This article was taken from the June issue of Wired magazine. Be the first to read Wired's articles in print before they're posted online, and get your hands on loads of additional content by subscribing online
In Chrouy Ampill a village in Kandal Province, Cambodia, 27-year-old Kakda Sun works as a weaver, making traditional shawls or kramas. Kakda Sun learned her trade from her mother on looms in the airy basement of the family home. But in 2007 she fell ill with appendicitis, and her family had to scrape together $300 for medical treatment. "I was working in a garment factory then," she says, via a translator. "It was a very harsh working environment. Sometimes when you were sick you couldn't ask for time off." Once she'd recovered, Kakda Sun sought out a loan so that she could buy rolls of thread, which cost $300 each, and a new loom ($500), which would enable her to work from home.
Banding together with fellow villagers, she turned to Angkor Microfinance Kampuchea (AMK), a microfinance organisation that specialises in microloans to the rural Cambodian poor. Operating in 6,253 villages across every province in the country, AMK has a loan portfolio worth $24,795,880, with 217,477 active borrowers, 85 per cent of whom are women.
Of the loans it makes, 91 percent are below $300. "When AMK started, we went to different provinces around Cambodia, looking at who was there, how poor they were and what their needs were," says Paul Luchtenburg, AMK's CEO. "We used those to go where there was the greatest need.
Most of our clients are trapped in their situation, and the opportunity to get a loan can be life-changing." It certainly was for Kakda Sun, who has since taken three further loans from AMK. In 2008, acting as president of her village "bank", she borrowed $4,975 which was divided among 35 villagers.
The loan was repaid in full and on time, with the last payment of $414.60 received in August 2009. In May last year, she and 15 neighbours took out another loan of $2,200, 75 per cent of which has been repaid.
The group's first eight repayments of $183.35 have been made on schedule. "Since borrowing money, I've been able to work at home," says Kakda Sun. "I can rest when I need to. We've also been able to buy a motorbike. Business is going well."
A total of 64 strangers financed her latest loan -- including Scott, a US marine from Dumfries, Virginia, who has served in the military for 24 years and believes in "helping others to help themselves". Scott came across Kakda Sun's profile on Kiva.org, a San Francisco-based peer-to-peer (P2P) non-profit, which uses the principles of social networking to connect individual or group lenders to entrepreneurs via microfinance institutions (MFIs) around the world.
When Kakda Sun applied for her loans, AMK approved her, advancing the cash, before creating a profile for her on Kiva's website. There, potential lenders browse through the pages, choosing the profiles they find appealing and uploading funds -- via PayPal, a credit card or their Kiva account -- to underwrite the loan. (Typically, a $1,000 loan on the site will be funded by 20 to 30 lenders around the world.)
Kiva borrowers and lenders such as Kakda Sun and Scott will never meet, yet the company has a remarkably high success rate:
98.46 per cent of loans made via Kiva are currently repaid. The company was founded in 2005 by Matt and Jessica Flannery after they attended a lecture by microfinance guru Muhammad Yunus at Stanford University in California.
Yunus argued that, given the chance, the poor -- particularly women -- would repay loans in order to preserve their access to credit. Kiva -- which means "unity" or "agreement" in Swahili -- made its first seven loans, a total of $3,500, in late 2005. Since then (at time of going to press), 330,560 entrepreneurs have borrowed $129,881,885 -- via 181,431 loans, administered by 111 field partners in 52 countries. "It's the personal story which makes me want to lend," says Russ Schoen, 36, a corporate trainer from Chicago, who has made 35 Kiva loans to entrepreneurs in 12 countries. "All my money has come back so far, and when I log on to Kiva and see there's $200 in my account, my first thought is, 'Who should I lend to next?'"
When Pierre Omidyar founded eBay in 1995, the company was inspired by his belief in "the potential of individuals and the power of the market". Starting from the premise that people are basically honest and seek to do the right thing, he built a social network around buying and selling and shared user interests. By incorporating incentives -- such as mutual feedback and transparency -- eBay became a marketplace based on informed strangers trusting one another. This ethos is the basis of Kiva's business and underpins a host of emerging ventures in sectors ranging from personal loans to venture capital, all based on informed co-operation.
Just as eBay shook offline retail to its foundations, P2P lending models such as Kiva, though still marginal, threaten to disrupt high-street banking. Although the public's faith in banks has been damaged and credit remains hard to come by, evidence suggests that a new trust-based economy is proving more efficient than traditional lending.
