The Invisible Bank: How Kenya Has Beaten the World in Mobile Money | Hacker News Reader

The Invisible Bank: How Kenya Has Beaten the World in Mobile Money

newswatch.nationalgeographic.com

146 pointsramabk68 comments

(Kenyan here) I think that the fact that Kenya did not have such a large banking industry helped M-Pesa succeed. By the time the local banks realised that it threatened to eat into their business, it had become too big and popular to beat. We had the local banking association petition the Central Bank to introduce regulations on it, the cost of compliance to which would have made it a lot more costly than it is, and subject to a lot more bureaucracy. Luckily, the proposal did not go through.

Does m-Pesa work securely enough to not need those proposed regulations then? The consumers are protected against fraud for the most part already?

There is no real protection against fraud at the moment, the operator Safaricom is fighting to block conmen from the system hence the push for subscriber registration

I've never heard of a technical exploit compromising the M-Pesa infrastructure. Social engineering cons are on the rise, though (like sending a dummy text telling you that someone sent you money, then calling and requesting that you send it back to correct the error. If you fall for it, and have money in your account, you send them your money). Those are much harder to police, though a proposed requirement that all phone numbers sold be registered to an individual with ID may reduce these in the future. Short of sharing your PIN, it's hard to think of a way to reach into one's account and steal money.

When you send money to another individual what kind of check is there that you've sent the money to the right address? I read there's a confirmation step but in the event you accidentally sent the money to a random person that person could confirm too. Maybe a CRC on the phone number?

My bank has no check that you've sent money to the right account when you do a deposit (and I hate that). (Banco Itaú in Uruguay).

So it's no different in that.

You receive a text notification of the transaction when it goes through, which has the recipient's full names. In case the names are different from what you expected, you can call a customer care number and have the transaction reversed (as long as they haven't withdrawn the money yet). Interestingly, this facility leads to another type of confidence con: a stranger walks up to you and tells you that they need cash but cannot get an agent nearby. You have them send money to your phone and you give them your cash. After they walk away they call the operator and claim they sent the money to a wrong number, and the transaction is reversed.

Ah I guessed right. Regulations are the reason why this is not done well in other countries.

Anyhow, m-pesa is right thing done in right place. Props to Kenya!

Yes. I was thinking at some similar service. With a bank (regulations), you go there, make a deposit for $10 and you will have $10 in your account. With a mobile operator (no bank regulations) if you give them $10 for anything (like pre-paid credit) they will have to charge you VAT for that amount. So what you want is a mobile operator that can take your money but not apply VAT.

At the same time... Vodafone doesn't have any banking regulations like deposit insurance does it? That seems like a big deal to me. What happens if Vodafone goes out of business? What happens if they decide to just keep the money and shut down the service?

These are low likelihood events, but then, so was the financial collapse of 2008.

The point is more likely to be transactions than savings

If it ends up being big enough for long enough, then people will start storing value in it due to convenience.

The financial collapse was caused by highly-leveraged banks relying on complex opaque models to "manage" risk while making loans and purchasing financial assets.

Since neither Vodafone nor Safaricom are engaging in lending, borrowing or betting on assets those risks don't exist. It is more akin to pre-paid credit cards than traditional banking.

I don't think you can know what Vodafone does with money in it's care.

If they were doing those sorts of things, they'd presumably be regulated as a bank in Kenya.

Like most start-ups the established players underestimated the capacity of the new kid on the block to eat their lunch[1].

By the time banks ganged together to lobby against the service it was too late [2].

[1] http://www.moneyweb.co.za/mw/view/mw/en/page292681?oid=55103...

[2]http://kenyaaudit.blogspot.com/2008/12/banks-now-gang-to-fin...

Safaricom is a public company: http://www.safaricom.co.ke/fileadmin/Investor_Relations/Docu...

I read about halfway through this article thinking it sounds good, but they used insubstantive marketing language terms like "new innovations", "big ideas", "safely and securely", "flexible adaptable technologies", and "dreaming big but thinking locally" so many times that it gave the game away. In addition there are no down sides described, the alternatives are demonized, and there is no real technical explanation. It's indistinguishable from a full page magazine ad. It is clear is a paid placement by advocates for a concept, and not a real article from a journalist. I then glanced up to see what site is running this sort of puff article and was genuinely surprised to see National Geographic is now doing this sort of thing. I guess the journalist needs some practice so he can learn to hide his tells in future product placements.

"no real technical explanation" The article may not necessarily have been geared for technical readers. Instead the author chose to emphasize the ease of use. Also, I don't get the impression that alternatives were demonized.

I think you're looking at the article in an overly-cynical way.

"M-Pesa is not an attempt to recreate developed countries’ banking systems in Africa. Instead, it’s an idea which has been tailored to the Kenyan environment. Rather than giving up on poor, isolated communities as unbankable, it has extended financial services to their most apparently unlikely customers."

Where exactly are the alternatives demonized? The author simply suggests that tailoring a solution to an environment (in this case, Kenya) is ultimately more beneficial to those within the environment than a one-size-fits-all approach that only serves a privileged few.

