David Roodman

[Update: Matt Flannery, CEO and Co-Founder of Kiva, replied to this post as a guest blogger. Kiva has also changed its site, and I have blogged more.]

This post is so long it needs an executive summary.

Kiva is the path-breaking, fast-growing person-to-person microlending site. It works this way: Kiva posts pictures and stories of people needing loans. You give your money to Kiva. Kiva sends it to a microlender. The lender makes the loan to a person you choose. He or she ordinarily repays. You get your money back with no interest. It's like eBay for microcredit.

You knew that, right? Well guess what: you're wrong, and so is Kiva's diagram. Less that 5% of Kiva loans are disbursed after they are listed and funded on Kiva's site. Just today, for example, Kiva listed a loan fepor Phong Mut in Cambodia and at this writing only $25 of the needed $800 has been raised. But you needn't worry about whether Phong Mut will get the loan because it was disbursed last month. And if she defaults, you might not hear about it: the intermediating microlender MAXIMA might cover for her in order to keep its Kiva-listed repayment rate high.

In short, the person-to-person donor-to-borrower connections created by Kiva are partly fictional. I suspect that most Kiva users do not realize this. Yet Kiva prides itself on transparency.

I hasten to temper this criticism. What Kiva does behind the scenes is what it should do. Imagine if Kiva actually worked the way people think it does. Phong Mut approaches a MAXIMA loan officer and clears all the approval hurdles, making the case that she has a good plan for the loan, has good references, etc. The MAXIMA officer says, "I think you deserve a loan, and MAXIMA has the capital to make it. But instead of giving you one, I'm going to take your picture, write down your story, get it translated and posted on an American web site, and then we'll see over the next month whether the Americans think you should get a loan. Check back with me from time to time." That would be inefficient, which is to say, immorally wasteful of charitable dollars. And it would be demeaning for Phong Mut. So instead MAXIMA took her picture and story, gave her the loan, and then uploaded the information to Kiva. MAXIMA will lend the money it gets from Kiva to someone else, who may never appear on kiva.org.

Moreover, the way Kiva actually works is hidden in plain sight. On the right of Phong Mut's page, you can see that MAXIMA lent her the money on September 8 and listed her on Kiva on September 21. So while Kiva is feeding a misunderstanding, it isn't technically hiding anything.

And finally in Kiva's defense, its behavior is emblematic of fund-raising in microfinance and charity generally, and is ultimately traceable to human foibles. People donate in part because it makes them feel good. Giving the beneficiary a face and constructing a story for her in which the donor helps write the next chapter opens purses.

Our sensitivity to stories and faces distorts how we give, thus what charities do and how they sell themselves. What if the best way to help in some places is to support communities rather than individuals? To make roads rather than make loans? To contribute to a disaster preparedness fund rather than just respond to the latest earthquake? And how far should nonprofits go in misrepresenting what they do in order to fund it? It is not an easy question: what if honesty reduces funding?

The big lesson is that the charities we observe, the ones whose pitches reach our retinas, are survivors of a Darwinian selection process driven by our own minds. An actual eBay venture called MicroPlace competes with Kiva; but MicroPlace is more up-front about the real deal. Its page for sample borrower Filadelfo Sotelo invites you to "invest in the organization that helped Filadelfo Sotelo: Fondo de Desarrollo Local" (FDL). This honesty is probably one reason MicroPlace has badly lagged Kiva. Who wants to click on the FDL icon when you can click on a human face?

Nicholas Kristof once tweeted that he "Just made a new microloan on www.kiva.org to a Nicaraguan woman. Great therapy: always makes me feel good." We should not feel guilty about the pleasure of giving. It should not just be eating your brussels sprouts. Indeed, Kristof might argue that Kiva.org's ability to make the user feel good is its greatest strength, for it draws people into an experience that stretches their horizons, educates them about global poverty, and entices them to contribute money they might otherwise spend on potato chips.

