KAUSAR PARVEEN, of Chakwal district in the north of Pakistan’s Punjab province, is a star beneficiary of the work of Karandaaz, a Pakistani financial-inclusion charity. The owner of just one buffalo, she borrowed 75,000 rupees (about $650) to buy another one and started selling milk. The business has done so well she now has four buffaloes and an assistant, and has taken out another loan to install a biogas plant, saving on firewood and sparing her family the woodsmoke.

This was how microcredit, as promoted by Muhammad Yunus, a Nobel-prizewinning entrepreneur from Bangladesh who launched his Grameen bank in 1983, was supposed to work: credit would allow the poor to establish microbusinesses and improve their lives. The idea has spread across the developing world. Sadly, in many places it has not worked out that way. A big expansion of microcredit in India’s Andhra Pradesh province caused a crisis in 2010 when the lenders were blamed for an increase in suicides by farmers. A World Bank paper last November, written by Robert Cull of the bank and Jonathan Morduch of New York University, considered evidence showing that microcredit has had “only modest average impacts on customers”. It has often been used to cover the normal ups and downs of household spending, which is helpful but not transformative.

Part of the problem is that microfinance is very hard to provide on a large scale. Reaching, assessing and helping borrowers like Ms Parveen is time-consuming and labour-intensive, which makes it hard to keep interest rates at a reasonable level. Typical annualised percentage interest rates are in the region of 20-40%, cheaper than the traditional local moneylender or pawnbroker but hardly a snip. Digital money holds out the hope of improving things in two ways: by making it cheaper and faster to grant, disburse and repay loans and to provide other financial services, notably savings and insurance; and by harvesting data that should widen access to financial services for those with little or no history in the formal financial sector.

In Kenya, for example, Safaricom in 2012 launched M-Shwari, a paperless bank account offered by the Commercial Bank of Africa (CBA) via M-PESA. CBA takes the risk but can use the know-your-customer checks already done digitally by M-PESA to open the account, and the M-PESA payment history to gauge creditworthiness. Like M-PESA itself, it has grown like Topsy (CBA’s customer base increased from 50,000 in 2010 to 22m today) and has been much imitated across Africa and beyond. In Pakistan, FINCA, the global microfinance network, wants to use SimSim, its new mobile-money account, to offer “nano loans” (the equivalent of $5 or $10, say), thereby establishing a data trail for assessing bigger loans later.

M-Shwari and a few of its peers also offer services that pay interest on mobile-money accounts in credit. Indeed, the number of financial services available to poor people with a mobile-money account is exploding. Michael Schlein of Accion, a Massachusetts-based financial-inclusion non-profit, speaks of “a golden age of fintech”. Take life insurance. In Ghana, MTN, a mobile-network operator, offers a life-insurance product called Mi-Life linked to its mobile-money accounts. For about $0.23 a month users get cover of around $100. This is catching on across the developing world. In March Telefónica, a Spanish multinational network operator, announced a tie-up with Bima, a provider of mobile micro-insurance, to offer life insurance across Latin America, starting in Nicaragua. Crop and livestock insurance is also becoming available on mobile phones. A number of firms, such as Econet in Zimbabwe and Acre Africa in east Africa, offer farmers “index insurance” for their crops that will pay out automatically to a mobile-phone account, without the need to put in a claim, if, say, a rainfall index drops below a certain level.

Digital money should make it cheaper and faster to grant and repay loans, and widen access to financial services for those without a formal credit history

Ingenious pay-as-you-go schemes offer credit for purchases. The most famous is M-Kopa’s solar-panel technology, which has brought electricity to hundreds of thousands of homes in Kenya, Tanzania and Uganda. Buyers put down a small deposit and then make a daily payment from their mobile-money account until, after a year, they own the panel. If they miss a payment the panel is automatically locked, so if they urgently need money for something else, they have the choice of forgoing a day’s electricity to give them extra cash in hand. In February M-Kopa announced a partnership with MasterCard to help it expand through Africa, using the card firm’s QR technology. Again, good east African ideas travel: Easypaisa in Pakistan, for example, now has a similar offering. SimSim would like to use the model to finance smartphone purchases, but the technology to lock the devices remotely is not yet robust enough to rebuff attempts to outwit it.

The data generated by such accounts provide the nearest thing many of the holders have to a credit score. They are an invaluable aid to lenders trying to decide if a borrower can afford a loan. But a phone—and especially a smartphone—also provides all sorts of other information that some lenders may find useful for marketing or credit-assessment services. Positional data, for example, can show if someone has a steady job and a permanent address. Social-media activity can be highly informative. And shopping data can let on, say, if the user is pregnant.

Some firms specifically try to generate credit judgments in the absence of a conventional financial history. Lenddo EFL, a merger of two fintech startups, claims to have facilitated more than 7m assessments, allowing 50 financial institutions of all sizes to lend more than $2bn to people with limited borrowing histories. Lenddo relies on advanced AI-driven analytics. EFL provides “psychometric testing”—online quizzes that have a surprisingly good record in predicting a prospective borrower’s propensity to repay. Questions might be about how you are feeling; your view of the time value of money (“Would you take $10,000 now, or $20,000 in six months’ time? How about $17,000 now?”); how you spend your money; what you would do with a windfall; and how you view your community. If the questions seem easy to game, that is part of the point: the way that defaulters game it goes into the data. The algorithm will always be one step ahead. Lara Zibarras, a senior psychology lecturer at City, University of London, is working on another set of psychometric tests to be introduced by Oakam, a British subprime lender. They ask people to choose between photos to reveal personality traits. Early tests suggest they are as accurate in predicting missed first payments as an experienced human loan-underwriter.

The most extensive use of “alternative” data (which, unlike “alternative” facts, do have a basis in reality) is made in China. In 2015 the government awarded eight firms licences to develop consumer-credit ratings. Alipay’s is the most advanced. A good score from the firm’s Zhima (Sesame) credit agency may allow its holder to hire a car, use a bike-sharing service or book a hotel room without paying a deposit, and let him see a doctor without having to queue to pay. At one time, it is reported, it even allowed people to jump security queues at Beijing airport. Lonelyhearts flaunt their credit ratings in online-dating profiles. For those with a lower score, however, a Zhima rating may be risky. According to Xinhua, China’s state news agency, the database includes a list of more than 6m people who have defaulted on court fines, which has helped the courts catch up with more than 1.2m defaulters who found that their credit score had plummeted.

Open Sesame

Ant says that Zhima improves financial inclusion. As of 2015, the People’s Bank of China (the central bank) maintained credit histories for around 380m citizens. That is less than one-third of the adult population, compared with nine-tenths of Americans who have credit records. Zhima’s system, claims an Ant spokeswoman, is transparent. The five metrics on which it is based are indeed public: personal information, ability to pay, credit history, stability of social networks and “behaviour”. The meaning of this last one is not entirely clear. In 2015 Li Yingyun, a Zhima director, told Caixin, a magazine, that someone playing video games for ten hours a day might be rated a bad risk; a frequent buyer of nappies would be thought more responsible. 

As concern about the misuse of online data mounts in China, too, Ant now tends to play down such behavioural data. Douglas Feagin, its head of international operations (and a former Goldman Sachs banker), says its algorithms rely heavily on the debt-service and payment history: “Past repayment history is the best predictor of future credit performance.” In Lahore, Mr Shahid of FINJA is also sceptical of claims made for non-traditional data: “Everything is overrated except the payment history.”

For Ant, the credit score forms part of an “ecosystem” of online services that support each other. It also offers loans, and since 2013 has had a fund where Alipay users can earn interest on their surplus cash. The fund, known as Yu’e Bao (or “leftover treasure”), offers much higher returns than bank deposits. By the end of last year it had become the world’s biggest investment fund, with 1.58trn yuan ($243bn) in assets under management and 325m accounts, equivalent to nearly a quarter of China’s population. It has an estimated market share of 25%. Tencent has its own online fund, Licaitong, linked to WeChat, with 300bn yuan under management by the end of January this year. Lufax, a subsidiary of Ping An, an insurance giant, started as a marketplace for peer-to-peer lending but has turned itself into a financial “supermarket”, offering loans, securities, mutual funds, insurance and more.

These Chinese giants have shown that serving people who until recently were regarded as unbankable can be profitable. Greater financial inclusion, in effect, is a business opportunity. Institutions in richer countries are trying to heed that lesson.