More on M-pesa and e-money | FT Alphaville

More on M-pesa and e-money

We wrote about Kenya’s M-pesa mobile money model on Wednesday, which we think is a really innovative and encouraging development in the world of money supply.

The point we were trying to make at the time is that there are some interesting parallels between Safaricom’s role in the M-pesa e-money market and the role of central banks in conventional money markets.

And that to some degree Safaricom, which is 40 per cent owned by Vodafone, is now creating a money-like unit which competes directly with Kenya’s own national currency, the shilling.

(These observations also apply also to PayPal and/or other purveyors of e-money, it’s just that in M-pesa’s case the economic effects are arguably much easier to quantify and measure.)

Some of the feedback we’ve received seems to misunderstand our point. So we’re going to try to be a little clearer with our explanation. Specifically with regards to how Safaricom’s role is more akin to a central bank, than a conventional licensed lender.

A parallel currency system

The truth is that with M-pesa Safaricom has created something of a parallel currency system. A currency system which, as it happens, only Safaricom via M-pesa controls, since only it can create or extinguish M-Pesa units.

The operational details, according to Safaricom’s own documents, are actually very reminiscent to how ETFs work. In fact you could call it a giant Telecom-sponsored ETF that tracks the Kenyan shilling.

For example…

When you purchase e-money units, you deliver Kenyan currency to an agent, who creates fresh M-pesa units on your behalf with Safaricom/M-pesa.  The exchange of e-money for shillings is free, provided the correct cash-collateral (shillings) is delivered, and is something that adds to the M-pesa money supply. Assets under management go up every time shilling are deposited and units are created.

In that respect, the transaction can be considered an asset swap.

Why would anyone hand over interest-bearing shillings for non-interest bearing private e-money? Because, by and large, they don’t have access to a bank account — which means they have no ability to collect the interest that is owed to them themselves.

What’s more, the benefits associated with e-money  — liquidity and transferability — are seen to outweigh the lost opportunity to earn interest. There is, in economic terms, a preference.

And since every e-money unit is backed by collateral (Kenyan shillings) there’s also the reassurance that in the event of system collapse or bankruptcy, users still have a claim on the underlying. This, like we’ve noted before, is like having a central bank put option to guard against a fall in the value of the units, or from general mistracking. Like having the right to redeem gold in a gold-standard model, providing there is truly a gold unit backing every single monetary unit created.

Of course, while the M-pesa system is not leveraged (i.e. there is enough collateral to match every e-money unit created), there is an interest liability mismatch which the e-money issuer can (and does) exploit to their own benefit.

In the case of PayPal, the opportunity to earn interest is a key part of their revenue model. This sees Paypal distribute non interest-bearing e-money in exchange for customer cash balances (either as a custodian or as an outright claim) which it keeps in FDIC-guaranteed bank accounts. Paypal is only obliged to return principal upon redemption, meaning any interest it makes it can keep for itself.  This, incidentally, means Paypal’s revenues are very much exposed to any prolonged periods of low or zero interest rates.

In both cases the issuers are absorbing interest-bearing money from the system, replacing it with non-interest bearing e-money instead.

The unleveraged factor

As we’ve mentioned, Safaricom doesn’t leverage the cash balances it holds (borrow from them for its own purposes) but that doesn’t mean it couldn’t hypothetically borrow from them one day if it wanted to — though we appreciate their could be legal formalities to sort out first. Fractional reserve mechanics, however, do apply since it would take users redeeming units all at once to cause a collateral shortage (a.k.a bank run).

But the fact it chooses not to leverage the funds actually makes the system much more central-bank like. The collateral it holds provides the “backing” for the currency units extended. Another way of looking at it is that every M-pesa unit holder has an equity style claim on the collateral pool that backs the system.

The same theory, of course, applies to base money, which is core money produced by a central bank “backed” explicitly by collateral held in its account (and implicitly by the collective wealth of the nation). The fact that money supply grows beyond the monetary base, is less to do with central bank money creation on a leveraged basis and more to do with external factors linked to velocity, credit demand, and private banking.

And this really is how the M-Pesa system sets itself apart. Since only Safaricom can create M-pesa units, it’s much harder for intermediaries or licensed agents to add to M-pesa supply on a leveraged basis.

