From AfricaFinance Forum blog, Making Finance Work for Africa, 25 November 2013]
Let me say again that I see a huge gap between
the potential of new electronic channels and the results that are being
observed on the ground. Much as we might convince ourselves of the inexorable
logic of bringing financial services to the corner shop near where people live
(agent banking) and right onto their hands (mobile money), what I see as I
visit country after developing country is too much effort and too many
resources being expended in entirely sub-scale operations. Must getting there
feel so hard?
Commercial players: don't play hero
As in any network business, mobile money operations are
about numbers of customers and breadth of ecosystems. Unless you have the kind
of scale Safaricom had in Kenya, are you sure you want to tackle the whole
mobile money ecosystem on your own?
Are you sure you can convince people to get off using that
grimy physical cash which touches and is immediately accepted by everyone, and
instead hop onto your private, exclusive electronic cloud? Your cloud would
cast such a bigger shadow on cash if it was combined with all other similar
clouds into a single, interconnected electronic payment network for everyone.
Are you sure you want to make the management of cash in/out
(CICO) -that thing which wears you down and which you so dearly would like to
go away-the key competitive battlefield with everyone else who abhors cash as
much as you do? Your cloud would be much more accessible for those sadly stuck
on cash if you joined forces with all the other electronic types and worked
together to create CICO networks that work for all of providers.
Are you sure you want to take it upon yourself to sign up
every primary school who wants to bill parents, and to sign up every small
employer who wants to pay employees, one by one? They will not want to force
all their parents and employees to join your cloud, and yet they will not have
the appetite to sign up with every other cloud, so they'll feel it's easier to
just continue with cash like they have always done. They would be so much
amenable to e-payments if the various players empowered a few payment
aggregators to work on their collective behalf in signing up those schools and
employers and distributing the transactions according to who has which parents
and workers as their customers.
If players are able to leverage the collective scale, the
total will be more than the sum of the parts. But getting to this result
requires that the industry as a whole work out which areas they must
collaborate on and which areas they want to fight tooth and nail on. Now
competition between mobile money operators tends to be focused on size of
payment network, ubiquity and liquidity of cash in/out points, and the length
of the list of billers/bulk payers signed up. Those are precisely the aspects
on which scale matters most, but the resulting fragmentation is only making
cash loom more supreme. How about if these became areas of collaboration, and
the competitive field was shifted to brand, customer service, product
development and quality of user interface instead? Aren't these, in fact, the
things that should turn on modern digital-based services?
Regulators: tear down those walls
Many regulators have gone to surprising lengths to allow new
services to emerge, often without explicit regulation, against prevailing
orthodoxy. But still, when the supply response is so weak as it is in many
countries, policymakers have to wonder whether there are other tacks they can
take to spur the market on.
For one, free up CICO networks
from the clutches of banks and mobile money operators. Forcing service
providers to be contractually responsible for anything that goes on in
thousands of stores (the current regulatory mantra) is not only illusory but
counterproductive: how can such fuzzy liability not lead to smaller,
proprietary cash networks? Instead, create a license for CICO networks, with
clear consumer protection rules, and let them operate for any and all financial
service providers. All they would need to be a CICO point for a given bank or
mobile money provider, beyond having a CICO licence, would be to have a funded
account with that institution and access to their secure, real-time electronic
Many regulators can also give up on their aspirations to be
match-makers for happy bank-telco partnerships. These are not natural things,
they remain rare on the ground to this day. They may emerge in time, but don't
predicate the development of the sector on two species with different genetic
make-up mating and working together (India's RBI, take note). Let banks and
telcos compete, under clear guidelines and a level playing field. Let telcos
and other non-bank players contest the market with an e-money license that
exempts them from onerous prudential regulations for the very good reason that
(if they are licensed and supervised properly) they do not create new
prudential risks. Let banks compete on the basis of the same third-party CICO
and account opening regulations that apply to non-banks, for the equally good
reason that CICO and anti-money laundering risks are the same no matter who the
account issuer is.
Regulators need to offer pathways to mobile money providers
that require less commercial brute force. And providers need to develop a more
nuanced view of how they can cooperate to build a new market and compete to
gain share within that market. Such is the modern way with most network and
[From World Bank's All About Finance blog, 22 October 2013]
Hard cash certainly has its drawbacks. Poor people mired in
a cash economy find it difficult, in times of need, to support or seek support
from distant relatives and friends. The size of the market they can sell their
products and wares into or source their inputs from is limited by how far they
can easily and securely transport cash. They are captive to local financial
organizations and moneylenders, because more distant financial institutions
don’t find it cost effective to go collect their saved-up cash and have no
visibility of their prior cash-based financial histories on which they might
otherwise grant credit.
All of these are good reasons to expect that people
everywhere will embrace digital money, if only it is served up to them in a
convenient, understandable, reliable and secure way. Money is just information
–how much I have, how much I owe—and the short history of the internet shows us
that information wants to become free of physical impedances.
So will cash go the way of the compact disk, in a gradual
wind-down towards oblivion? Must we or even can we go for an accelerated
eradication of cash? Many hope so, but I don’t think so.
The CD is simply digital information bottled up for
convenient transport. But once devices became ubiquitously connected, there was
no longer any reason for musical information to be delivered through a physical
distribution network rather than online.
