Zap (for example) were to match deposited funds and capitalize to such a degree that they would be trading treasury funds as a minority % of actual matched deposits. it is a question of risk management and this policy is controlled by the central bank.

    To enable payment of interest mPesa would need to do what our new Kenyan project Mobikash (+ Zain to some extent) have done – fully, seamlessly link a savings, loan or cheque account to the SVA (with MFIs, banks) , or apply to the regulator for a deposit taking licence, or buy a bank outright. Both of the latter options are expensive.

    mPesa do allow money to be transfered to MFIs in a crude manner in Kenya and Tanzania and have done for some time now. However, it is not slick and slow to clear (2 days). Our projects are delivering SVA to account transfers as fast as RTGS.

    The big market opportunity is tying this whole thing together in one seamless independent national ecosystem – exactly what we are working on across Africa with 4 – 5 major regional rollouts.

    I have a few workds to say on this one “”If the network goes down, or the agent is out of float, customers cannot access their cash.”" Well, the fact is that mPesa does go down and customers are indeed left stranded without access to their money, or recipients waiting for money that doesnt arrive. Sadly MPesa is well known for being out of service (3 whole days in July) and it is a topic of widespread concern (re trust) amonst the wananchi and the press. This is not sustainable. mPesa is a great product – no question, but they will never attain a banking standard operation based on their reliability track recored. Imagine if Citi or Barclays was completely down for 3 whole days, everywhere! There would be a regulatory clamp down and battle of mammoth proportions.

    Another point of perception that caught my eye is – “”Some also made clear that M-PESA was actually safer than the banks. Given the history of bank crashes in the country, this is no surprise”" is not really accurate any more. Kenyan banking is now acclaimed as being as robust as European / US banking systems. In fact there was an article in the news last week by the World Bank I think saying that the standard of management and control in Kenya is world class. But, mPesa money is stored in trust by at least 4 Kenyan banks, not in mPesa’s pocket, so there is a misconception here. Significantly more chance every day of mPesa system problems (see above) than a bank failing and losing your cash. The government won’t allow that to happen.

    Your last point is exactly what we are doing across Africa – tying all this together, offering intrerest on money in mCommerce ecosystems and a whole lot more than the MNOs offer, because we can focus, move fast, offer niche solutions to stakeholders and operate independently of MNOs and banks etc = a great advantage going forward.

    “” If this is the case, then it could be a while before we see the mobilisation of savings via the mobile. This is unfortunate.”" – we and our clients are already in this space and about to go public on these services.

    How is Edinburgh? When are you next in Kenya?

    Cheers, Shaun

  • September 30th, 2009 at 3:25 am, Toffene B. Kama ()

    Hi,
    Paying Interests is a tempting shortcut and above all, remains a matter of benefits (real/perceived) VS costs (direct/indirect).
    From a very high level perspective, adding interests in existing mobile wallet deployments would mean, at least :
    - Updating millions users profiles with the Interest => DB migrations/services costs
    - For every debit/credit transaction, calculating the period interest in real time => more processing,
    - Modifying all user notification messages, statements with the interest amount,
    - Supporting more Balance queries from users willing to see their interest balance => additional transactions
    - Updating customers care training to answer Interest related questions
    - etc ..

    Interests from global MPesa account would be shared between Safaricom, Suppliers and the Users…. with such low balances per account is, one can wonder how many users would get out of poverty from such interests.

    An idea would be that SF, transparently uses, those interests to fund high impact local projects (schools, water supply, lighting for rural areas, …).

    Ultimately, MPesa users may submit their best ideas/wishes/issues, and vote for the few selected projects to be funded. Some “Tontines” have a similar approach.
    In general, I am a strong supporter of Group Savings (rather than individual low balance savings). This allows to raise enough funds and go after significants projects for the common group benefits.

    I guess Agents/Business Accounts should get interests (against management/maintenance fees ..)