The collateral pool

So what happens to the Kenyan shillings you give up to create M-pesa? According to Safaricom’s own literature, that money is reinvested via the M-Pesa Holding Company into a Trust Fund on behalf of M-pesa account holders of Safaricom.

As the literature notes, the Trust Fund then does the following:

(E) It has been agreed that the amount of conventional money paid in respect of the creation of E-money will be paid to the Trustee and that all amounts so paid, less all amounts paid out at any time by the Trustee and that all amounts so paid, less all amounts paid out at any time by the Trustee in the redemption of E-money (such net amount in conventional money being the “Trust Fund) shall be held by the Trustee on trust for the System Participants as holders of E-money in accordance with the trusts declared below.

Investment of the Trust Fund and Interest

6.1 The amounts constituting the Trust Fund (including interest and income thereon) shall be held by the Trustee in such commercial bank accounts and such Government of Kenya securities as the Trustee shall in its absolute discretion determine (provided, in the case of securities, these must be debt securities which provide a certain return of the principal amount invested and which are capable of being realized into cash at short notice and without loss of capital).

6.2 Any interest or income received in respect of any investment of the Trust Fund shall be retained by the Trustee for its own account and shall not form part of or be credited to the Turst Fund and the Trustee shall have no obligation (express or implied, and whether as trustee or in any other capacity) to account to any System Participant for any such interest or income. Any such interest or income shall generally be applied first to defray the Trustee’s own costs of its role in the Service but may be applied for such other purposes (whether charitable or not) as the Trustee, may in its sole discretion determine.

The first thing to point out is that the Trust Fund has a relatively loose mandate. It can invest in any cash-like instrument providing it is debt-based, liquid and does not risk principal.

The second point is that the Trust receives all the interest it generates and has no obligation to return it to M-pesa holders, since E-money is non-interest bearing.  Interest received covers operational costs and beyond that looks to be retained as possible profit by the company. How it’s deployed is discretionary.

Also mentioned in the document is the fact that the Trustee shall not be liable for any loss of the Trust fund “arising in consequence of any decision made in good faith or by reason of any mistake or omission made in good faith or any other matter or thing except for willful fraud and/or wrong-doing or bad faith on the part of the Trustee”.  In other words, it runs the risk of investment failure and correlation breakdown (correlation with the net asset value of the fund versus a Kenyan shilling) like any other fund that aims to track a collateral basket.

By using a trust system, Safaricom is also itself removed from any liability in the event the Trust’s investment fails.


So what makes this such an exciting development?

First, the system is a private one rather than a public one. While this may leave the system exposed to the corporate fortunes and bankruptcy risk of its owner, we’d argue that doesn’t need to be a bad thing. Who says money has to be centralised at all? Looking to history, central bank money is actually a relatively new phenomenon, pre-dated by private money systems.

Furthermore, since the company is not leveraged, there is theoretically no “bank run” exposure. The units are fully backed. And if anything was to happen to the provider, the system could easily be transferred to another owner — providing the Trusts had been run properly — with no-one being less well off as a result.

The fees charged by Safaricom for transfers and redemption, as well as the interest earned by the Trust, obviously come at the disadvantage of the users. But since it is a corporate interest, we think that’s understandable. In an ideal monetary system, however, there’s something for the idea that such transactions fees would cover operational costs only.

Of course if the the system was to deployed universally with the strict objective of improving money velocity in conventional money markets, it might make sense to drop these transaction fees. As it stands the e-money units depreciate over time the more they are transacted. That said, the lack of interest on e-money counterbalances that effect, by making it less desirable to store or hoard e-money.

Ultimately, the beauty of the system is that it can be easily applied anywhere. All you need is a user-base that is happy to provide collateral in exchange for non-interest bearing units.

And given that, we think there may be some benefit to the ECB looking at the system more closely with respect to its own needs. A scheme that perhaps offers to exchange euros for M-euros which are then backed by the ECB’s existing collateral pool explicitly — a some small improvement on keeping euros as unsecured deposits at distressed periphery banks?

Related links:
M-euro, a lesson in money supply from Kenya
– FT Alphaville
Space opera, beyond finance edition – FT Alphaville
How Kenya became a world leader for mobile money – World Bank