Still, we should take responsibility for how our pursuit of that pleasure plays out. Surely it is better to invest in an institution such as FDL without requiring it to incur the expense of posting pictures and stories of every borrower. Historically microcreditors have scaled to reach millions of people by cutting costs to the bone. Surely it would be better for us to give in a way that allows the microfinance institutions to put more of their limited energies into helping poor people manage their difficult lot and less into making us feel good.

I do not know the full answer to this conundrum, this tension between the need to draw donors and operate efficiently. Still, subtle dissembling makes me uneasy, perhaps because good intentions so often go awry. If a charity obscures how it operates, should we trust its claims about its impacts?

My wife Mai heard someone say that the world needs both playwrights and critics—if more playwrights. I treasure this observation because, as this blog must make obvious, I'm a critic. I can testify that being a critic can be bruising, especially when the playwrights you critique are alive. It's solace to think that the world needs me.

But the observation also helps me appreciate playwrights. They are the people who create things that weren't there, the people who are a tad insane in the sense that they confuse fantasy and reality. They see something in their mind's eye and believe they can make it real. Precisely because I am not like them, I hold playwrights---visionaries---in some awe. The most skillful, passionate, and lucky of them "put a dent in the universe" as Steve Jobs said. (An early employee described Jobs's uncanny ability to create a reality distortion field that altered bystanders' perceptions of the technologically possible.) Without playwrights, we might be still living in caves. At least, we wouldn't have iPhones.

We also probably wouldn't have the Grameen Bank, BRAC, and dozens of other successful microfinance institutions (MFIs) made by driven visionaries. And we wouldn't have Kiva, the person-to-person microcredit web site founded by Matt Flannery and Jessica Jackley.

On the other hand, without critics---analysts driven to understand the world rather than change it---we might not have mastered electricity. So we needed them too to get to iPhones. Critics and playwrights are yin and yang. Of course the two essences exist within all of us.

Critics seem to parse matters into quantities and concepts while playwrights seem to speak, and perhaps think, more in pictures and stories. (Or am I over-reaching here?)

Like most innovations, Kiva is not entirely new. Rather, it is an ingenious fusion of older ideas. One is child sponsorship, which Save the Children pioneered in 1940. A family in a rich country sends $10 or $20 each month to a designated child in a poor country via a charity. In return, the family receives a photo and an update at least once a year. When I was perhaps eight, my family sponsored Constance, a Greek girl about my age, through Save the Children. I remember looking at her solemn face in two successive black and white portraits, trying to judge how much she had grown in a year.

Child sponsorship grew explosively in the United States in the 1990s, thanks mainly to groups with names like the Christian Children's Fund, Children International, and Childreach (now Plan International). Then an exposé in the Chicago Tribune in March 1998 brought it crashing down (hat tip to Tim Ogden). Starting in 1995, editors and reporters at the paper sponsored a dozen children in such countries as Guatemala and Mali. Then the reporters tracked down the children:

The Tribune's yearlong examination of four leading sponsorship organizations...found that several children sponsored...received few or no promised benefits. A few others received a hodgepodge of occasional handouts, such as toothpaste, soap and cooking pots. Some got clothing and shoes that frequently did not fit.

Sick children were sometimes given checkups and medicine, but not always.

One child, a 12-year-old Malian girl sponsored through Save the Children, died soon after being sponsored, although the charity continued to accept money on her behalf for nearly two years after her death. A subsequent investigation by Save the Children found that at least two dozen other sponsors had sent the charity money on behalf of dead children in Mali for varying periods of time, in two cases as long as five years.

There was more to the story. Clover and John Dixon of Bellingham, Washington, received faked New Year's letters from a West African child who had died in a donkey cart accident. Sponsorship peddlers sent heart-string-tugging appeals for extra $25 contributions on birthdays, Christmas, Easter, and the purpose-built International Hug Day. Childreach ran a disastrous experiment in Ecuador with a novel intervention called "microcredit." Local workers embezzled funds; in protest, borrowers burned loan documents.