Bank of England data reveals that UK credit card companies wrote off £1,568 million in the third quarter of 2009, up from £747 million in the same quarter of 2008. In 2004, the equivalent loss was just £488 million. In the US, the delinquency rate on credit cards at America's largest bank holding-company, Bank of America, was 7.35 percent in January 2010. Kiva currently has a default rate of just 1.5 percent. According to Flannery, "We've lent $100 million, and one per cent of that was stolen."
Kiva's success lies in its stringent vetting procedures. Unlike banks and other lenders, the organisation does not vet its borrowers directly. Entrepreneurs in the developing world are digital ghosts, for whom electronic data does not exist. Instead, Kiva rigorously checks, evaluates and monitors its field partners -- the 111 MFIs scattered across 52 countries that it uses to disburse its loans. The selection process for prospective field partners begins with a desk review, followed by a site visit.
Only 42 percent of all MFIs considered for partnership make it to the "active review" stage. "There are about 10,000 MFIs on the planet, reaching about 100 million people," says Premal Shah, Kiva's president. "The estimated total demand for microfinance services is about 500 million people. The top 250 MFIs now attract commercial capital from Citibank and Deutsche Bank and so on, which means there's a very long tail of MFIs below commercial-investment grade. Those are known as Tier 2 and Tier 3 MFIs, and Kiva's capital is intended for them. "When selecting MFI field partners, Kiva looks at two factors: risk and "social performance". Shah says Kiva examines 71 variables on the risk side alone. "I've argued internally that we've gone too far," he laughs. "Those 71 variables include management and board, the audited financials of the institution, who else has funded them, sustainability, their internal controls and operation systems and MIS [management information systems]. Based on all that, we give field partners a one- to five-star risk rating. That rating then determines your credit limit. It's a lot like eBay, in the sense that as you improve your performance over time, if you demonstrate fantastic repayments on the Kiva platform, then you can improve your star rating."
Kiva also sends specially trained "fellows" to assess MFIs in the field. "Right now we have people in about 40 countries and they randomly sample ten borrower profiles," Shah continues. "They ask the loan officers at the MFI in question to take them to the ten borrowers that day. They verify that the borrower exists, that the dollar amount raised, the loan terms and the purpose of the loan are all accurate. If a borrower took a loan to buy a sewing machine, we need to see it. If one of those four things is inaccurate, the fellow writes to all the lenders letting them know." On the social-performance side, Kiva's model uses the Cerise audit tool, which gauges the social value of MFIs by assessing their principles through four "dimensions".
Dimension one examines outreach to the poor and excluded: how many clients are below the poverty line? Dimension two: adaptation of services to target clients. Dimension three: economic and sociopolitical benefits for clients and their families. Dimension four: an MFI's responsibility towards staff, clients, community and the environment. When assessing applicants for loans, MFIs use a process known as character-based lending. "Our field partners' credit committees do something banks used to do a hundred years ago," Shah explains. "When they first get to know a client, they literally ask the neighbours whether they are trustworthy. The algorithm for MFIs I've visited in west and east Africa, parts of the Caribbean and South America is character lending. They do reference checks, because there's no collateral or credit history. But you can get a very good sense of people that way."
Similarly, when MFIs grant loans to groups -- such as Miss Kakda Sun Village Bank Group -- they identify a group leader, who is generally one of the most trusted and respected members of their community. "We find that groups know and trust each other and are willing to guarantee each other," says AMK's Luchtenburg. "But with all loans we use the same principle -- start small, start simple.
Our loans are small, but that way, even in the worst-case scenario where someone gambles it all away, they can still find a way to pay it all back." And Flannery sees the current model as just a start. "[You can have] disintermediated payments via mobile phones, you can have borrowers becoming lenders, you could create a credit bureau for the poor," he says.
Since the credit crisis, venture-capital cash for start-ups and tech companies in the Series B (second-round) growth phase has been scarce, leaving some entrepreneurs scratching around for alternative ways to fund expansion.
In the digital hub of Shoreditch in east London, Trampoline Systems -- which describes itself as "a social-analytics business" -- may have found a solution with its very own twist on the trust economy. Trampoline had followed a conventional financing pathway for a tech venture, raising its initial finance from family and friends.
The business went on to raise $6 million from a US hedge fund in 2007 to invest in its Series A stage, which paid for advanced network and semantic analysis, with the understanding that as long as Trampoline hit its targets for that stage -- validating the fact that customers wanted what it was developing -- Series B investment would follow. "Then fate intervened," says chief executive Charles Armstrong. "The disruptions in the global financial system caused such devastation for the hedge fund that had worked with us, it withdrew completely from early stage venture investment. It was early 2009 and probably one of the worst periods in history to attract a new institutional investor. We were faced with a pretty stark set of choices and we looked at a bunch of alternatives which under normal circumstances we wouldn't have considered."