  • September 30th, 2009 at 3:40 am, Jerome de Mallmann ()

    As a strong supporter of mobile deposits and micro-loans, I share this view. However, we need to understand the challenges to make it a reality. We are asking mobile operators to become bankers.
    First of all, the regulation often does not allow this. In Kenya, the central bank is actually developing a legal framework to better control mobile financial transactions.
    Second, operators are not willing to become bankers and vice-versa, as these are complex businesses requiring specific know-how. So mobile operators and financial institutions need to partner. Building such a partnership to deliver an integrated service is a complex challenge. This is why, at this stage, most operators are testing mobile money, simply as another VAS, and, in the same way, bankers have launched simple mobile banking services such as alerts and bank statement requests.
    Today, we are still a very small number of experts with both mobile and banking skills and a real experience in mobile finance. The market will mature in the next two years, in some emerging countries first, as this is where the needs are and regulators will probably act more swiftly than in developed markets.

    Jerome de Mallmann

  • October 1st, 2009 at 7:00 am, Tony Gaunt ()

    I propose we change the thinking here – M-Pesa and equivalent systems may not provide the type and range of financial services needed by the poor, but they do provide an excellent vehicle for helping delivering these services.

    Focus should (and is at CARE) move to the provision of tools to allow the financial services providers to quickly and easily accept and make payments via existing mobile payments networks. In this way, the “last mile” cash movement is handled by the MNO and their agent networks, but the providers of savings, insurance and credit in these markets get the opportunity to change their business model, reaching more clients at a lower per transaction cost.

  • October 1st, 2009 at 11:42 am, Olga Morawczynski ()

    Many thanks to everyone for their excellent feedback. I published this blog because I wanted to start a discussion on savings. It looks like my goal has been attained. I wanted to address a few comments:

    “I propose we change the thinking here – M-Pesa and equivalent systems may not provide the type and range of financial services needed by the poor, but they do provide an excellent vehicle for helping delivering these services”

    I absolutely agree with this point. I am not suggesting that M-PESA become a bank or MFI. I am simply suggesting that they work with these institutions and offer a link between the poor user and formal financial services.

    “As a strong supporter of mobile deposits and micro-loans, I share this view. However, we need to understand the challenges to make it a reality. We are asking mobile operators to become bankers.”

    No, we are just asking mobile operators to figure out ways in which they can work with banks.

    “CBA is not the Central Bank of Africa but a retail bank called Commercial Bank of Africa”

    You are absolutely correct! This was a mistake on my part.

    “MPesa don’t get the interest (or treasury income) on funds as the money is not theirs (it is the user’s money) and so the money is held in a Trust account on behalf of the user. In fact the bank keeps the Trust account interest!!”.

    The point here is that poor Kenyans should be reaping the benefits of their limited savings, and not large institutions. We just need to strategize and figure out the best way in which this can be done.

    “Some also made clear that M-PESA was actually safer than the banks. Given the history of bank crashes in the country, this is no surprise”” is not really accurate any more. Kenyan banking is now acclaimed as being as robust as European / US banking systems.”

    Yes, I also agree with this point. However, there was a distrust of banks amongst many of my informants especially the poorest. These bank crashes were not easily forgotten. As a result, numerous informants would spread out their savings to decrease the risk of loss—putting some cash in their home bank, some in the bank, and the rest in M-PESA.

    I think most of us agree that mobilizing savings via the mobile phone is no easy task. But I like the suggestion put forth by Shaun, “The big market opportunity is tying this whole thing together in one seamless independent national ecosystem”. This way, bankers do not have to become banks, or vice versa. They just need to work together and figure out how best to serve the needs of their poor users. This is certainly a challenge that this industry will soon need to address. The main thing I want to do here is highlight the need for this type of service.

  • October 4th, 2009 at 6:24 am, Elena Mastava ()

    Olga, very interesting blog. Are there any good examples of savings products that can be accessed through the mobile phone?

  • November 6th, 2009 at 11:54 am, MNOs as Depository Institutions? « FinVentures ()

    [...] http://technology.cgap.org/2009/09/28/why-m-pesa-should-offer-savings-accounts  [...]

  • Leave a Reply