We started this post before a Greek deal rendered the discussion of a digital parallel currency moot. Nevertheless, it’s still worth looking to the Kenyan M-pesa for a better understanding of why it’s dangerous to treat fintech solutions as panaceas for economies struggling with productivity, poor credit profiles, tax collection issues and overall corruption without understanding what’s really at stake.
Kenya is often presented as an example of a country which has flourished thanks to mobile money adoption — the poster child that “digital payments can make the world a better place”.
But often forgotten is Kenya’s unique circumstances. The M-pesa mobile money system, owned and operated by Safaricom which is 40 per cent owned by Vodafone, was allowed an unchallenged monopoly in the country for a very long time.
This is important because Safaricom’s gain became the government’s monetary power loss due to seigniorage revenue contraction and the transfer of revenue from a licensed banking network to a single private telecoms company.
As a GSMA (association of mobile operators) report noted in January 2015, M-pesa faced a great deal of resistance from banks in the first two years precisely because of this monopoly oversight:
For instance, the banks complained they were not allowed to offer banking services through agents, yet MNOs were offering ‘deposit-like’ products through agents. The banks pitched for stringent regulations to ‘level the playing field’, and their lobbying efforts provided temporary reprieve. Even though the CBK had provided a monitoring framework for M-PESA in its letter of no objection, the Minister for Finance asserted that the absence of a legal framework to regulate and supervise mobile money was a gamble that might not pay off.32 The Minister then ordered the National Treasury and the Central Bank to conduct an audit of M-PESA.
Nor do people appreciate that M-pesa is a fully collateral backed money, meaning its “technology” is underpinned by the very same tech that underpins money-market funds everywhere: disciplined float management.
For every M-pesa digital unit in circulation there is a Kenyan shilling in collateral and/or (in the event that Safaricom wasn’t fully collateralising its digital money units) a Safaricom liability for a shilling. This collateral sits in a segregated custodian account in the Kenyan banking system backed either by “safe” government securities or other commercial assets deemed fit for purpose by the Trustee. As with every float-based system, the fund is exposed to both liquidity and collateral mark-to-market risk.
So how did this deposit-taking money market fund succeed in bringing digital banking to the financial excluded where banks failed? Largely, we’d argue, because it already owned the telco infrastructure that banks would otherwise have to pay to use and because it brought a telco brand name — Safaricom — to the Kenyan money market. This is important, because in terms of trust a telco sits somewhere between the incumbent banking system and the government in Kenya (if not above the latter).
But the key point remains. The only reason Safaricom was able to achieve success in Kenya is because the government in part turned a blind eye to its quasi monopoly status, not to mention the company’s veiled stab at operating as a better central bank than the Kenyan central bank. Had the system been rolled out in a competitive framework, chances are the network effects would not have been sufficient to overcome the costs of deployment.
But first a bit about how mobile adoption in Kenya was always going to be different to Greece
For one thing, Greece is a well banked country with a functioning digital payments network. The current Greek crisis was never connected to the banking network as much as it was to the quality of the state collateral used by the banking system to back its liabilities. The health of the bank network was in this way intrinsically linked to the health of the state’s ability to manage its debt load.
Second, there was never any scope for a mobile company to come in and issue fully collateralised units to stimulate economic activity with — because the problem was a lack of quality collateral to keep system liabilities in check with the value of euro-system liabilities, not the lack of a digital payments infrastructure or a need to unlock new internal trade relations.
To the contrary, if M-pesa has proved useful in Kenya it’s because it has brought digital payment services to areas that were previously unbanked, something that has arguably unlocked trade relationships which never existed before. This in turn has generated new economies of scale which have led to productivity and output growth.
The success side of the M-pesa story — and many academic studies agree — is closely associated with the output effect. Greece, on the other hand, is a mature economy with fewer unexploited synergies of this sort.
Yet even in Kenya, M-pesa’s output effect is nuanced, meaning the jury’s still out on whether an explicit link between mobile money and growth can be made. Some studies even suggest there could be unintended effects impeding monetary policy — and thus broad economic management — as well as those stimulating output.
Indeed, spare a thought for Kenya’s economy which despite the phenomenal adoption of mobile payment technology is still facing the same old economic challenges.
Kenya’s central bank, for example, is still worried about inflation. Last week it raised rates unexpectedly by 150 basis points to 11.50 per cent, barely a month after an emergency 150 basis point hike. As analysts at Barclays noted last week:
The decision to hike so aggressively was driven by the Committee’s view of heightened inflation risks as a result of the pressure on the exchange rate.
Meaning whatever output synergies mobile money may be encouraging in the country they haven’t yet “solved” the inflation problem. More from Barclays on that matter:
With the June 2015 Market Perceptions Survey showing increased inflation expectations among the private sector as a result of the anticipated pass-through effect of the recent exchange rate depreciation, it highlights a general concern across our coverage area of the likely effect of weaker FX on inflation. While Kenya’s inflation rate remained fairly stable at 7.0% y/y in June from 6.9% in the preceding month, fuel prices were pushed higher as a result of the weaker shilling and pressures are expected to intensify further after the KES moved to its weakest level since late 2011, moving above 100/USD.
The KES has weakened 10% year-to-date against the USD and weak tourism inflows, a likely stronger USD ahead, and firm imports do not bode well for the FX outlook and consequently inflation prospects. As a result, the bias remains towards further policy tightening.
