The Portfolio Yield indicator is the best available proxy for the actual annual percentage rate (APR) for most loans disbursed by Kiva’s field partners. Portfolio Yield was actually introduced by Kiva in 2009 to replace an even worse proxy for interest rates.^[6] Microfinance Information Exchange (MIX), a non-profit organization, sampled several MFIs and found an adjusted Portfolio Yield to be within five percent (not percentage points) of the reported APR.^[7]

However, in the MIX analysis the Portfolio Yield indicators were adjusted upward to account for defaulted and at-risk loans still on the books (which inflate the portfolio value and so deflate the reported Portfolio Yield). In addition to overvalued portfolio, the Portfolio Yield can easily underestimate the actual APR of microloans for two other reasons:

So the nominal interest rates suggested by the Portfolio Yield of an MFI, which can be outrageously high, in many cases significantly underestimate the interest actually charged to microfinance borrowers.

Hugh Sinclair, author of Confessions of a Microfinance Heretic, wrote an article titled “What’s Wrong With Kiva’s Portfolio Yield Statistic?” exploring both of these (and other) problems with Kiva’s Portfolio Yield metric. He compared the actual APR for a sample of loans from ten of Kiva’s field partners with the Portfolio Yield reported by Kiva: for every MFI he looked at, the APR on most of the loans was higher (sometimes as much as double) the reported average Portfolio Yield.

So why doesn’t Kiva require MFIs to state the actual cost of loans made using Kiva money rather than relying on the oblique “Portfolio Yield” figure? Sinclair thinks it is because if people saw the actual interest rates collected on loans they are subsidizing then they would think twice before using Kiva:

While the high rates charged by some of Kiva’s partners are a bit shocking and look outright usurious on first sight, Portfolio Yield alone doesn’t indicate whether the interest being collected is exploitative. Those high rates could reflect the actual costs of administering credit in some regions of the world. Small loans are more expensive because fixed costs tend to dominate the price. And some regions have high inflation, poor or non-existent infrastructure, difficult to reach populations, or high crime and other instabilities, all of which contribute to the cost of credit.

If high interest rates were simply the result of price gouging due to lack of competition, then we’d expect that MFIs which charge high rates would tend to have correspondingly high profitability. However, as illustrated in the graph below of Profitability plotted against the Portfolio Yield for all 149 MFIs (as of January 2018) with nonzero Portfolio Yield and available Profitability data, there is no such correlation (the Pearson correlation coefficient is 0.09).^[3]

A 2013 report based on a sample of 193 MFIs (none necessarily partnered with Kiva) gave similar results but found a slightly negative correlation between return on equity and portfolio yield (r = -.117). The author of the report hypothesized that the negative correlation, indicating that more profitable MFIs charge lower interest rates, is the effect of MFI lifecycle stages: young MFIs try to reach financial stability by charging high rates and fees while mature MFIs can afford more competitive rates.^[9]

In any case, the lack of correlation between Portfolio Yield and Profitability shows that at least some markets are competitive, though likely very geographically uneven. We might hope, then, that even if we concede high interest rates are a necessary evil in some regions to cover operating costs, we could use the Profitability metric alone to identify MFIs that are overcharging for loans (say, try to avoid giving to MFIs who make more than 15% on their assets). Unfortunately, as is noted in the next section, the Profitability indicator seems to be rather noisy and unreliable as an indicator that an MFI is overcharging (or undercharging) for its products.

In addition, it should be kept in mind that low profitability can be an indicator that an MFI is struggling with internal inefficiency including under-utilized investments, over-paid executives, and fraud (none of which are especially unheard of in the microfinance sector).

Other metrics Kiva provides that can help here are the Default Rate, Delinquency Rate, and Loans At Risk Rate. Field partners which have trouble collecting repayments may be charging more than their clients can afford.