One of those alternatives was crowdfunding -- a P2P method of financing a venture by raising small sums from a large group of informed strangers -- which had hitherto only been used in niche industries such as movies and music.
Funding Series B development via crowdfunding had never been done before by a tech business of Trampoline's size, Armstrong says; he spent a month with lawyers to ensure he complied with Financial Services Authority (FSA) regulations. "We decided to see whether we could make it work for us and potentially create a template for others. We wanted to raise £1 million from up to 100 investors, with a minimum stake of £10,000."
It would have been against FSA legislation for Trampoline to have solicited backers; instead it could provide details only to people who expressed an interest and who self-certified as high-net-worth individuals. "Previous crowdfunding models tended to be based around membership of a club, for which you pay a fee, where you don't end up as an investor with shares but you get rights to a share of profits," Armstrong says. "We wanted to offer equity. In my view, for venture investment, the crowdfunding model is only going to be successful if it ends up putting equity in the hands of investors." Armstrong split the process into three tranches, closing the first tranche of £250,000 in October 2009, with shares priced at £1.30. This valued the company at £5 million. At that price, a minimum stake of £10,000 acquired 0.2 per cent of the company. Trampoline's investors gained full information and voting rights in the business. "We saw a ripple going through our own social networks," he says. "Introductions were made and that chain of trust extended.
A number of people who got involved were experienced investors who were connected to friends of friends." Crowdfunding of existing businesses is unlikely to replace the institutional "hub-and-spoke" model, but will become a significant counterpart, Armstrong predicts: "There are very clear inefficiencies in the conventional funding model, not least because they have an inbuilt bias towards certain types of transaction. The cost to a venture capitalist of due diligence and completing a transaction of £100,000 is more or less the same as for one of £10 million. So there's a pressure to invest larger sums in a single transaction. People have been talking about the funding gap for the past 20 years -- but the last 18 months have exacerbated that. Transactions from £100,000 to £1 million are going to turn out to be very well matched for crowdfunding and other innovative models."
SL75 is the user-name of a 34-year-old software developer in Telford. Two years ago, having tired of the high-street banks, with their sneak penalties and low rates of return, he stumbled across
Zopa.com, a British matchmaker for borrowers and lenders. "I'd always felt dissatisfied with the way the financial system worked," he tells Wired.
<img src="http://cdni.wired.co.uk/674x281/w_z/Wired Finance 140051.jpg" alt="Trust economy"/> "Here was a novel approach [to banking] in which end-users benefit from each others' cooperation. It just seemed a better use for spare funds than a bank, which would make more profit from my money than I would."
SL75 tentatively experimented by loaning a total of £500 through the site in small chunks to multiple borrowers. But when he read posts on the Zopa chat forum saying that higher and higher rates were being matched, he began to lend strategically.
Since joining Zopa (an acronym of the economics phrase "zone of possible agreement"), SL75 has lent a total of £32,540. He has had £10,426.30 of his capital returned, written off £1,889.31 in bad debt and earned £3,364.54 in interest. The rest of his money is still lent out, accruing interest.
In the summer of 2008, as the credit crisis gathered pace and surging demand for loans allowed lenders to offer their money on the site at interest rates topping 16 per cent, SL75 cleaned up.
His ledger shows that in the first two months of 2010 he achieved average interest rates from borrowers of 13.58 per cent and 16.10 per cent.
To date Zopa -- which launched in 2005 and is perhaps best described as an eBay for personal loans -- is the only UK service overseeing a significant P2P loans marketplace. But with about 25 similar companies popping up across the world -- ranging from Smava in Germany to DhanaX in India -- it's fast becoming clear that this boutique web industry may yet steal a slice of Britain's unsecured personal-loans market.
Digging into Zopa's numbers reveals that, at the time of writing, the site had issued 14,489 loans, worth £72,128,800; 104 borrowers had defaulted on a total of £471,591, giving it a repayment rate of 99.35 per cent. There is firm evidence that Zopa's business is showing prodigious growth: £33 million of that £72 million was lent last year, which means its 2009 loan book was worth the same as those of the previous three-and-a-half years combined. The company says that it is currently well on target to reach £50 million in 2010.