I didn't get that impression. The company I work for sells services in the area of mobile payments, and M-Pesa really is the "holy grail" that many companies worldwide are trying to imitate. I visited Kenya in April, and it is everywhere. In the middle of nowhere along the roadsides you see little green shacks with the M-Pesa logo where you can collect money that has been sent to you. However, M-Pesa is not without its downsides. Safaricom is stagnant on adding features, and getting a "pay bill" number for your business so that you can accept payments from customers is a lengthy and somewhat complicated process. On the other hand, this opens new doors for entrepreneurs to fill in the gaps.

I don't think the argument was that M-Pesa isn't something real and substantive, it's that the article was a poorly done puff piece. Honestly, I got as much from your one paragraph write up about M-Pesa as I did from the original article.

Read itentirely and it's as vague.

Does not explain that it only works because the tellco is a public company (otherwise your money with them would be invested in risk) and the usabilities issues that allow a.lot of theft via social engineering.

I previously founded a credit and savings bank using mobile money in the Philippines. M-Pesa is a fantastic success in Kenya but replicating it's success in other countries has been very difficult. There are several reasons for this: 1) There were no alternatives when it was launched, eg Western Union. The best way to send money home to your family was literally to trust it to a bus driver, you can imagine how well that worked 2) The regulators were very 'soft' touch and allowed Vodaphone to launch a service which was illegal in many other countries (for money laundering reasons) 3) Vodaphone had an 80-90% market share in most segments, this prevented the 'cross-network' problem

I believe that until android (or other) smartphones become affordable to poor people, freeing innovation from the carriers, we won't see M-Pesa's success get replicated around the world. Then we will see rapid growth shortly after that tipping point is reached.

Stereotypical POV about Africa(that bus driver shit). There was Western Union & Moneygram before M-pesa. M-pesa was just better(cheaper, faster, cellphones everywehere,).

He is still right on on being completely locked in one network with dumb phones.

There was no Western Union in Kenya villages before M-Pesa launched...

In Africa, many homes do not have electricity but increasingly everyone has a cell phone. Prepaid cell phone minutes have become a defacto currency in many African countries. Africans are in many ways more acclimated with using privately issued digital currencies (such as M-PESA)...

Where do they charge their phones? Are portable hand-powered generators common, like wind-up radios?

http://www.viewchange.org/videos/kenyan-solution-for-chargin...

video2text: They extend the light machine on bicycles with just enough circuitry to make it suitable as a charger.

Hooking up a bicycle dynamo to charge your phone is a popular african hack: http://www.ted.com/talks/william_kamkwamba_on_building_a_win...

An unreliable grid, car batteries can also store off-grid power, fancy people have generators

Can anyone explain how it actually works?

I am guessing you give vodafone 10 dollars, they give you a txt with a random code anyone with that random code and your number can request the 10 dollars to come off your account and onto theirs

but ...

What is the strength of the code ? What is the security around the transmission to your phone? It sounds bruce schneir might no like it

You register with Safaricom (The Telco) with your Identification Card Number/Passport Number. You get a 4-digit PIN to use as a password. After this your phone number is your 'account number', in a manner of speaking.

An M-Pesa application comes installed as a SIM application (I don't know the technical details, I access it from the "SIM Toolkit" menu in my phone) on all Safaricom SIMs. The application allows you to Send money, check your balance, and pay bills to registered corporate accounts. You need to know the account PIN to perform any of these actions.

To deposit money in your account, you need to go to an agent (these are numerous all over the place) and pay them. These are registered business who can deposit money in your 'account' (a normal customer cannot deposit money directly into another person's account), and you get a Text message confirming that the transaction was successful, and your account balance has been updated. When you receive money, you receive a text message, and you may withdraw it from any agent. You need photo ID to withdraw money from an agent.

-- edited for clarity.

Is the value purely stored in the SIM? ie does this all work without any form of cellular connectivity, or is there some sort of synchronization over cellular.

How do they deal with rogue agents? Presumably a rogue agent can claim you have handed over a large sum of money but it would take a while to discover they have lied.

No value is stored in the SIM, the transaction record and balance is stored in Safaricom's servers. Agents have a registration number prominently displayed, and all transactions are immediately confirmed by SMS to both parties. I think agents also need deposit a (significant) amount of money to get registered. Daily transactions for individual accounts are capped at 70000 shillings (about 900 dollars), so is the balance in the account. It would be difficult to carry out fraud of any significant scale with the system.

So that means all transactions have to take place where there is cellular connectivity then?

I live in the US - our connectivity is highly volatile and connectivity concerns are still a major part of the cellular consumer experience.

I was on the lookout for lions in a wildlife park in Uganda while checking my email... It's all possible.

So they don't need bank accounts, which were never required for cash purchases anyway and was a red herring, but they need government id numbers now to be tied to and track their every purchase.

Most banks require a Government ID to open an account, and additional IDs when giving out credit card. I don't see how it is different. And I think you can still use cash if you want to.

Does anybody have a term for the phenomenon where a technology achieves enough critical mass among early adopters for network effects to arise only to have the technology improve as it matures so the late adopters wind up leapfrogging the early adopters?