In addition to the information provided by Kiva itself, financial data about many of Kiva’s partner MFIs can be found on MIX Market (themix.org), an online clearinghouse for microfinance information run by the World Banks' Consultative Group to Assist the Poor. The data available on MIX Market is largely self-reported but claims to be independently reviewed and includes many more indicators than Kiva provides (including a portfolio yield figure which has been adjusted for inflation). Unfortunately, in June, 2016, while I was doing research for this essay, the MIX Market website was restructured to commercialize most of its data and publications behind a paywall making future research efforts using MIX datasets prohibitively expensive. Prior to the redesign, the MIX Market profile for Kiva provided a convenient list of MFIs currently and formerly associated with Kiva.^[10]

In a weblog entry mourning the change, Phil Mader notes that the privatization of the MIX Market data “mirrors one of the darker trends in microfinance as a whole, where institutions are first set up with public or charitable money and supported for years (MIX was funded with millions of dollars in charitable, tax-deductible donations), but then are turned onto a revenue-maximising, commercial course, confronting their users with a hard-nosed commercial lender. Even though in practice this restructuring often fails to yield truly commercial returns (and behind the scenes the institution continues to be supported with soft money) the beneficiaries still must deal with what poses as a for-profit business, stripped of the more ‘social’ promises that lured them in.”^[11]

One issue in trying to assess an MFI’s financial characteristics is that the data reported by Kiva is sometimes very different than the data provided by MIX Market. To get an idea of how well the available data agrees, I sampled^[12] the Portfolio Yield and Return on Assets indicators from both Kiva and MIX Market for several MFIs and plotted them in the figure below.^[3]

While both sources tend to report similar Portfolio Yield figures (left graph), Kiva’s numbers can occasionally vary greatly from MIX Market’s. When the two obvious outliers are ignored, the correlation coefficient is nearly linear (r=0.96). In fact Kiva actually uses MIX Market data directly when reporting Portfolio Yield for some of its field partners, so that strong correlation is expected.^[13] I do not know if there is a list of which partners Kiva uses MIX Market data for and which it calculates itself.

The Return on Assets figure (right graph), on the other hand, is much less consistent between Kiva and MIX (r=0.52).

I suspect the main cause of the discrepancies between the data reported by Kiva and MIX Market is timing: each organization receives and releases information from the MFIs on different dates (the plot above is based on the most recent data from both Kiva and MIX at the time of retrieval — but they aren’t necessarily updated at the same time). If that’s the case, then it indicates that the reported Return on Assets of MFIs tend to be rather volatile, which is another reason the Profitability metric may not be of much use in identifying non/exploitative MFIs.

Keeping the limitations of Kiva’s metrics in mind, the best strategy a Kiva user can adopt in order to find loans from a good MFI on Kiva.org depends on their humanitarian philosophy: a user who is concerned that high interest does more harm than good should seek out loans through partners with low Portfolio Yield; a user who is most concerned about exploitation should look for partners with low Profitability; a user who wants a balance of sustainability and humanitarian efficacy might look for partners with low Portfolio Yield and moderate Profitability; etc.

The best way to explore individual loan offerings is to use Kiva’s lending tool (https://www.kiva.org/lend). As part of a June 2016 redesign, Kiva implemented improved filtering of their lending tool including the ability to filter based on “Average cost to borrower” (Portfolio yield, APR, or MPR, whichever is available) and “Profitability.”

Unfortunately Kiva does not provide the option to sort their field partners table by Portfolio Yield or Profitability. But they do helpfully provide field partner information through a programmable web interface. I used that service to build KivaSort (https://kivasort.americancynic.net/) which provides a fully sortable and filterable table of Kiva’s field partners.

By default KivaSort displays field partners with the lowest Portfolio Yield at the top to facilitate finding inexpensive lenders. But in order to illustrate and briefly investigate the high interest rates charged on some Kiva loans, the table below lists the active field partners with the highest Portfolio Yield (as of January 2018).

Kiva was founded in 2005 by Matt Flannery and Jessica Jackley, a husband-and-wife team working out of San Francisco. They developed the first version of the Kiva.org website while Flannery was employed as a programmer at TiVo and Jackley was employed at the Stanford Business School (where she attended a lecture by Muhammad Yunus which ignited the initial spark of inspiration). Two years after launching, Flannery published a retrospective about Kiva’s origins and development called “Kiva and the Birth of Person-to-Person Microfinance.”^[16]

In that article, Flannery described how even in the early days of Kiva there was a fundamental tension about “whether it was better to be seen as a charity or as a business.”^[17] Neither Kiva nor its users have ever earned interest on the loans facilitated by the website, but that charitable nature is more an accident of history (encouraged by the bureaucratic hurdles erected by the SEC) than the design of its founders.