But cash is more than just bottled-up information on
financial value: it is value that is readily and universally recognized and
accepted on mere visual inspection. The physicality of banknotes makes it easy
for people to make snap judgments on how much value it embodies and whether
it’s a real banknote or not. Cash is a visual acceptance instrument,
in contrast to electronic money which requires electronic acceptance (an
ATM, a point-of-sale terminal, a mobile phone) in order to be recognized and
Electronic acceptance introduces risks that when you want to
pay with electronic money there may not be a device available, that it may not
work properly, that it will delay you in some way, or that the information on
that payment will reach someone you’d rather didn’t know about it (including
the taxman). It takes a long time to overcome these fears, which is why the
shift to electronic money is so slow and gradual, even in the most developed
So here you have the basic trade-off: electronic money is
superior to physical cash in transport and storage (lower transaction costs),
but cash has advantages over digital money in acceptance (immediacy,
universality and privacy).
Digital music, in contrast, requires electronic acceptance
(i.e. translation of stored digital signals into sounds) whether it is
delivered as a compact disk or online. So the compact disk involves much
transport pain and no acceptance gain. That is why it’ll end up going away, as
have done the specialist stores that sell them.
While digital money will tend to expand inexorably over
time, both forms of cash will find their sweet-spot uses and will co-exist.
That presents a major problem, because the interface between the two is the
most expensive and operationally complex component of mobile money systems. The
lack of interworking between these two forms of money creates a last
centimeter problem (you can get the bill so close to the phone, but
not quite into it) which we only know how to solve by turning into a last mile
infrastructure problem (building networks of ATMs and cash in/out agents).
That’s a drag for everyone.
I predict a long future for banknotes, but with the
difference that they will come to be accepted both visually and electronically,
indistinctly. Future smart banknotes will
retain the physicality that allows them to circulate and be traded like normal
banknotes, on mere visual inspection, but also the smarts to allow their value
to be transferred at will from the paper note to your digital bank account (and
vice versa) using an appropriate acceptance device like a mobile phone. Smart
banknotes may be turned on and off through encrypted messages sent to a chip
embedded within them, and their state can be checked both digitally (e.g.
reading the chip with a mobile phone with near-field communication) and
visually (e.g. by means of digital ink or other visual displays). Thus the last
centimeter will vanish.
Far fetched? For sure, but maybe not quite as far-fetched as
imagining that stores everywhere will become ubiquitous and reliable cash
in/out points for banks and mobile operators.
[From IMTFI blog, 21 October 2013]
Myths as retrospective reinterpretations
Every movement has its founding myths. These stories are
rarely manufactured during the early struggle to catch attention. Rather, they
develop at the onset of success, and tend to be retrospective reinterpretations
designed to give a greater sense of purpose and method to what happened along
the path to success. Myths generally have a strong basis of fact, at least
initially, but they always play very selectively with those facts, and usually
get disconnected from the facts through noisy repetition. Never let the facts
get in the way of a good story. Founding myths serve to give a sense of
inevitability of outcome; to create a sense of order in the inherently chaotic
process of challenging and innovating; to create inspiring figures
(individuals, organizations) that others can seek to emulate. In modern
parlance, we call them case studies.
Mobile money is now cool, and it has attracted many of the
trappings of a movement. Never mind that it’s just applying the logic of
prepaid cards and internet banking on mobile phones, which are the only digital
devices that most people in developing countries happen to have. In a cold
business analysis, mobile money is simply the application of a secure,
real-time, digital infrastructure permitting banking o shift from costly direct
channels (branches, ATMs) to much more distributed indirect channels (stores as
agents) and even to self-service (mobile) channels.
So what are the founding myths of mobile money, those origin stories which have
inspired the industry? I would say there are two: agent banking in Brazil (even
though agents tend to be connected through bulkier point-of-sale [POS]
terminals rather than mobile phones, but that’s just a technicality because POS
terminals carry a SIM card within them), and Smart Communications’ mobile money
service in the Philippines. These two aspects, plus excellent messaging of
customer benefits, are the three legs that Safaricom stood on to build its
wildly successful M-PESA service in Kenya.
The early Brazilian experience
A critical enabler or component of mobile money is the
conversion of cash to electronic value through non-bank retail outlets, and
here Brazil is touted to this day as the poster child. Agent banking
regulations date back to 1999, and by end-2008 (the dawn of the M-PESA era),
Brazil had over 140,000 agents (correspondents, to use their term). Sure,
Brazil is a large country, but not even Kenya today comes close to that. Here
we see the key ingredients of a founding myth.
One element of myth-making is the injection of purpose. What the Central Bank
of Brazil sought to do with its agent regulations was not so much to create new
channels but rather to find a place within the law for a constellation of
existing semi-formal and outright informal players who resold or brokered bank
services to clients. The central bank was driven by consumer protection
concerns (clarifying obligations and responsibilities across existing players)
rather than by aspirations of financial inclusion.
Equally, banks did not see agents as a way to offer more convenient service to
existing customers or to seek out new ones, but largely as a way of getting
non-customers out of their banking halls. That’s because of an unusual
provision in the Brazilian banking law, which stipulated that bill payment was
a banking service and hence could only be conducted at banking outlets. Banks
largely sought to shift this mass of bill payers onto non-bank retail outlets
while respecting the letter of the law – agents, bingo! By end-2008, 75 percent
of agent transactions were still bill payments, and only 5 percent were
Another element of myth-making is the misleading interpretation of data. Of
those 140,000 agents in end-2008, only some 37,000 offered cash in/out
services. The rest were a plethora of brokerage services for loans, mortgages,
pensions, insurance and such. And two-thirds of these 37,000 I’d be hard
pressed to count as a novel kind of retail agent. This figure includes the
lottery houses, directly owned by Caixa – no arm’s length there. And it
includes the post offices which Bradesco was running as a postal bank – as if
postal banks didn’t have a long tradition elsewhere. If you exclude these
agents, you are left with far fewer agents than there were bank branches (circa
19,500) at the time.
The early Philippine experience
Now onto the mobile phone. Smart Communications is rightly
credited as being the daddy of all mobile money systems in developing
countries, launched in 2001 and hence predating M-PESA by an impressive six
years. Smart Money was the first operator to create a secure platform permitting
customers to access a bank account securely from any mobile phone. Smart Money
was much celebrated in its day, but its shine has now worn off, largely because
of its paltry performance compared with M-PESA.
Was the touted success of Smart Money a mirage? No, it’s simply that Smart
Money was not intended to be what later mobile money enthusiasts have wanted it
to be. Smart Money was conceived not as a person-to-person money transfer
system but rather as a mechanism to broaden the base of electronic airtime
distributors across the operator’s customer base. Here was the thinking: if
customers’ prepaid airtime account could be linked to their bank account, they
would be in a position to buy as much airtime as they needed, on the spot.
Combine that with the ability that users already had to transfer airtime to
each other, and now any Smart Money user could become a Smart airtime reseller
– in the process earning airtime commissions, which have historically been very
high in the Philippines.
So Smart Money was designed as an enabler for airtime distribution, rather than
as a consumer product in its own right. The design implication was that a cash
in/out network was not so important: the money would come from linked bank
accounts rather than from cash, there would be little cash-out since the money
would go primarily into buying airtime, and in any case juicy airtime
commissions to business-minded customers would compensate for any inconvenience
from having to go to a distant or crowded bank branch to deposit cash. But
precisely because Smart Money (and G-Cash, which came after it) underplayed the
importance of the cash network, transitioning it to a consumer product à la
M-PESA proved very difficult.
Their enduring legacy
All this is not to downplay the importance of Brazilian
agent banking or Smart Money as precedents for what has come later in
branchless banking. Their significance will be better understood if we look at
them from the lens of the problems they were trying to solve at the time,
rather than the problems we wish they had been trying to solve. If you don’t do
this, you’ll be puzzled if you go to Brazil and notice the low level of banking
(non-bill payment) transactions still happening through its agents. And if you
go to the Philippines, you’ll be surprised to find so few people who send
mobile money to each other. It’s only fair that we tell their story their way.
My respect goes to them – and to M-PESA which was the first to combine these
two stories –the agent and the mobile side of things—to great effect.
[From IFMR blog,
18 October 2013, with Abhishek Sinha]
When you have two systems running in parallel, the hardest
part is always managing the interface between the two. Customers don’t usually
all migrate to the new system entirely and at the same time, so there is a need
for the new system to offer backward compatibility with the older, more
established system. Without that, the stakes to migrating to the new system may
be too large, and adoption will lag.
So it is with money. The new electronic form of money must
integrate as seamlessly as possible with legacy paper money if new-to-banking
people are going to be at all comfortable in experimenting with and using it.
That’s why the retail agent or Business Correspondent (BC) bit is the
cornerstone of any branchless or mobile banking proposition.
But that is where the economics and operations of branchless
banking ventures get really tricky. The technical bit is easy: as long as it’s
about managing electrons and bits, scalable, low-cost systems can be put in
place. But a cash in/out channel is an entirely different story: it needs to be
carefully constructed bit by bit as a patchwork of stores. Each store needs to
be vetted, supervised and supported. Each store needs to be fed with enough
revenue day in and day out to justify setting liquidity aside and running to
the bank to rebalance when liquidity runs out.
You don’t see many large-scale, branded retail chains in
India in any sector (think groceries, pharmacies, agricultural inputs,
whatever), and that’s for a reason: managing retail channels is really hard
business in a country with such a disperse geography and where there are
already so many local shops running on razor thin margins.
If that’s the case, why should banks, who have no experience
operating beyond their own branches, be expected to succeed in creating their
own networks of BCs? Some are trying hard, but how many can claim to have a
sufficiently dense network of stores doing brisk cash in/out business every
You should let specialists manage retail networks, and let
banks ride over them. Specialist store aggregators, operating under specific
rules and guidelines issued by the RBI, could then offer BC service for any and
Think how such specialized shared BC networks could
transform the problem. Entrepreneurs with experience in managing indirect
channels would develop business where banks themselves are not willing to go.
Banks would simply sign up whichever shared BC networks they feel are in
relevant locations and offer good service to their clients. Individual stores
would be able to offer service to all their clients, whichever
bank they happen to bank with, in the same way as they sell a range of
toothpastes to suit their clients’ various toothpaste preferences. Banks
wouldn’t each need to deploy essentially the same systems to reach pretty much
the same stores – a costly duplication that makes the already precarious
economics of cash in/cash out altogether unachievable.
If this seems far-fetched, that is essentially what is
happening today under the RBI’s new policy allowing third-party ATM networks.
Why should banks themselves have to run with the operational hassles of
installing and managing hardware and replenishing cash boxes? That can easily
be delegated, as long as certain ground-rules are met, especially on aspects of
technology platform and consumer protection. These are properly identified and
clearly laid out in the RBI’s guidance on the topic.
A BC is functionally equivalent to an ATM. An ATM is
essentially a point-of-sale (POS) terminal (the card reader, the screen, the
keyboard) with a cash box attached. In a BC setting, the only difference is
that the cash box is physically separated from the POS terminal (which might be
as simple as a mobile phone), and the cash gets dispensed or accepted manually
by the shopkeeper rather than automatically. But the transaction is governed
electronically through a banking technology platform, because the POS informs
the customer how much cash to hand over to or to take from the shopkeeper. What
is important is that transactions always occur on a technology platform
controlled by a bank (not the third party network manager) and hence under the
clear supervision of the RBI, and that the network manager be bound by clear,
specific consumer protection rules.