Undoubtedly some hard-sell charlatanry was at work. But the problem was deeper than that: a tension between creating the psychological experience of connection that raised money and the realities of fighting poverty. Often the fairest and most effective way to help poor children is by building assets for the whole community such as schools, clinics, and wells. Often charities contract with locals to build these things. Often things go wrong because of corruption, bad luck, or arrogance among outsiders thinking they know what will work. In the best cases, charities learn from failure. All these factors break the connection between giving and benefit, sponsor and child. But admitting that would have threatened the funding base:

"For a segment of the public, there will be nothing else that will reach those people the way that child sponsorship does," says Charles MacCormack, the president of Westport, Conn.-based Save the Children, the nation's oldest and best-known sponsorship agency.

...

As MacCormack puts it, "An awful lot of people who sign on to a personal human being will not sign on to a well."

...

"[The charities] are addicted to it, because if they stop, they lose their identity as Save the Children," says Michael Maren, a veteran aid-agency worker in Africa and author of "The Road to Hell," a book critical of private foreign assistance organizations including Save the Children.

"That's their thing," Maren says. "They invented it. That's their problem. The Catch-22 is that the only way to raise money is sponsorship, but that is not the way to development. The show is the biggest part of what they do. So, they say, let's keep the show going, but try to find ways to make it better."

Within a year of the Tribune series, the Missouri attorney general had slapped restrictions on Children International while the non-profit umbrella group InterAction committed to developing a set of voluntary industry standards. Many of the rule changes related to how clearly the charities disclosed how they operated.

Matt Flannery penned a history of Kiva's first two years for MIT's innovations journal in 2007. Two years later, he wrote a second installment in the same periodical. Flannery's authentic, conversational voice makes both articles readable and engaging. As he tells his own story, he comes across as an approachable man of vision, passion, and action.

Flannery tells how another Kiva ingredient, microcredit, first mixed in his mind with child sponsorship. Fittingly, it happened through hearing a story:

One night, [Jessica] invited me to come hear a guest speaker on the topic of microfinance, Dr. Mohammed [sic] Yunus. Dr. Yunus spoke to a classroom of thirty people and shared his story of starting the Grameen Bank. It was my first exposure to the topic and I thought it was a great story from an inspiring person. For Jessica, it was more of a call to action that focused her life goals.

Some months later, Jessica went off to East Africa to perform "impact evaluations" for the Village Enterprise Fund, which works intensively with poor farmers, providing grants (not loans) and training to help them start business activities. Jessica's work gathered data on indicators of poverty among participants, asking "questions like 'Do you take sugar with your tea?' and 'Do you sleep on a mattress?'." The couple kept in touch by phone. Then came the epiphany:

When the words “Sponsor a Business” entered our phone conversation, it set off a chain of ideas. We had both grown up sponsoring children in Africa through our church and families. Why not extend the core of that idea to business? However, instead of donations, we could focus on loans. This seemed like a dignified, intellectual, and equitable extension that appealed to us at this point in our lives. Instead of benefactor relationships, we could explore partnership relationships. Instead of poverty, we could focus on progress.

Soon after, Matt joined Jessica in Africa. He brought his video camera, which embodied the third key ingredient in Kiva, information technology. "I planned to spend most of my time making a short documentary of small business stories. I was also intent on investigating the viability of our new idea."

Back in the United States Matt and Jessica began their impressive passage across the desert in pursuit of their vision. She networked for advice and support. He built the website after-hours, and eventually quit his job. Together they wrote the business plan.

Once the site was ready, we needed loan applications in Africa to post on the site. That’s where our friend Moses came in. Moses Onyango is a pastor in Tororo, Uganda, whom Jessica had stayed with after I left. Moses is a community leader in Tororo and is highly connected to the Internet. We had been in close contact over the past year and Moses was ready to post and administer the loans of seven entrepreneurs in his community....

