IT IS A measure of how fast and unpredictably technology and finance have developed that the two most influential new payment systems of the 21st century so far both came about more or less by accident. M-PESA, Kenya’s mobile-payment system, evolved out of a pilot scheme in 2005 by Safaricom, the country’s biggest mobile operator, financed by DFID, the British government’s aid agency. Its researchers had noticed that Kenyans were transferring mobile-phone airtime between each other as if it were money. They thought this might offer a way to handle microcredit repayments, reducing costs.

Alipay, a smartphone-based payment system now ubiquitous in China and spreading fast abroad, has its origins in a service devised for Taobao, an online platform run by Alibaba where small businesses sell direct to Chinese consumers. Customers were reluctant to pay for goods before they had received them. So buyers would send their orders by fax to Alipay to hold their money in escrow and release it when delivery was confirmed. In 2008 this system was transformed into mobile “wallets” in which the money is held.

On Singles’ Day last year, a day of frenetic online commerce, Alipay processed $25bn in transactions, 90% of them via mobile phones

Safaricom turned M-PESA into a general money-transfer system which became the most popular way of moving money around in Kenya. Account-holders (who now number nearly 30m) pay money in by handing cash to one of Safaricom’s 148,000-plus agents, typically corner shops that were already selling scratch cards to top up mobile phones. The cash can then be withdrawn at another agent or transferred to another M-PESA account-holder. That allows people working in the cities to send money back to their home villages faster, more cheaply and more securely. Other services have been added over the years. M-PESA has expanded abroad and spawned dozens of imitators.

Almost all of them are tiny compared with Alipay, which has 520m active users, almost as many as all the mobile-money accounts held in the rest of the world put together. It hopes to increase its customer base to 2bn worldwide by 2025. Ant, founded only in 2014, is expected to list on a stock exchange next year. It is reported to be seeking an earlier round of funding which would value the company at $150bn (for comparison, Goldman Sachs is valued at about $100bn).

The volumes its systems handle are staggering. On Singles’ Day (November 11th) last year, a day of frenetic online commerce, Alipay processed $25bn in transactions, 90% of them via mobile phones. The only mobile-payment service that comes close to Alipay’s scale is WeChat Pay, offered by its Chinese rival Tencent, a social-media giant. It has reduced Alipay’s share of the Chinese mobile-payment market from above 80% to just over half. Most Chinese use both systems.

M-PESA and Alipay follow very different models. M-PESA was designed for a simple feature phone, working from a text menu of options (though it is now also available as an app). Alipay is available only as a smartphone app, linked to a bank account, reflecting the rapid uptake of internet-enabled phones in China. Payments are made by Quick Response (QR) codes, the square black-and-white dot matrices that have become ubiquitous in China. Even some beggars accept them.

Both systems have made big inroads into financial exclusion. A study in Kenya quoted in the Findex by two economists, Tavneet Suri of MIT and William Jack of Georgetown University, found that access to M-PESA increased consumption levels and lifted 194,000 Kenyan households (2% of the total) out of poverty. In China the absolute number of adults without an account, at 225m, is still larger than anywhere else in the world. But 82% of the unbanked have mobile phones, compared with about two-thirds globally. Already 40% of adults in China make mobile payments, and 85% of those who make purchases on the internet pay for them online (globally, more than half of online buyers pay cash on delivery). In a recent paper the Consultative Group to Assist the Poor, a partnership of development groups based at the World Bank in Washington, DC, pointed out that 44% of China’s people live in rural areas, where connectivity can be a barrier. In the countryside 71% of residents still do not use the internet, compared with 33% in urban areas.

Both the “Chinese” and the “Kenyan” models have crossed borders. Most developing countries have a mobile-payment service, but Sub-Saharan Africa is the only region where the share of adults with a mobile account exceeds 10%. Tencent has an e-payment licence in Malaysia where it plans to launch WeChat Pay—its first foray outside China and Hong Kong. Alipay has taken a higher-profile approach, enlisting merchants in Europe and America to accept it as a means of payment for the benefit of Chinese residents and tourists. And in Asia itself, Ant Financial has been investing in local mobile-payment services in India, Indonesia, Malaysia, the Philippines, Singapore, South Korea and, most recently, Pakistan.

The Chinese are coming

This last investment, of $184.5m, is to buy 45% of Telenor Microfinance Bank (TMB), which manages Pakistan’s biggest mobile-money service, Easypaisa. Owned by Telenor, a Norwegian multinational mobile-network operator, TMB launched Easypaisa in 2009. Competitors in Pakistan view Ant’s arrival with some foreboding. “They are here not to save the poor Pakistani but to promote e-commerce,” says one local microfinance lender.

It is hardly surprising that many in this industry, rooted in charitable development work, feel ambivalent about vast commercial enterprises entering the payment business. The suspicions are not confined to Pakistan, and are likely to become more acute as American and Chinese tech giants slug it out for market share in poor countries (see box). As a still largely nascent market of enormous potential, Pakistan also illustrates many of the other tensions affecting the payment business.

One is between the desire of both governments and businesses to digitise payments swiftly and the capacity of the population to go along with that. Moving away from cash payments reduces costs, cuts leakages through corruption, discourages the informal economy and increases the tax base. The poor may be equally quick to realise that mobile money is more secure from robbers, can save them hours of travelling and queuing and may open up a range of financial services. But they may struggle to afford even a simple feature phone, and the illiterate and innumerate especially may find using it daunting at first.

In Pakistan that covers a big chunk of the population. The overall adult-literacy rate of 58% hides lower shares in the countryside (49%) and among women (45%). The drive for financial inclusion may not narrow the gender gap. Pakistan’s Benazir Income Support Programme (BISP), which offers cash transfers to the neediest women, seemed a good way to do that, but making it work has not been straightforward. Agents delivering the cash would take a cut. Almost all Pakistanis have a digital identity stored in a national database that helps them open a bank or mobile-money account. But giving BISP recipients debit cards linked to this so they could get the money from an ATM also sometimes meant that middlemen took the cards, withdrew the money and skimmed a commission. Mobile money works better, but it still usually involves a visit to an agent. The number of mobile accounts held by women in Pakistan rose by an impressive 4m in the 12 months to September last year, to 7.3m, but those held by men increased by an even more remarkable 12m, to 25.6m.

Similar obstacles slow down the move from “cash-in-cash-out” (CICO) systems to those in which mobile money is accepted for day-to-day purchases. (Easypaisa is largely an “over-the-counter” system in which both sender and recipient use cash and the digital money moves from one agent to another.) The aim is to increase the number of individual mobile accounts, and then of mobile payments. But so long as other shops accept cash, an individual shopkeeper has little incentive to accept electronic payments. And a new study by McKinsey finds that CICO is still crucial to current business models for mobile money, accounting for about 60% of profits (see chart).

Cash is here to stay. “It works quite well,” notes McKinsey’s Ms White drily. Even in Norway, where digital payments have a bigger share than anywhere else, 17% of all payments are still in cash. But digital payments will become easier and more common. “Tap-to-pay” methods using near-field communication technology that have taken off in Europe, and the EFTPOS (electronic funds transfer at point of sale) machines ubiquitous in rich countries, may be supplanted in many developing ones by an app on a small retailer’s smartphone. In Pakistan, as in much of the world, this is likely to be one that can read a QR code. M-PESA in Kenya, for example, is rolling out “scan-to-pay” as well as “tap-to-pay” services among its merchants.

Although cumbersome, electronic payments are possible on a feature phone, and some such phones have cameras that can read QR codes. But a smartphone makes them much easier, raising another tension: between feature-phone-based services and internet-enabled phones. In Pakistan the local subsidiary of FINCA, a global non-profit microfinance network, has a joint venture with FINJA, a local fintech, marketing a mobile wallet for smartphones called SimSim. That seems perverse in a country where smartphones account for only about a quarter of mobile connections. As elsewhere, however, that proportion is rising rapidly thanks to cheap Chinese handsets. Qasif Shahid of FINJA argues that in the modern world those without a smartphone lack a digital identity and are not really “included”. Designing systems for them that rely on feature phones is “a ploy for people to continue to belong to the have-nots”.

A final tension is between competition and the concentration implied by network effects. Rich countries have burgeoning choices. Sit in a taxi in Singapore and the window is obscured by stickers advertising different ways to pay for the ride—credit cards, debit cards, stored-value cards and any number of smartphone apps. Shop counters groan under the weight of all the EFTPOS machines. But in frontier markets the brave pioneers of mobile money tend to become near-monopolies. M-PESA has 80% of the market in Kenya. In Bangladesh the central bank has licensed 27 services, and Kamal Quadir claims a market share of only 60% for bKash. But his network of 176,000 agents is hard to match. As he says, “you need network effects and scale to be effective.” In Pakistan, Easypaisa and JazzCash, its biggest rival, have a market share of about 85% between them.

One way of both fostering greater take-up of digital finance in the short term and mitigating the long-term risks of monopolies is to embrace “interoperability”, allowing payments across different systems. To this end the Gates Foundation has collaborated with a number of fintechs, including Ripple, a highly valued distributed-ledger developer, to create free open-source software. The result, a system called Mojaloop (moja means “one” in Swahili), makes it easier to deploy interoperable payment platforms. The idea is to ensure that the very poor will have access whatever happens to the rest of the market.

For now, intense competition in most countries means that disadvantaged consumers should indeed benefit from the rise in mobile money. But competition is fierce in part because network effects imply that the winner takes all. And as transferring money gets ever closer to the goal of free, frictionless, real-time payments, what will matter is not so much the process itself as the additional services the provider is offering.