If Trampoline's experience proves VC firms can be sidelined in favour of a trust-economy model based around groups of informed strangers, the success of Zopa raises another question: why bother with personal loans from banks? Launched by part of the team behind internet bank Egg, Zopa is backed by Bessemer Venture Partners (backers of Skype) and Balderton Capital (its portfolio includes Love Film) and is based near London's Oxford Street.
Unlike Kiva, which achieves rock-bottom default rates by vetting its field partners, Zopa's closely guarded algorithms effectively screen out nearly all risky borrowers. When a prospective client applies for a loan on Zopa (which has an upper limit of £15,000), the company does a credit search. It takes a score from UK credit-reference agency Equifax and details from an online application form, and then gives applicants an internal credit rating, grading them from zero to five stars. "We're on our sixth iteration of risk models since we started -- and it's a proprietary piece of intellectual property," says Giles Andrews, Zopa's unstuffy CEO. "It uses a load of inputs -- some of which are given by the customer, such as 'I earn X', and some derived from the credit bureau." Red flags attach to those who are unemployed, or who have had long periods of absence from the UK or county-court judgements against them.
Of the applicants who reach this stage, Zopa declines "more than half". Andrews adds: "Most of the people we say no to, actually nobody should be lending money to them." Surviving applicants are then assigned one of five categories or markets -- A*, A, B, C, and Y (Y refers to 20- to 25-year-olds for whom there is little digital data). The "riskier" the market, the higher the typical APR charged.
At the time of writing, to borrow £5,000 over 36 months, an A*-rated borrower would be paying back £157.55 a month (a total of £5,671.95), at an APR of 8.7 percent, whereas a C-rated borrower would pay £165.00 per month (£5,939.96 in total) at 12.2 percent. "The reason we created these markets is that we want to create subsets of people to which we can assign a sensible risk measure,"
Andrews explains. "We tell our lenders, if you're going to lend to the A* market you should price in a certain amount to allow for defaults, and if you're going to lend to the Cs you need to price in more. As a borrower, you're told what market you're in and are given a quote. That quote is based on the cheapest offers made by the group of lenders on the market at that moment."
The cash is then put into a "silo", while the aspiring borrower is vetted even more thoroughly. "The problem with credit data is that it's all retrospective," Andrews says. "It's a whole lot of data saying that, for example, you've borrowed money 23 times and repaid it on the nose each time. What it doesn't do well is predict whether someone can afford a loan. We look at things like credit card data. Is there a pattern of credit-card debt going up? If there is, that means they're living on debt and they're not people you want to lend to."
A team of Zopa underwriters then verifies answers given in the application. For their part, Zopa's approximately 20,000 lenders -- who commit £1,902 on average -- set their own rates, and choose whether they want to lend for three or five years and to which market(s).
They typically lend in units of £10 or £20."So, when a lender joins, he sends us, say, £1,000 and wants to lend across all the markets at a rate of seven per cent," says Andrews. "We then offer his money in £20 units, because we believe that for our predicted bad-debt numbers to make any sense, you need to lend on a portfolio basis." Zopa takes its cut by charging borrowers a flat fee of £124.50 when a loan application is agreed, plus lenders pay a one per cent annual service fee. Despite its proven success on a variety of platforms, doubts linger whether P2P finance is truly viable as a scalable business.
Kiva can be considered a unique model -- a non-profit that uses existing MFIs to reach the world's poorest, which cannot be replicated commercially in the developed world -- and Zopa's story raises a key question: can this ever be more than a niche business?
Although issuing £72,128,800 worth of loans since 2005 might sound like a lot, it barely registers as a blip in the UK's unsecured personal-credit market. The outstanding amount held on UK-issued credit cards alone totalled £63.9 billion in December 2009. "This is not a boutique industry," Andrews insists, although he concedes that his business has yet to make a profit. "There is no reason why we can't be a multi-billion-number business."
But one expert on social-networking models demurs."Zopa's secret is that it uses a market place system combined with intense screening of borrowers," says associate professor Mikolaj Jan Piskorski of Harvard Business School. "It does not give its lenders the choice to decide on the creditworthiness of a borrower -- Zopa itself decides where to place the borrower. But it pays a very heavy price for this careful selection of individuals -- veryslow growth. The company essentially is only attractive to people who are low-risk but appear as high-risk to banks - and if truth be told, there simply aren't very many people like that. So even though its default rates are low, the business will take a substantial amount of time to scale, if ever. It is easy to see how most banks are not afraid of peer-to-peer lending."
If P2P finance has yet to prove scalable or profitable, it's also true that, not so long ago, the same was said of other web ventures which went on to change the world.
This article was originally published by WIRED UK