Once Moses had posted the seven businesses, the site was ready to go. We sent out an email to our wedding invite list and waited to see what would happen. We emailed about 300 people, and all seven businesses were funded in a weekend. That was April 2005, and we raised $3,500 in a few days. We were blown away; everything worked.

Right there, Kiva hit the tension in the sponsorship---more currently, "person-to-person" (P2P)---model: the need to find and post enough stories to keep up with demand. It led instantly to fraud, though Matt Flannery didn't know it when he wrote the two-year history. As he recounts in the four-year history, a Kiva Fellow (volunteer) sent to Uganda discovered that Moses was producing many stories about individual borrowers the easy way, from whole cloth. Flannery flew to Uganda:

I spent two weeks organizing a clean-up operation. We hired accountants and lawyers. I spent hours with Moses, trying to figure out exactly what happened. He was very apologetic, but our conversations didn’t go anywhere. The money had vanished into a series of bad investments and a new house. Moses had a growing family. His new son was named for me: Matthew Flannery Onyango.

Admirably, Kiva went public with the information:

...we alerted our users that not all of their funds made it to the intended recipients....The reaction from our user base was telling. Overwhelmingly, they thanked us for our honesty and poured their refunds back into loans to other MFIs on the site. They reinforced an important lesson: whenever possible, be completely transparent. Transparency pays huge long-term dividends.

...

If you are running an organization and are considering withholding valuable information from your customers, just don’t. There are a million reasons to withhold information. Lawyers will warn you about liabilities. Marketing people will preach about tarnishing the brand. Investors will encourage you to look bigger and better than you are. Most of this is just tired and outdated thinking.

Operating transparently is a great way to keep an organization accountable for its actions. Before you act, ask yourself: would you be OK doing this if you had to tell your entire user base about it? Would you be proud if your actions were described on the front page of the New York Times? These are great tests that I often use to vet a decision.

Flannery describes the "story factory." Running one---collecting and posting stories---imposes a significant expense on MFIs but is evidently offset by the low 2% 0% (hat tip Ben Elberger) interest rate that Kiva charges on capital:

Out in Cambodia, I got to watch firsthand how a sophisticated MFI gets content on the site. It is quite an operation….

In the field, loan officers carry Kiva questionnaires along with a host of other loan documents. When they visit a village, they gather women and tell them about the opportunity to apply for a loan. If a woman decides to apply, the loan officer takes down information on paper---some for the Kiva site and some for other business purposes. The Kiva questionnaire asks for information that interests lenders. For instance, how many children do you have? And how will the loan make an impact on your family? This is all done in the local language---Khmer. They also take photos of the applicants.

Returning to the branch, the loan officer enters the data into a computer and sends the information---via Yahoo! Messenger---to the Kiva coordinators at the headquarters in a major city. Kiva coordinators are typically young, Internet-savvy males who get paid a few thousand dollars a year. It is a desirable job and about ten of them are now working in Phnom Penh. We train them in the art of synthesizing the Kiva questionnaire into a readable narrative; then they spend their days writing stories and uploading pictures.

...

As a kid, I would write letters to [sponsored] children a few years younger than me in Africa and South America. I imagined my letters being delivered to a thatched-roof hut halfway around the planet. It sparked my imagination and gave me a sense of connectedness. Through Kiva, we can provide some of that to a new generation of kids.

Looking back now, I imagine that the transaction wasn’t as simple as I had thought. A lot of intermediaries were involved, lending a certain production quality to the experience. Plus, it was expensive. Delivering the child sponsorship experience was often as expensive as the child sponsorship itself. At Kiva, it’s not as simple as it seems, either…

innovations invited to Sam Daley-Harris, who was central to teaching Americans about microfinance and serves on Kiva's advisory board, to comment on Flannery's four-year retrospective. While praising Kiva's "profound contribution to the field of microfinance and international development," he worried about the transaction costs, and noted one other concern:

...there is still a bit of deception in the notion that the moment that a loan is funded, the client in Kenya or Cambodia receives his or her microloan with those particular dollars. Indeed, there are real people receiving real loans to start or grow real enterprises, but if a client in a remote village qualifies for a loan, the MFI will not likely make that client wait for the Kiva lenders to put up that last $25. Said another way, loan funds are fungible, and a larger MFI on Kiva’s website will use Kiva’s loans as one important source of their lending pool, but it’s not actually those precise dollars going to that precise client.

