It has been about a year since we transitioned control of interest rates from borrowers to lenders, and we now have enough data to assess how the interest rates lenders choose affect the financial performance of their loan funds. Over the past few weeks, we have been working with data scientist Tracey Li to analyze the relationship between interest rates, repayment rates and financial sustainability for lenders. Given the assumption that higher interest rates increase financial returns on Zidisha loans, we wanted to identify more precisely the interest rate that, averaged over many loans, would be most likely to help lenders break even, without excessive profit or loss. The analysis yielded a surprising finding: higher interest rates were not, in fact, associated with higher financial returns for lenders. The reason is that the incidence of non-payment increased when higher interest was charged, and even at the higher rates, the interest income from the repaid loans was not sufficient to compensate for the greater amount of principal lost to non-payment. In fact, the value of lending funds after interest payments was preserved best at 0% interest, and losses were highest for the loans that had the highest interest rates. This finding holds true for Zidisha's five years of history in aggregate, as well as over the past year when interest rates have been set by lenders without input from the borrowers. It holds true for both first-time borrowers, and repeat borrowers: all else being equal, individuals were more likely to repay loans that carried lower interest rates, and the difference in repayment rates was substantial enough to override the difference in interest income. Both the the data scientist and I performed multiple checks of the data, using different methods, in order to be certain of the conclusion: For loans at Zidisha on average, applying higher interest rates has not increased the value of lending funds, and may be counter-productive. The main purpose of interest at Zidisha has been to help lenders preserve the value of their lending funds. Since a substantial body of data now show that interest has not fulfilled its intended purpose, it is our responsibility to share these results, and to adjust our service in response. Starting this month, we will be piloting an alternative approach to loan fund sustainability: Instead of paying interest over the course of each loan, new borrowers will be asked to pay a larger membership fee at the time they raise their first loan. The fee ranges from about $12 to $40 depending on the country. (It is highest for the fast-growth countries of Kenya and Ghana, and temporarily reduced or waived in other countries where membership numbers are still small, in order to facilitate early growth in those locations.) It is only paid once (for lifetime membership), and only in the event that a loan is successfully funded and disbursed. In the event one of the zero-interest loans defaults, we will use funds from the membership fee pool to reimburse lenders, up to a maximum of $1,000 per lender, as long as funds are available. Since this is subject to funds availability and does not cover currency exchange rate losses, it is not an insurance policy and does not eliminate risk. Zidisha is a philanthropic platform, not a place to store financial assets, and as always, we recommend lending no more than you would be prepared to give away to charity. The intent is simply to do our best to help lenders preserve the value of their lending funds over time, not to make lending accounts a risk-free place to keep one's money. How will the reimbursements work? Here is the way they are currently programmed in our website: When a new borrower joins, a Loan Loss Reserve fee payment will be due after his or her first loan is fully funded. The payment must be made before the first loan is disbursed. This fee is only paid once, at the time a new borrower joins; it is not paid for each loan. Once received, the payment goes into a Loan Loss Reserve account. The funds in this account will be used exclusively to reimburse lenders for loan losses. In the event that one of the zero-interest loans falls into arrears (defined as more than ten days overdue, by an amount greater than $5 or greater than the value of one repayment installment, whichever is lesser) a "Receive early repayment" button will appear in the "Your Loans" page if the following two conditions are true: There are sufficient funds remaining in the Loan Loss Reserve account. You have not received more than $1,000 in Loan Loss Reserve payments in the past. If you click the "Receive early repayment" button, you will receive a refund for the full dollar amount you contributed to the loan, minus any repayments you have already received. The intent is that the amounts refunded to lenders are as close to possible to the amounts paid by borrowers. In order to achieve this balance, we may adjust parameters such as the amount paid by borrowers, the lateness threshold at which refunds will be available, or whether lenders are refunded automatically instead of opting in. However, the overall structure will remain the same for these loans. We believe shifting borrower costs to the initial membership fee will have several advantages: It increases the financial incentive to repay, once a borrower has joined Zidisha. It makes joining Zidisha less attractive to anyone who does not intend to repay, because most of the financial benefit is deferred to later loans. Taking interest out of the picture may improve the relationship between lenders and borrowers. It may make participation more attractive for lenders, who commonly report choosing an interest rate to be a less enjoyable aspect of funding a project. It makes cost of loans more predictable for borrowers. That said, we will monitor performance carefully during the pilot. If the expected benefits do not materialize, or there are unexpected problems, we will reevaluate. In the event we decide not to continue the pilot, the offer to reimburse lenders for defaults of the zero-interest loans made during the pilot will still be valid. Notes: You may read an overview of the data science project here: https://tcl270.github.io/ Below is an illustrative subset of the data, depicting the changes in loan fund values for Kenya loan amounts that were due to be repaid six months ago or more, for the loans that were disbursed in the past year. The Change in Loan Fund Value is calculated as the difference between the principal that was due to be repaid during the period, and principal + interest that has been repaid to lenders.