Under a shared BC model, each store might operate from a
single bank account using the technology platform provided by one bank (we
could term this the “acquiring” bank). Any transactions it performs on behalf of
customers from other banks could be settled through the NPCI’s real-time mobile
Creating third-party networks of shared BCs is the logical
next step in the process of enabling scalable branchless banking services. Free
up banks from the burden of managing the operational aspects of cash-in/out
networks. Aggregate the cash in/out transactional volumes across all banks to
make the business case easier for individual stores. Flash forward to the
inevitable step of infrastructure sharing, like telcos have come to accept with
tower sharing and banks with ATMs. And let banks concentrate on their core
mission: developing, selling and managing a variety of electronic financial
services that solve people’s broad financial needs.
[NextBillion Financial Innovation blog, 15
Do you realize to what extent formal savings services are
irrelevant to many of the world’s poor?
The first step in any recovery program is to acknowledge the
problem. So before we talk about how innovation can encourage savings at the
BoP, we need to confront two uncomfortable facts:
1. Half the world’s adult population
doesn’t have an account in a formal institution
2. More than half of those in
developing countries who have an account admit that they haven’t saved in it at
all in the last year (source: Findex).
In reducing finances to a bunch of numbers (account
balances, amounts owed and due, available credit), formal financial services
tend to become dissociated from the mental models, social relationships and
customs that dominate people’s financial lives. But allowing for some of that
nuance invariably adds complexity to formal products (more accounts, new terms,
longer forms, more levels on the mobile user interface). People are put off by
this tussle between relevance and complexity. They might pick out individual
products that solve particular large pain points for them (sending money long
distances, access to cheaper credit) and avoid the rest, leaving in particular
a wake of empty savings accounts.
In an attempt to make savings products both relevant and
simple, researchers and providers often narrow their focus to specific use
cases that appeal to defined customer segments. This may lead to simpler
individual products, but the accompanying product proliferation presents an
insurmountable marketing challenge. It also risks pigeonholing customers into
prescribed behaviors. For instance, many commitment savings products feel
somewhat patronizing in their attempts to develop good savings habits by
securing promises of regular deposits.
So how can we establish customer relevance, avoid product
complexity and preserve broad product appeal? We need to
First of all, we must think long and hard about the core
customer proposition around savings in developing markets. It’s less about what
to do with money people have now (wealth management) and more about how to
secure the money they will need (building discipline). It’s less about managing
stocks (a storage problem) and more about regulating flows (a payments problem).
It’s less about fillable vessels (the account itself) and more about adjustable
rules and frictions (the restrictions on it).
Frictions are mechanisms that help you keep money
saved. In truly innovative savings products, those frictions cannot be tacked
on as features on an account; they need to define the savings
Take locking money up, for example. What if instead of
opening a time deposit that reaches maturity after a fixed term, people could
simply send or push money to a specific date (or a month) based on when they
will need it, much as they can send or push money to someone else?
Another friction is avoiding instant liquidity by having a
waiting period. What if instead of pushing money to a fixed date, people could
push some money to Fridays or Sundays? They’d set up different pots of money
that would become liquid weekly, and which they might associate with different
purposes (e.g.: play money versus money for a family trip).
Yet another friction is indivisibility, to avoid small
temptations. Imagine if people could open an account dedicated to saving for a
chicken or a goat, and could only push money in and out in lumps of $4 or $40,
because that’s what a chicken or a goat cost, respectively.
Or, as a final example, consider peer pressure as a
friction. People often send money to someone they trust (informally referred to
as a money guards) for safekeeping so that they can mentally set it aside.
Imagine a product where the designated money guard can’t do anything with that
money other than sending it back to the person when they ask for it. This is
saving through a friend; there’s peer pressure because dis-saving can only be
done in connivance with the friend/moneyguard.
These are examples of how one can define savings products as
payments to oneself, by adding a time, social or other dimension to them. One
can go further by adding rules, which are mechanisms to help people
to save more regularly (as opposed to frictions which help people keep the
money saved). Common rules consist of recurring deposits (or self-payments) and
automatic sweeps, in which accounts automatically transfer amounts that exceed
a certain level into other savings or investment options. A more interesting
one is a “personal ROSCA” (Rotating Savings and Credit Association),
an informal, community-based savings group. Its logical opposite, “paying
yourself a salary,” lets people distribute a windfall over a number of weeks.
However, figuring out these frictions and rules as drivers
of savings is not enough. For poor people, all moneys need to stand ready to do
double duty. A jewel may be a stepping stone to a good marriage and a pawnable
asset in case of a medical mishap. A cow is an income supplement and a buffer
against emergencies. Savings (building up assets and social status when
circumstances permit) is not separable from credit (being able to extract value
out of assets and relationships at times of need). So our frictions and rules
must have outs, such as being able to borrow against some future
money that customers have pushed forward or set aside.
In developing innovative savings products, we must think of
financial services as a Swiss Army knife of management tools which people can
use and adapt in their own way. It’s the principle of mass customization: a few
well-designed tools can be used to support a variety of use cases and shape
very personal experiences. People’s sense of ownership over the relationship
with the financial service provider - and their feelings of control over their
money - will be enhanced if they are able to define how they use the service.
The key challenge is to make these tools intuitive, both in
terms of what they do and how they are used.
[From Center forFinancial Inclusion blog, 30 September 2013]
Remote payments usually are the beachhead for mobile money
as it struggles to create a role for itself alongside cash in developing
countries. It’s easier to create customer awareness, induce customer
experimentation, and generate customer willingness to pay when you are
addressing situations in which cash presents the biggest pain points – sending
money home, paying bills at distant and busy utility offices, travelling with
or delivering larger stashes of money.
But that’s a limited market. Those are not daily occurrences
for most people. And in many markets, especially outside of Africa, the reality
is that people already have decent domestic remittance or bill payment
mechanisms – through networks of pawnshops in the Philippines, courier
companies in Colombia, or Hawala in Afghanistan.
Now mobile money providers are itching to get into the
merchant (i.e. in-store) payment space. That’s where you can drive daily as
opposed to monthly usage, so that’s where the volume is.
Of payment volumes and value
Much of the discussion on mobile retail payments centers on
pricing models (merchant fees, interchange, potential cannibalization of P2P
pricing), usability issues (customer convenience, speeding up transactions at
the store), and acquiring strategy (how to roll out devices at stores and at
what cost). These issues arise because cash is a much stronger competitor here
than in a remote setting: paying some cents and waiting some seconds for an SMS
confirmation are an irrelevance when you are sending money halfway across the
country, but can tax people’s sense of fairness and patience in a retail
setting. The less value people and stores see in electronic payments the larger
these issues will loom.
But treat these as hygiene factors. Imagine that you could
buy half a kilo of rice at any corner store from your mobile wallet at no cost
and as conveniently as you flash out a banknote. Still, the fundamental
question remains: why would you take out of your pocket your mobile phone
rather a banknote?
There is only one robust answer to this question: I want to
pay electronically because electronic money is what I have. The only reason why
you and I are likely to pay with a card next time we go to a restaurant is to
spare us having to visit an ATM before going to the restaurant. But if our
mattress was our bank, we’d look at it differently: we’d want to pay at the
restaurant with cash to spare us having to visit the bank branch to deposit the
cash we’d need to fund a card payment. It’s converting the money – not paying
it – which creates the real cost and inconvenience.
And here’s the rub with mobile money. Most people’s accounts
are empty or have very little value. Cash is what they have. If that’s the
case, why would they want to pay electronically in situations where cash is
easy to pay with? The daily merchant payment opportunity will remain largely
unaddressable unless people start seeing value in storing their money
Of payment chickens and eggs
We are frequently told that retail payments present a
chicken and egg problem: there is not much point in people getting cards if
stores don’t accept them, but stores won’t want to get terminals if their
clients don’t have cards. The solution: flood the market simultaneously with
cards and terminals. It takes a big push to flip the market to electronic
Mobile money providers have taken this to heart. Since
customer accounts are what they have issued in large numbers to-date, now all
they need to do is to go on a merchant acquiring spree. The happy balance will
be struck, and torrents of retail payments will be unleashed.
For sure there is a chicken and egg situation between
issuance (cards in customers’ hands) and acquiring (terminals in merchant
outlets). But keep in mind that issuance and acquiring refer only to the
linking of payment instruments to accounts. But what can possibly be the
significance of linking a payment instrument to an empty account?
Here’s where the experience with mobile money in developing
countries differs from the experience with debit cards in developed countries.
When debit cards were first issued in developed countries, they were linked to
accounts that had balances in them. Cards fell in the hands of people who had a
natural long position in electronic money, and hence had an
in-built bias towards paying electronically. Electronic storage of value
preceded electronification of retail payments. Not so with mobile money in
So just as there is a tension between electronic issuance
and acquiring, there is one between electronic payments and storage of value.
But these are two very different problems. The former tension, between the
deployment of cards and terminals, truly is a chicken-and-egg problem that
needs to be solved simultaneously. Issuance will not be successful as long as
acquiring is not successful, and vice versa.
But the second tension, between payments and storage of
value, is not a chicken-and-egg problem and needs to be
addressed more sequentially. For all the reasons I have mentioned, if you get
people to store value electronically it will lead them to naturally want to pay
electronically (if only the issuance and acquiring chicken-and-egg problem is
solved). I would find it extremely hard to make the argument backwards: that
people’s desire to make small payments electronically when they go to the
corner shop will lead them to want to save larger balances electronically.
People’s savings decision is driven by larger questions of sense of control,
safety, and manageability of money, not by fine points of convenience at the
corner store. We must change the metaphor here: retail payments are the tail
that can only marginally wag the savings dog.
By the way, just getting people to be paid electronically
more often (per the other ongoing mobile money gold rush: G2P) won’t do the
trick while their practice is to withdraw any incoming money immediately and in
full. The money has to stick in the wallet, and for that to happen mobile money
needs to have more money management tools and behavioral hooks that connect
with people’s mental models and make them feel more in control of their money.
So here you have it: the savings dog is where issuance
chickens and acquiring eggs come from (sorry!). We can no longer neglect the savings
agenda of mobile money. Savings may not be hugely profitable by itself, but it
is the engine that fuels retail payments. And credit too, since savings and
payment behavior ought to be primary inputs into credit scoring models.
[From CGAP Microfinance blog, 24 September 2013]
There is so much talk now about how electronic, and in
particular mobile, payments are replacing cash. Some see it as a war on cash,
the more circumspect talk about transitioning through a cash-lite state.
I tend to look at it from the opposite perspective: far from replacing cash,
mobile money has made cash so much more efficient. Even in Kenya.
Mobile money, and in particular the agents it spawns, pushes
cash collection points deeper into local markets, slums and rural areas where
before the only path to cash was to take a long bus ride out to an urban
center. If cash had feelings, it could hardly be affronted by all these
attempts to make it easier for it to be dropped off and picked up locally, at
will. If I were cash, I’d be happy to concede long-distance travel and entrench
myself where it matters: in daily life. With mobile money, cash works better.
Indeed the predominant practice the world over is for
senders to deposit the cash just prior to sending it, and for recipients to withdraw
the cash immediately and in full. From the customer perspective, it’s a
cash-to-cash service, and who cares what happens in between as long as the cash
gets there. The electronic bit in the middle is just a matter of efficiency for
the service provider processing the remittance or the government paying out
social welfare; all the customer knows is that she now needs to walk a shorter
distance to collect the cash.
Or look at the macro trends. M-PESA money transfer volumes
may amount to an awesome 50% of Kenyan GDP, but because there’s cash at both
ends just think of how many more cash transactions it has spawned. Nobody
quotes those numbers. Because of M-PESA, cash does a brisker business; it
enjoys higher velocity, to use the proper economic term.
Given all this, it’s not a surprise that many mobile money
systems, particularly in Asia, are cutting right through all this and simply
processing payments over the counter (OTC), i.e. processing the transfer
through agents’ rather than customers’ mobile accounts and sparing customers
from having to press buttons on their mobile phones or open an account. Like
pawnshops such as M Lhuillier have long done in the Philippines, like courier
companies such as Efecty do
in Colombia, like specialized bill payment companies such as PagaTodo do in Mexico, and like
Western Union does globally.
What a pity that is.
Because the fight against cash is the good fight. Cash is
costly to move around, unfair on the poor (disproportionately expensive for
small transactions), unsafe, smelly, a hiding place for germs and criminals,
To truly go towards a cashless or cash-lite state, what matters
is not so much how money gets moved around but how it gets stored. Again,
look at it from the customer perspective: cash, that I go to a local shop to
drop off/pick up, is not electronic money, no matter how it left/arrived at the
store. Cash is only electronic if it’s not converted back into paper.
is what you have, cash is what you’ll use to pay at local shops. The unfortunate
corollary is: you can’t go cashless on empty
accounts. Mobile payments will have a glass ceiling as
long as mobile money providers continue to neglect
savings. The mobile money strategy ought to be to
infiltrate cash: making it so easy to go in and out, in and out of cash, that
eventually you don’t bother going out any more. That option is shut with OTC
because OTC must always start and end with cash.
The rising wave of OTC services amid mobile money
deployments reflects two failures. In the first instance, a failure of banking
regulators to make account opening simple and painless enough for people to
even contemplate acquiring an electronic repository of value. A central bank
that enables mobile and agent transactions but insists on cumbersome procedures
for account opening (I have a few in mind) is not a progressive regulator. But
more fundamentally, it’s a failure of providers to convince customers about the
storage-of-value benefits of mobile money accounts. The harsh reality is that
the majority of electronic accounts are empty. More precisely, it’s a failure
of innovation, because mobile money accounts fail to incorporate the subtle
discipline mechanisms that people like to wrap their savings in.
I sympathize with mobile money providers who feel they have
their hands full just building out a basic mobile money platform and an agent
network, and leave the challenges of dealing with burdensome regulation and
innovation to others. But policymakers and donors make a mistake in promoting
OTC-heavy deployments like they are just another flavor of mobile money. Sure,
today they do pretty much the same job as wallet-based services do today, but
their evolutionary path is stunted. OTC systems offer no possibility of
electronic storage of value and hence no path to cash substitution. Because the
agents are left to do all the pressing of buttons on their own mobile phones,
OTC services undermine one of the biggest unique selling propositions of
electronic cash which is privacy – agents get to know each and every
transaction customers do. (With wallet-based services, agents know their
client’s cash in/out transactions but not the destination/source of those
funds.) With OTC, customers can never become self-providing, because it is the
shops and not the customers who are mobile-enabled.
An account also forms the basis for recording a financial
history, which can help spread responsible credit. OTC services can only
generate credit scoring information based on what I have spent; with an account
one can also capture what I have not spent (i.e. saved). Accounts are therefore
a likelier stepping stone into broader financial services.
Some might argue that OTC services may be a valid stepping
stone to account-based services, but I don't see an obvious, customer-friendly
migration path from OTC to wallet, beyond mere price discounting. Opening an
account should be the (albeit, small) price to pay to get the benefit of
OTC services are commercially understandable but represent a
large missed opportunity policy-wise. We need to strive to put mobile money
technology directly in customers’ hands so that they can feel empowered and in
control over their financial decisions.
[From Next BillionFinancial Innovation blog, 19 September 2012]
You may be tired of hearing about mobile money, but the
world is not. The Google
Trends chart (also see below) on M-PESA, the poster
child service which launched in April 2007 in Kenya, shows that it keeps increasing
market share of global Google searches. And the GSMA keeps ramping up its count of mobile money
deployments across the world: today’s number is 191.
Let me dig under the hype and give a personal assessment of
where things are. In this post I’ll focus on the potential of mobile money as a
retail payment platform, leaving to a subsequent post the discussion of how it
links to the delivery of broader financial services. I’ll highlight 10 key
points: some reaffirm my hope (the good), some I find disappointing but fixable
(the bad), and some fill me with deeper-seated concern (the ugly). The former
points represent the key reasons for mobile money’s success, and the latter
hopefully lay some markers for its future agenda.