Which leads to the obvious conclusion that mobile money tends to increase the velocity of money, and with it inflationary effects, as much as it increases trade and output. The velocity problem, meanwhile, is heightened if mobile money leads to greater remittance-fuelled purchasing in the local economy without associated import or output increases.
It’s worth noting that the the African development bank flagged the link between mobile money and inflation as early as 2012:
The increase in the velocity of money induced by these activities may have in turn propagated self-fulfilling inflation expectations and complicate monetary policy implementation. The monetary authorities may inadvertently follow looser monetary policy if the stock of e-money grows more rapidly than projected.
Should we be surprised then that as the central bank grows more concerned about inflation, the government becomes equally more focused on squashing M-pesa’s monopoly position.
When monetary policy means busting a payments monopoly
As Quartz noted last week, new regulations could soon force Safaricom to separate the M-Pesa business from its mobile phone services and infrastructure businesses, potentially weakening its position in the market and opening the door to competitors.
But here’s why busting a payments monopoly in a place like Kenya might have massively paradoxical effects, if not squash the M-pesa system entirely.
To begin with, there’s the issue of liquidity management and the additional cost to the consumer (or someone) of having to hop between multiple platforms. The slow take-up of M-pesa-like systems in other African countries is in part related to the poor economies of scale for users and providers operating in competitive frameworks.
For example, transferring payments within the M-pesa system may be cheap as chips for users, but gaining access in and out of the closed network (particularly out of) and into another comes at a cost. Think of it as a redemption/liquidity charge which masquerades as an exit fee for users and discourages them from redeeming wherever possible.
This fee system makes sense for Safaricom because it helps the telco protect itself from an outright run-on-the bank liquidity crisis whilst also helping it manage mismatches between the market value of entrusted funds and the promissory value of the e-money units it has issued.
But other challenges come in tow. For example, according to the GSMA report, as M-Pesa grows in popularity so does the size of the associated trust fund and with that the telco’s exposure to local custodian agents. Safaricom must now spread funds across several banks to reduce the risk of a single custodial bank failure. Furthermore, it must invest those funds in “safe” government securities or high quality loans linked to the local economy if its units are to keep tracking the shilling faithfully.
But how many safe assets are there really in Kenya, a B rated economy?
To dodge the sort of collateral risk that burned a hole right through Greek bank balance sheets M-pesa must eventually either take to making disciplined loan allocation decisions of its own (thus turning Safaricom into a bank, equally backed by self-created assets) or to investing in higher grade foreign collateral. The latter might be good for managing Vodafone’s credit exposure and defending the value of its own e-money units, but it would also amount to the formation of a privately-managed FX-backed reserve system which would undermine the power of the central bank and lead to potentially depreciating effects for the Kenyan shilling.
But let’s get back to the liquidity and credit risks M-pesa faces on a day to day basis.
As the below system chart from GSMA helps explain, operationally, the longer funds stay entrapped within M-Pesa’s own closed network (blue section) the more fluid and cheap M-pesa’s service is likely to be. To the contrary, the more redemptions in and out of the system, the bigger the liquidity and credit evaluation exposures for M-pesa, thus frictions and costs for users:
M-pesa’s incentive, consequently, is to turn itself into a liquidity-issuing monopoly backed by the country’s best quality collateral that sits tight on its balance sheet.
Sadly, with that model, there is a limit to how much high quality collateral can be absorbed.
When the limit is reached — unless the government is prepared to keep borrowing, putting the quality of its own collateral at risk — a telco like M-pesa would either have to entrust its own agents to issue liabilities against newly forged assets in accordance with quality-control and data/information capture conditions or, alternatively, absorb only those assets from the banking network which met its quality and data criteria. On the latter point, M-pesa might even be inclined to share its own data with the banking network, to help them make better loan decisions, and withhold M-pesa liquidity from those that fail to make the grade.
In short, over time, M-pesa would become the central bank.
As an aside, back in 2008 when this document was drafted, Safaricom’s competitive edge was further heightened by its ability to use interest revenue from the funds it held in trust “to defray the Trustee’s own costs of its role in the Service” but also “for such other purposes (whether charitable or not) as the Trustee, may in its sole discretion determine“, meaning for proprietary use, if not possible integration into the profit and loss account.
In August 2015, the Kenyan government ended this practice by establishing a formal legal framework for mobile money which banned e-money issuers from earning interest or any other financial return from funds held in its trust and forcing them to hand it over to charity.
Yet, if fragmentation is forced upon the market, not only could the cash-in cash-out savings associated with operating on one platform be lost, the additional liquidity and credit cost of transferring funds across rival platforms could up the cost of mobile payments services for everyone as a whole. Someone, somewhere will have to absorb the cost and risk associated with more players in the market and the associated breakdown of network effects.
Second, if fragmentation increases payment friction, there’s a good chance money velocity decreases with it. That may be a good side-effect for a government facing inflationary pressures, but it’s not so good for the profitability of telcos which depend on collecting transaction fees or rents.
There are plenty of EM governments seeking to mimic the “M-pesa” miracle on their own turf in a competitive framework. What few understand is that, unless, like in Kenya, the tech firms are allowed to achieve monopolistic control over the money supply, in a way that cedes state control of state scrip to an external private authority, none of them will prove successful.
Our simple point is M-pesa is not a technology. It’s a stealth political coup by a private operator which profits only from enforcing discipline, control and transparency (via massive data capture) over a wayward system.