As I noted at the top, Sam is right. In fact, I wrote a little program in Excel to extract data from kiva.org. It shows that for September 2009, only 4.3% of loans were disbursed after Kiva users had fully funded them through the site. And probably some of those the local lender had already committed to make before Kiva users had funded them. And in a new report on what happens to investors when microfinance institutions collapse, Daniel Rozas computed from data on kiva.org that the failure of just three lending institutions caused 93% of all Kiva defaults to date. No doubt many of those institutions' borrowers were faithfully repaying at the time of collapse. Conversely, if a borrower defaults, the lender will often cover for him in order to maintain a good reputation on Kiva. So whether you get your money back as a Kiva user depends overwhelmingly on the solvency of the lenders, not the borrowers.

Kiva deserves kudos for being transparent enough for Rozas and myself to extract such data. But I wondered whether Kiva might become the Save the Children of P2P microcredit, the reasonably responsible pioneer who is imitated and overtaken by less scrupulous actors who pull the whole industry down a muddy slope into hucksterism. So I checked out MYC4, Wokai, and Babyloan (motto: "micro credits, great stories"; and no, it doesn't make loans to babies: it's French). To my surprise they were more honest about the P2P relationships they (seem to) forge. Here's Babyloan in enjoyably imperfect English:

Note: Babyloan works as a REfinancing platform and not as a direct financing system. It can happen that the MFI already "advanced" the microcredit to the entrepreneur when you make the online social micro loan. Indeed, as we are still in a launching phase and particularly for seasonal projects , we did not want to make the realisation of the project  "dependant" on the Internet users' good will and click. Babyloan is no microfinance reality show of ! However, we limit the funding time of the project not to create too much time discreprency between the projet and your micro loan, so your money is really used to finance the project. After the delay of 3 months maximum , we send all the money even if the funding has not been completed by the Internet users. The MFI will complete the funding.

So these sites are refinancing mechanisms. Kiva-linked microlenders make loans, then "sell" them to Kiva and its users. Might we rescue the P2P conception by observing that the lenders make their loans anticipating refinancing on Kiva? Yes, but only partly. Kiva limits itself to providing at most 30% of any lender's capital. So a lender will make at least three loans for every one it chooses to post on Kiva (hat tip to Molly's dad).

Kiva brings microcredit and microchips to child sponsorship. Like sponsorship charities, it is all about stories: it was inspired by them and it succeeds by telling them. As a result, it operates in a pincers between the giver's desire for personal connection and the costs and constraints that imposes on business of serving poor people. In fact Kiva can be seen as an ingenious finessing of this old tension. Technology has brought down the cost of transmitting stories and images.

Indeed, Kiva's P2P connections are more solid than those of child sponsorship 15 years ago. The people in the pictures, we can assume, really do get microcredit. Following in the Tribune's footsteps, Nicholas Kristof tracked down one of his borrowers, a Kabul baker, with little difficulty.

On the other hand, the P2P connection comes at a cost, is one-way, and is partly synthetic. The baker was surprised by the encounter because he had never heard of Kristof. For his part, Kristof might be surprised to learn that most of the Kiva loans he helped fund were disbursed before he saw them on Kiva. And the cost of collecting the baker's story, translating it into English, taking his picture,and uploading it over a balky Internet connection may still be significant relative to the small loans and the great needs in Afghanistan.

Is it so terrible that Kiva modestly misleads in order to raise money for a cause about which it is passionate? No. But as a critic I offer these points: