Some thoughts on Kiva's interest rates

The following is based on an excerpt from my much longer and more complete essay “Microfinance and the re-exploitation of women”. The full-length version includes citations for all references and data.

1. Two Misconceptions About Kiva

Kiva Microfunds is a nonprofit organization whose website allows users to provide money toward filling small personal and business loans to individual borrowers around the world. The stated mission of Kiva, whose name is a Swahili word meaning unity or agreement, is “to connect people through lending to alleviate poverty.”

Users of the website browse profiles of borrowers and select those to whom they wish to lend. Each profile includes the borrower’s location, some brief biographical information (usually including a photograph), and a description of what the loan will be used for. Users can then lend as little as $25 to a borrower, which is pooled with money from other users to reach the full loan amount. When the loan is repaid, the user’s account is credited back the amount they gave which can then either be lent to another borrower, donated to Kiva itself, or withdrawn. In its first ten years of existence (October 2005—​October 2015), Kiva’s users lent $774 million to 1.8 million borrowers (75% of whom were women) in over 80 countries.

The goal is to improve the quality of life in developing and conflict-torn regions where it is hoped that entrepreneurs can make effective use of even very small, expensive loans. The loans are disbursed by Kiva’s field partners called microfinance institutions (MFIs). Despite its founders' original intention of allowing users to realize gainful returns on their loans, Kiva itself does not collect interest on loans, and Kiva’s users do not receive any interest on repaid loans.

While the concept behind Kiva is simple — use a website to crowdsource cheap credit to subsidize MFIs operating in poor neighborhoods — the way it is presented can be confusing and tends to result in two misconceptions:

That Kiva users lend money directly to individual borrowers

Because the Kiva website is built around the biography-oriented borrower profiles, it is easy for users to assume that the money goes directly to the borrower after the loan is filled. Both assumptions are wrong for most loans. In fact the money goes to intermediary MFIs who usually disburse loans to borrowers even before their profiles are posted to kiva.org. By the time Kiva users have filled the loan amount, MFIs have not only already disbursed the loan but have likely started collecting payments (and interest) from borrowers.

So Kiva is not a direct person-to-person lending system. Kiva users are in fact lending risk- and interest-free money to Kiva’s partner financial institutions, not to individual borrowers.

The Kiva website was once even more misleading about how disbursements worked, but after David Roodman of the Center for Global Development published an editorial titled “Kiva Is Not Quite What It Seems” (October 2009) which gained significant attention online, Kiva updated their “How Kiva Works” page to make it clear that loans are pre-disbursed.

And that borrowers are not charged interest on loans received through Kiva

Because Kiva itself is nonprofit and presents its operations as philanthropic, and because users do not receive interest on the money they lend, users might easily assume that borrowers are not charged interest. That is also an incorrect assumption. Most of Kiva’s partner MFIs do collect interest and other fees on the loans, sometimes at very high rates, and sometimes acting as explicitly for-profit investor-owned banks.

2. What Are the Interest Rates Charged by Kiva’s Partners?

Kiva does not directly provide information on interest rates for most individual loans. Instead, it lists two measures of field partners' performance which can be used to roughly indicate the average annual interest rate on that institution’s products and how much profit those rates produce:

  • Portfolio Yield is the primary measure of the cost of loans from a given lender and can be treated roughly as the annual interest rate. Portfolio yield is defined as all interest and fees paid by borrowers to the field partner divided by the average portfolio outstanding during any given year.

  • Profitability (Return on Assets) is the field partner’s net income divided by its total assets. It indicates how efficiently a field partner turns its investments into profit.

Kiva has started listing a calculated APR for some of its partners. They’ve introduced an “Average Cost to Borrower (PY/APR)” metric which displays either the Portfolio Yield (PY) or Average Percentage Rate (APR), whichever is available.

According to the Kiva website the average Portfolio Yield for all of its field partners in January 2016 was 29.09% (down from 35.21% in January 2010), but some of its field partners have yields near 100%. In 2010 the charity evaluator GiveWell surveyed one of Kiva’s field partners, MLF-Malawi, and reported that the actual APR on its most popular loan was 144%-149%.

As of January 2018 Kiva provides both Portfolio Yield and Profitability for 196 out of its 270 active field partners. The table and graph below summarize the range and distribution of those metrics.

Minimum 1st Quartile Median Mean 3rd Quartile Maximum

Portfolio Yield

0%

0.975%

22%

23.839%

37%

101.6%

Profitability

-170.2%

-1.653%

1.5%

-0.262%

4.333%

124.76%

Box plot of field partner Portfolio Yield and Profitability (data is also given in table)
Figure 1. Distribution of Portfolio Yields and Profitability among Kiva’s field partners (whiskers extend to 1.5 IQR). Click for PDF version.

As a comparison to interest rates in countries with established and widely available financial services, the average credit card APR in the United States is about 15%, and the maximum allowed interest rate on loans backed by the Small Business Administration is 8.75%.

2.1. The problems with ‘Portfolio Yield’

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. 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.

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:

  • Many of the microfinance institutions partnered with Kiva require borrowers to save a portion (usually around 20%) of the loan they receive. This requirement is known in the microfinance world as “forced savings.” Borrowers from MFIs which use forced savings are effectively paying interest on a larger loan than they receive, so the Portfolio Yield in those cases will be misleadingly low.

  • Because the Portfolio Yield is calculated based on an institution’s average outstanding portfolio, it will tend to reflect the larger (usually cheaper) products offered by that institution rather than being an accurate estimate of the (almost always higher) rates on the tiny loans provided to the struggling rural borrowers featured on Kiva.org. The MIX report identified this (the use of the wrong APR to represent an institution’s microfinance loans) as the main reason Portfolio Yield and APR diverge.

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:

Kivans like to believe they are helping the poor, and in order to achieve this Kiva needs to provide them with minimal, but reassuring information. Some nice photos, a little story, and as favourable an impression of the actual interest rates as possible, as this is an emotive topic that will irritate many Kivans. They can get away with rates of 30%, 40%, even 50%, but they have to avoid rates which will raise too many questions, and by citing a statistic known to be deeply flawed, but reassuring the Kivans, is the best way to do this.

2.2. The use and difficulties of ‘Profitability (Return on Assets)’

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).

Scatter plot of Profitability against Portfolio Yield
Figure 2. Profitability plotted against Portfolio Yield. Click for PDF version.

The lack of correlation between Portfolio Yield and Profitability shows that at least some markets are competitive, though likely very geographically uneven.

Furthermore, 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%.

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, some preliminary time-series investigation reveals that the Profitability indicator seems to be rather noisy and unreliable as an indicator that an MFI is overcharging (or undercharging) for its products. (This is something I have not explored further.)

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.

3. Finding good lenders on Kiva

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).

The ten field partners with the highest Portfolio Yields. Visit KivaSort.americancynic.net for a full, up-to-date table.
Name Portfolio Yield Profitability Country Default Rate

3.1. Example: Thrive Microfinance

If we read the Kiva profile for the MFI second from the top of the list of field partners with the highest Portfolio Yields, we learn that Thrive Microfinance is an independent MFI in Zimbabwe which lends exclusively to women using the group loan approach (in which small groups take out a loan collectively and keep each other accountable for payments).

Note that after suffering record hyperinflation from 2007—​2009 (by the time the central bank stopped issuing currency, prices were more than doubling every 25 hours), Zimbabwe stabilized prices by abandoning its national currency and switched to the US dollar. During most of 2014, Zimbabwe was experiencing slight deflation caused by a shortage of cash. That means the credit provided by Kiva was all the more valuable to Thrive, and its loans that much more expensive than their nominal rate.

The main reason Thrive charges so much for loans is apparently because they provide four weeks of mandatory training to borrowers before disbursing a loan. There is a note from Thrive addressing the high interest rates on their Kiva profile which concludes, “Even though we could reduce the interest rate if we reduced the amount of training, we do not believe that it is in our borrowers' interests to do so.” That rings hollow to me: why not provide cheaper loans and then sell training to the groups who find that service valuable enough to pay for it?

4. Philanthropy or Business?

It is immoral to use private property in order to alleviate the horrible evils that result from the institution of private property. It is both immoral and unfair.

4.1. Kiva’s origins: an accidental charity

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.”

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” (36). 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:

I architected the database, software, and user experience around the idea of returning interest to users. There was never any question that we wanted interest rates on the site. (37)

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” (53-54).

4.2. Charity for whom?

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?

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.

Some MFIs have shown themselves to be nothing more than predatory banks, as typified by the 2007 IPO of Compartamos Banco in Mexico which raised millions of dollars of equity for investors with its business model based on charging groups of women very high interest on microloans.

In October 2010 the microfinance industry in the Indian state of Andhra Pradesh self-destructed in a frenzied lending bubble accompanied by aggressive collection practices. Hundreds of suicides in the region have been linked to microfinance debt and harassment at the hands of loan agents.

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.”

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 2009 Flannery wrote a second retrospective for Innovations in which he recounted some of the hardships the startup organization faced in its first four years of operations. In that time 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.

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. 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.

4.3. Kiva’s apology

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.

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” (54).

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.

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.

Even four years into Kiva, Flannery 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.”

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. No doubt this new enterprise provides Flannery with better opportunity than Kiva to more directly “empower” cash-strapped sub-Saharan workers.

5. Direct and interest-free: Kiva’s future?

In the final analysis, putting questions of profit and exploitation aside by assuming that MFIs operate efficiently and at some optimal balance between charity and sustainability, the root question which confronts the Kiva user is whether giving poor women in Zimbabwe expensive debt is a good way to help them — and if it is not, then what else can someone on their computer in a rich Western country do to help?

Fortunately Kiva has a few initiatives in the works which may help side-step these difficult questions in the future. In 2013 they launched Kiva Zip, a pilot program which offered loans directly to entrepreneurs (without an intermediary MFI) in the United States at 0% interest and without any credit score requirement. In other words, Kiva Zip set out to be exactly what many people thought (and continue to think) Kiva was to begin with (albeit limited to the USA). In 2016 the Kiva Zip program was integrated into the main Kiva website as Kiva U.S.

In the summer of 2016 Kiva also announced the Direct to Social Enterprise program which provides interest-free loans directly to medium-sized enterprises (too big to be customers of MFIs and too small to benefit from commercial loans), which has brought the benefits of Kiva Zip to countries outside of the United States (albeit on a person-to-enterprise rather than person-to-person model). I haven’t found a complete list of participating social enterprises, but at least a partial list can be found by searching KivaSort for field partners whose names contain ‘direct to’.

6. Deeper questions

On one hand the demographic and geographic distributions of Kiva’s borrowers and lender-users, wherein sympathetic people in parts of the world with extra money are providing charitable loans to people in parts of the world with not enough money, are not surprising. Those are exactly the sort of relationships Kiva exists to facilitate per its mission statement, after all. But Kiva’s entire model of microfinance takes for granted that there are a great number of women and peasants at the developed world’s periphery who are in desperate need of financial services without attempting to explain why the world’s wealth has become so stratified by lines of geography and gender, and without any introspection into its own role in the greater processes of global capitalism arising from and transforming those lines.

For a closer look at those processes see the full essay at “Microfinance and the re-exploitation of women”.

Comments

comments powered by Disqus