The value of real time: Mobile money is about
bringing immediacy to electronic transactions. With real time access, users
feel an unprecedented level of control: they are able to act upon their money
here and now. And providers can eliminate credit risk across the service fulfillment
chain, which greatly facilitates its propagation: deposits and withdrawals can
be accepted through loosely affiliated third party stores because cash movement
between customer and store can be offset electronically on the spot.
Give them the tools: The story of M-PESA is an
antidote against the natural customer-centric urge to develop
specific solutions to specific needs presented by specific people. The
remarkable thing about mobile money is how basic it is: you can hold electronic
value, you can pass it along to others, and you can exchange it against cash.
M-PESA is like a
tool, a minimum viable solution. It may seem paradoxical, but the more
basic the capability that is being offered, the more versatile and universal it
Inventiveness at the base of the pyramid: Given the
power of immediacy and a flexible payment tool, the people of Kenya took to
using it with remarkable inventiveness. Marketed as send money home,
it didn’t take long before people discovered that it also let you travel
without having to carry cash on your person, conduct informal business with
more remote counterparties, and squirrel money away for some other purpose.
People figured out how to incorporate M-PESA into their routines, supporting what they’d
The power of network effects: If you look at
development with a broad historical sweep, much of it has to do with
externalities and network effects: cities flourishing, trade routes opening up,
ideas propagating – defeating distance in all senses of the word. M-PESA has
demonstrated again the power of connecting people by reducing transaction
costs, in this case through a national retail payments network. Once enough
people move onto a new thing, the snowball effect can become unstoppable.
Operational troubles: Most mobile money systems I
know are blighted by operational challenges which limit their ability to build
on early successes. Their IT systems may go down with indecent frequency; their
agent channel management structure may not be sufficiently scalable or
adequately policed; they may be battling (internal) fraud cases; their call
center may be deluged by confused customers. All this is usually the result of
providers’ chronic underinvestment in systems, people, processes and marketing.
Lack of business interfaces: You’d be surprised to
see how much of the formal economy in Kenya turns on checks. Yes, those
non-real time, paper-based things. Mobile money systems are simply too skimpy and inflexible to
suit the authorization, traceability and business analysis requirements of
modern firms. They lack the kind of interfaces that would let developers,
business consultants and system integrators work with corporate clients to
adapt mobile money to their needs, which would entrench mobile money more
firmly at the heart of the (formal) economy.
Unrealistic expectations: It is a gargantuan effort
to rally sufficient numbers of people into simultaneously adopting new ways of
doing things and moving onto new digital platforms. But the shining case of
Kenya, the hype surrounding mobile money which feeds on the tiniest flashes of
success uncovered anywhere, and the sudden interest of the payment majors all
conspire to create a false impression of the mobile money business as a natural
extension of mobile operators’ core business.
Unsupportive regulation: Many regulators are inflexible and
refuse to see how new technologies and business models transform the nature of
risks involved. Policy agendas may be conflicting: for example, the need to
appear tough in the global fight against crime and terrorism causes many
regulators to insist on unreasonable account opening procedures which dissuade
millions from leaving the cash economy. And policymakers may be pressured to
protect banks from competition from others with disruptive models.
Closed-loop systems: Reaping network effects requires
building a platform for growth. Most operators simply do not have the size to
pull it off. Consider a mobile operator with 40 percent market share in an
African country with 50 percent mobile penetration. If half its customers used
mobile money, that would constitute a payment network spanning only 10 percent
of the population (40%x50%x50%) – not terribly useful. Systems need to
interconnect so that they can reach everyone.
accounts are empty.
While mobile money has demonstrable utility as a money transfer and collection
mechanism in developing countries, the stark reality is that most digital
accounts hold little or no value. This will severely limit the scope for
electronic payments to spill over into daily life,
because people have a natural tendency to want to pay at local shops in
whatever form they keep their money. As long as people eschew electronic
storage of value, cash will continue to rule.
David Wolman’s blog, 4 September 2013]
Hi David, I wish I
could give you an upbeat follow-up, but I don't think I can, at least for
emerging markets. After the developing world took the lead on mobile payments
on the back of the spectacular growth of M-PESA in
Kenya and the ridiculously large number of copycats that popped up seemingly
everywhere, I think the action is now firmly back in the developed world,
particularly the US. That's because mobile payments there are at the
intersection of two ballooning trends: the spread of smartphones, which let
third parties develop good customer experiences affordably and in a
telco-independent way, and the sweeping belief in Big Data analytics. The Big
Data crazy has led many people to believe that payments embody a treasure-trove
of actionable customer information, and a battle of the titans is brewing. It's
not banks vs. telcos as it was a few years ago, but Google vs. Amazon vs.
Facebook vs. Visa (part owner of Square).
markets, there has been precious little service innovation on mobile money.
Most of the schemes we see are still struggling to put some legs on a basic
M-PESA-like service based on remote payments and/or bill payments. That isn't
generating enough volume of transactions to sustain the service in most
countries. I would say it's still only a handful of countries where mobile
money has legs, which is to say has gained critical mass to sustain the big
ecosystem it requires. (Tanzania and Uganda are top of mind; Bangladesh and
Pakistan to some extent, but I tend to discount them because the majority of
payments are for cash over the counter, akin to Western Union, rather than
account-based; Somalia and Zimbabwe are supposed to be experiencing substantial
growth in mobile money, but I haven't been there, and I tend to discount all
stories of success until I've seen evidence with mine own eyes.)
To garner many more
transactions, there is now a lot of attention on electronic payment for goods
at stores. This will mean underwhelming results for some time to come.
M-PESA-like services are still too expensive and too fiddly or time consuming
to be convenient at everyday shops where poor people live. More fundamental
still, people will not have an inherent preference for paying electronically
for their daily stuff as long as their money is not held electronically. The
problem, or a problem, is that not many people are taken by the store-of-value
function of mobile money. One of my recent mantras is "you can't go
cash-lite on empty accounts."
implicit assumption is that people don't save money electronically. Or worse,
they immediately withdraw whatever electronic money they receive because
of a lack of financial literacy. That is an incredibly facile and incorrect
assumption. People don't want to keep any money in their mobile money wallet
for exactly the same reason they don't want to keep many banknotes in their
wallet: It's hard to hang onto that money. What they do instead is separate out
their money in jars, or in various savings instruments, based on how they plan
to spend the money. As long as mobile money fails to deliver a convenient form
of money separation with enough self-discipline nudges and hooks, it will not
serve a useful store-of-value function for consumers. And without
store-of-value there will be limited use of mobile money as a means of payment.
You'll recall that
my deferred payments idea --being able to send money to yourself on some future
date--was one early example of how we might achieve this digital earmarking
simply and intuitively. While most people I've talked to in the business like
the idea, I have not been able to get anyone--bank, telco or third party-- to
try it. They are all too busy sorting out their basic operational issues and
copying each others' services to have any energy left for innovation.
To put it kindly,
we are going through a 'consolidation' phase where mobile money providers are
ironing out their programs, learning how to mitigate fraud, and digging in for
the long run.
Over in the world
of banking, meanwhile, zzzz. I know only a few banks in this big world of ours
that are pursuing agency banking (getting closer to people by letting them do
basic things like cash in/out at local shops) with any sense of urgency. The
only exception seems to be in Kenya.
[From MicroSave Financial Inclusion in Action
blog, August 2013]
I’ve been arguing for a while that the mobile phone
shouldn’t just be a payment instrument: it should be a planning
or accounting tool. A mechanism that captures how people
keep tabs of their various pots of value, goals and amounts due is more likely
to then in turn capture the transactions that derive from them. That’s what
triggered the Money Management Metaphors (Metamon)
work I did with MicroSave earlier this year. I thought I was
making a somewhat original point. It turns out that I was just echoing
millennia of practice. Or so I have learned from David Graeber’s fascinating
and provocative Debt: The First 500 years.
Having been trained as an economist, I assumed the logical
sequence of events has everywhere and always been: first barter, then currency,
then credit systems, finally double-entry book-keeping – a natural evolution
towards higher levels of abstraction and complexity in trading arrangements. I
am now informed that there is no evidence that human societies ever worked on
barter. In the beginning there was debt, as people variously shared, gifted and
loaned each other stuff. The fabled “coincidence of wants”
problem that makes barter so impractical (the fact that at the market you and I
can only transact if I want your chicken and you want my goats) was
solved by separating transactions in time (now I take something from you, later
you’ll take something from me), developing simple debt tracking devices (such
as the tally stick),
developing various moral codes to guide the sizing and fulfillment of these
dues, and periodically netting out the various debts across people in the
In Graeber’s words, “abstract systems of accounting emerged
long before the use of any particular token of exchange.” The primary need was
to create common notions of value, not necessarily to harmonise how value got
stored or passed around. So in the beginning money only fulfilled a unit of
value or accounting function; means of payment and storage of value came later,
The startling conclusion is that “there’s nothing new about
virtual money. Actually, this was the original form of money.” Yet here we are
now worrying about whether poor people in developing countries can make the
transition from hard cash to dematerialised electronic money. Might
this begin to explain why it is that, given the right conditions, mobile money
can just take off as it did in Kenya?
Think about it: people everywhere seem to have no problem
managing the “artistry” of gifting – an even more intangible and convoluted
practice than exchanging digital money. You can see generosity and balanced
reciprocity leading to mutual insurance. But you
can equally see dependence and charity preserving hierarchy. In Graeber’s
eloquent words, gifts “are usually fraught with many layers of love, envy, pride,
spite, community solidarity, or any of a dozen other things.” There’s nothing
simple about that, but somehow people work out a proper response to gifts
(whether they are an honor, a provocation, or a form of patronage) intuitively.
So there is no reason to believe that dealing with abstract
notions of (mobile, digital) money should, in itself, be a barrier for ordinary
people who are used to informal debt and reciprocity arrangements. The real
challenge will be the formalisation of finance: making them accustomed
to reducing financial arrangements to a bunch of numbers and financial
relationships to impersonal arithmetic.
Removing the social context from transactions may obliterate
much of the intuition and survival strategies people have developed around
money matters for centuries. As Susan Johnson vividly explains, the social
dimension of informal finance allows for much more open-ended
negotiability of resources in case of exceptional
need. And it’s not all casual: reading the typologies of informal financial
mechanisms documented by Stuart Rutherford,
others, one wonders at how inventive and recurring certain structures are.
Those can only be the result of a natural evolutionary process based on
variation (fed by the inherent flexibility of social arrangements) and
selection (the disciplinary and insurance benefits they bring).
With formal finance, all that is replaced by binding credit
limits, inflexible terms made up by someone, and an imposed moral requiring you
to repay your debts on time (the “criminalisation of debt [non-repayment],” to
use Graeber’s graphic if hyperbolic language).
The core problem of digital finance for poor people is then
not how intangible it is, but rather how explicit everything becomes. Being
more discreet may be an advantage, but must it all become so discrete too?
To end with Graeber: “When matters are too clear cut, that introduces its own
sorts of problems.”