I see this all the time, a simple example being self check out machines. A chain in Atlanta adopted them in the late 90s (this is from memory, could be later) and they were fairly common in the city in early 2000s. They were awful/clunky to use but 5-6 years later I'm in the middle of nowhere South Carolina and find the process to be comparatively painless on an obviously newly-installed machine. Meanwhile the original technology is still installed/operational in the city.

I'm not too sure, but it sounds like you mean the law of the handicap of a head start[1], maybe?

[1] https://en.wikipedia.org/wiki/Law_of_the_handicap_of_a_head_...

Last mover advantage?

Sort of like magnetic strip credit cards in the US vs chip cards in Europe. Or the reasons Eastern European countries have the highest broadband speeds, having never dealt with investments in ISDN and other networks.

I am not sure cards are a good example.

In the UK at-least we had magnetic stripe and signature cards everywhere for years and years.

When chip and pin was introduced all the merchants had to get chip and pin terminals and all card holders had to get compatible cards.

http://en.wikipedia.org/wiki/Worse_is_better ?

https://en.wikipedia.org/wiki/File:Whytryharder.jpg

The Leapfrog effect: http://en.wikipedia.org/wiki/Leapfrogging

That is a very interesting article, I love to see the innovation in these scenarios. I also kept clicking around on articles there, it seems national geographic has some very good/interesting articles. I particularly enjoyed this one involving pay as you go solar power, for lights and phones: http://newswatch.nationalgeographic.com/2012/04/17/pay-as-yo...

It look like the fees can be summarized as ~2% (with a ~$1 minimum) split between sender and receiver with the full fee covered by the sender for out of system transfers. http://www.ifc.org/ifcext/gfm.nsf/AttachmentsByTitle/Tool6.7...

I thought the Philippines had cash transfer and payment by sms since 2006 (article is 2007) my friends use this so it is not vapour.

http://www.nytimes.com/2007/10/19/news/19iht-20oxan-PHILIPPI...

While this Phillippines SMS mobile payment service looks superficially similar and 7 years older, it is much more tightly linked to existing financial infrastructure, including bank accounts (optional) and cards (required).

Kenya's M-Pesa only requires a mobile.

Both country services require registration with government ID documents.

In addition, what has been partly astonishing about M-Pesa is the incredibly rapid growth in its use. Within 8 months of its launch in April 2007, it had over 17 million subscribers (50% of the population), i.e. on average over 2 million subs per month for a non-Internet service!

I'm not sure if it's exactly the same thing, but this sort of service has been available in Pakistan for a while now - http://www.youtube.com/watch?v=5kL--YSnFPo

M-pesa is awesome, it's been intriguing to me that at Kiva we've been able to do mobile payments for microloans in Kenya long before they will be easy enough to transact in the U.S.

With this system, you're giving your location, real social graph and financial transaction data to a single company. I think I'd pass.

Was I the only one avoiding this article because they thought it's about Kanye West?

This trend of third world countries taking a technology and running with it doesn't surprise me one bit. It's much more difficult to disrupt an established industry with new technology than to adapt a mature technology to create new industries.

I thought I read a story recently about how the Kenyan system was being plagued by mistrust. I'll see if I can find a link.

Edit: can't find the story.

It's a halwa network. It moves debt. It's not at all innovative. The only reason we don't do this here is because it's illegal.

In the us you'd have to pay $30/mo for unlimited payment via mobile phones in a 2yr contact.

And it would only work for the first $1000, then it will cost extra per dollar.

And it doesn't matter it works over means already provided/charged by the telcos, they will go out of their way to make it billable.

No offense, but Bitcoin already did this.

I don't know much about Bitcoin, but I think the distinction is in how many people are using the service. A big part of that seems to be platform compatibility. These people are able to use it on their cheap Nokia cell phones rather than needing a computer to use Bitcoin. The impression I got from the article was that payments are done via text message, so you may not even need a data plan for it.

I would agree, totally different use-cases.

Also, M-Pesa was launched in 2007, Bitcoin in 2009. So the "already" is a bit imprecise, too.

What's offensive is that you apparently didn't even bother to read the article or think about the differences -- both theoretical and practical -- between Bitcoin and this, but instead just jumped to a known buzzword.

@DiabloD3 INFORMATE BEFORE YOU SPECULATE...BRO

The fact that he is catastrophically wrong does not justify flaming him. Just downvote and move on.

This has nothing to do with bitcoin. Bitcoin is a completely separate currency with its own value.

This is just a digital solution to transmit existing currency.

A more apt comment would be that paypal (or any number of similar services) did this already.

In fact, though, none of these similar services have done anything close to this.

When you have 50% of the population using a digital payment solution for all of their money needs... then maybe you can make a comparison.

Internet hipster facepalm.

Actually, creating SMS bitcoin wallet is possible (however it won't be that easy to use). I will probably hack together SMS wallet support for easywallet.org . Meanwhile there is coinapult SMS wallet, but it is only for canada/US.

Reply on news.ycombinator.com