The original plan was to charge MFIs interest on Kiva-financed loans, and then share a portion of that interest with users:

After grudgingly settling on the interest-free approach for the first two years, Flannery wrote that “we would still like to realize our original vision of having interest rates on the site. The fact that we had to remove them is a sore spot with me […​] Kiva thus continues its effort to allow our partners to post businesses to the site with interest rates attached.”^[19]

Kiva’s founding tension was resolved by keeping Kiva itself purely charitable, supported exclusively by donations and grants, but partnering it with remote field partners who charge interest. That way the SEC was satisfied and the dirty business of collecting interest from people with no money was pushed to a more comfortable distance from the users of Kiva’s website.

But the impedance-matching function of Kiva, which converts interest-free loans made by Kiva’s users into interest-bearing loans collected by microfinance institutions, understandably produces a cognitive dissonance in its users. That double nature of loans given through Kiva, which are simultaneously charitable credit for MFIs and expensive debt for poor borrowers, is the source of the misconceptions I outlined in the first section of this essay, and it raises a question at the heart of the matter: who benefits from the free credit raised by Kiva? Does the incidence of Kiva users' charity fall mostly on the borrowers they intend to help? or does it fall more on the MFIs who accept the free credit and then turn around and loan it for gain to those borrowers?

Profitability among Kiva’s field partners tends to be rather modest (suggesting that Kiva prefers partners on the charitable/non-profit as opposed to self-sufficient/commercial side of the spectrum): over three-quarters of currently active field partners have a Profitability at or less than 4.3%. By comparison, one report that looked at interest rate data from hundreds of MFIs (not limited to, or necessarily including any, Kiva field partners) over seven years found that three-quarters of MFIs in 2011 had a rate of profit up to 20%. That same report noted that if every MFI set their interest rates to their break-even point (so that profits = 0), then the average interest rate would fall by only 2.6 percentage points (in other words, even if every MFI were non-profit, interest rates would still be quite high in many cases).^[20] That finding underscores the fact that microcredit is simply expensive.

The efficacy of microfinance at alleviating poverty has been a matter of research and debate since Muhammad Yunus founded the Grameen Bank in Bangladesh in 1983 (Yunus and the Grameen bank were jointly awarded the Nobel Peace Prize in 2006). The early anecdotal reports of success and the prospect of a business-friendly cure to poverty created an increasing excitement around microfinance for over two decades. But in recent years expectations have sobered.

In the past five years or so several rigorous studies which use a randomized method to compare the effects of microfinance on borrowers have appeared in the academic literature. The result of one recent survey of six such studies found that “The studies do not find clear evidence, or even much in the way of suggestive evidence, of reductions in poverty or substantial improvements in living standards. Nor is there robust evidence of improvements in social indicators.” But the same survey also found “little evidence of harmful effects, even with individual lending […​] and even at a high real interest rate.”^[21]

Kiva performs a degree of due diligence and monitoring of the MFIs it chooses to partner with which provides some protection against abuse. The measures Kiva uses to evaluate field partners developed out of some hard-learned lessons. In their first four years of operations they discovered six “situations involving severe fraud,” including one involving their very first partner, a Ugandan man named Moses Onyango, who was so instrumental in getting Kiva started that he is sometimes referred to as the third co-founder.^[22]

Among the MFIs Onyango signed up as partners in Kiva’s early days was one he founded himself, the Women’s Initiative to Eradicate Poverty (WITEP). It turns out that not all of the WITEP borrowers were real: some loans were being disbursed in the names of fictional people and pocketed by Onyango. Due to Uganda’s unresponsive legal system, Kiva never recovered the stolen money (but they did pay back Kiva.org users out of their own expense account, as well as maintain a policy of transparency about the fraud when it was discovered).

The kind of fraud Onyango perpetuated is not particularly worrying. He used the stolen money to buy a house for his family (and he was so grateful to Kiva for its influence on his life that he named his new son Matthew Flannery Onyango). What Onyango did was to cut out the interest-charging middleman and transform Kiva into the version of itself that users imagine it is: they lent money, helped out a Ugandan family, and got repaid.

Far more worrisome are legitimate MFIs who might use Kiva’s website as a place to sell feel-good stories to naive Americans for free capital with which they can go about their business of robbing the poor. It strikes me as much less likely that Kiva would discover and terminate its partnership with MFIs who overcharge borrowers, and in fact Kiva’s entire structure of funneling interest-free credit to interest-charging lenders almost encourages it.

As an example consider the case of Kiva’s former partner in Nigeria Lift Above Poverty Organization (LAPO). In 2010 the New York Times published an exposé about microfinance interest rates which specifically mentioned the high rates and forced savings of LAPO.^[23] Kiva had initially defended the interest rates and high profitability of LAPO with its usual explanations (inflation, high operating costs, and the need for sustainability), but then after further investigation prompted by the media attention decided to terminate the partnership. LAPO overwhelmingly targeted women with its high-interest loans. The LAPO incident doesn’t exactly breed confidence in Kiva’s other partners who have not been investigated by journalists.^[24]

To questions about the high interest rates charged by its field partners, Kiva has responded by pointing to the high costs of administering microloans, emphasizing that they annually evaluate each partner “to ensure that there is a sound justification for each relationship,” and that they are continuing their efforts to bring in more charitable and 0% lending partners.^[25]

In his article, Flannery also outlined a justification for the proposed collection of interest on loans to the global poor. He rejects pure charity, what he calls the benefactor relationship between people in developed and undeveloped countries, because “recipients resent benefactors even as they consume the aid.” He also rejects the defeatist notion that poor people cannot be helped, what he calls the colonizer relationship, because it is just the other side of “the same destructive mentality” as the benefactor. To these dialectical endpoints he applies a liberal dose of Silicon Valley Logic and derives the perennial insight that the best way to help poor people is to find a way to make money off of their circumstances. He calls this the business relationship, from which he formulates the precept that “interest rates, which turn a charitable relationship into a business relationship, empower the poor by making them business partners.”^[26]

Invoking the concept of colonizer without any discussion or hint of awareness about actual colonialism or the socioeconomic context within which microfinance works is emblematic of Kiva’s tone-deaf approach to structural issues. This decontextualization is also inherent to the way the Kiva website packages and presents borrower profiles by “displacing them from local contingencies” thus providing to lenders a ‘flat’ and placeless perspective on development and poverty. “By obscuring the geographic contexts of microlending and borrowers, Kiva.org squanders an essential opportunity to engage the public with the ongoing conversation about poverty, debt, development, and the roots of contemporary inequality.” And the faux equality presented by Flannery’s “business relationship” obscures the differentials of wealth between the site’s lenders and borrowers. “Kiva not only renders its own workings unproblematic, but also justifies the worldwide expansion of micro-credit by much larger conventional aid and financial organizations.”^[27]

Of course Flannery’s “business relationship” has never been unambiguously achieved by Kiva. The tension between philanthropy and business remains, and is in fact what is attractive to users of the platform. By promising a way for users to help the global poor without personal gain to themselves and without just giving money away, Kiva harmonizes the warring moralities of compassion and fiscal responsibility.^[28] By blurring the differences between lenders and borrowers to maintain the illusion of person-to-person business deals,

In 2009 Flannery wrote a second retrospective for Innovations in which he expressed his lingering unhappiness that Kiva was not generating profit for comfortable Americans (who make up most of Kiva.org users) off of the hard work of poor women in remote agricultural villages (who make up most of Kiva’s borrowers): “I repeatedly tried to get the interest rates back on the site. […​] To me, taking the rates off the site was an accident and I was determined to undo that temporary concession.”^[30]

Fortunately, Flannery was never successful in his efforts to turn Kiva into a for-profit lending platform which would have had the effect of making loans targeted at the global poor even more expensive while sucking more wealth out of impoverished regions. In 2015, after serving as Kiva’s chief executive officer for ten years, he stepped down (but remains on the board of directors) to co-found the for-profit Branch International. Branch provides a branchless banking service (based on Vodafone’s M-Pesa) which allows people in Kenya to use a mobile phone app to receive small loans. Branch appears to determine credit worthiness and interest rates on an individual basis by an algorithm (which looks at borrowers' Facebook profiles, among other sources of data). By July 2017 Branch is reported to have disbursed $35 million (KSh3.63 billion) in loans to its 350,000 users. Interest rates start at 163.2% per year, and can get as low as 14.4% for users with the best credit rating.^[31] No doubt this new enterprise provides Flannery with better opportunity than Kiva to more directly “empower” cash-strapped sub-Saharan workers.

The data used to generate the table and graphs presented in Section 1, “Two Misconceptions About Kiva” are available to download as:

The data files above can be automatically updated and the figures re-generated using the following scripts: