Digital Credit: Data Sharing Can Improve Product Diversity

How innovative is digital credit? While the data analytics, mobile technology and delivery channels that make digital credit possible are innovative, most digital credit products are quite basic — short-term, high-cost unsecured debt to individual borrowers. Even five years after M-Shwari launched, much of the market feel like clones of this first pilot, with the main difference being the interest rate charged.

Beatrice Kimwele conducts business on her phone in Kenya.
Beatrice Kimwele conducts business on her phone in Kenya. Photo by Mwangi Kirubi, 2015 CGAP Photo Contest.

Many lenders do not have sufficiently deep or comprehensive data to offer customized products that go beyond small-value loans for personal consumption (e.g., business loans to microentrepreneurs based on wholesale purchase or merchant payments activity data). Lenders may have access to very different sources of data and customer information, such as mobile phone records, social media or financial statements. Their data are either not shared across lenders or, if shared, are not reliable. Even within formal data-sharing channels like credit bureaus, data are often incomplete, outdated or even inaccurate. This limits digital credit providers’ ability to segment borrowers and offer more products that are tailored to them.

Data around digital credit can be especially unreliable. A survey in Kenya showed digital credit users their credit histories via SMS and asked them whether the reports were accurate. Ninety-two of the 420 consumers surveyed (22 percent) indicated that portions of their credit histories were either inaccurate or missing, and 86 percent of these discrepancies concerned digital credit products.

Credit history inaccuracies reported by customers
Credit history inaccuracies reported by customers
Sources of inaccuracies reported by consumers
Sources of inaccuracies reported by consumers

The fact that only about 9 percent of these consumers had ever accessed their credit history before the survey suggests that Kenya has a challenge not only with data accuracy, but with consumer access to and control of their data. This is true not just for credit bureau data, but for all the data consumers generate that could lead to better segmentation and customization. Going forward, there should be a more open and consumer-led data-sharing environment for digital credit and other financial services, such as data portability standards that allow consumers to easily access and share their financial histories across providers. Such a system would not only enhance consumers’ ability to share — or restrict sharing of — their data, but empower them to ensure their data are accurate.

This would also be a smart business strategy for digital credit market development. A recent CGAP pilot with M-Kopa in Kenya shows why. M-Kopa offers digital loans and voluntarily reports borrowers’ histories (which are almost always positive) to the credit bureaus in Kenya. CGAP and M-Kopa tested an interactive SMS system that allowed M-Kopa customers to access their credit histories for free and notify the credit bureau, TransUnion, if there were inaccuracies. The SMS content explained to customers that credit history is linked to positive repayment and helps people access future loans.

Initial evidence from the six-week (February to March 2017), prepilot showed strong engagement: 384 of the 1,632 (24 percent) invited customers opted to receive learning content, 225 requested their credit histories, and 53 requested follow-up from TransUnion regarding the data shared in their credit history.

More important, M-Kopa customers who opted into the SMS content took up more credit, were less likely to be blocked by failing to make a payment within 90 days and were more likely to have paid off in full.

Payment behavior during credit history SMS pilot
Payment behavior during credit history SMS pilot

If the M-Kopa pilot showed the power of giving consumers access to and control of their digital credit data, then why is so much digital credit data either inaccurate or hidden from anyone but the providers who collect it? For digital credit products to innovate beyond short-term, high-cost unsecured lending, consumers need to be able to more easily let lenders know who they are and what products may best fit their needs and ability to repay. The following five immediate solutions can make this possible:

  1. Expand credit bureaus' coverage across all provider types in the digital credit landscape.
  2. Increase the frequency with which digital loan data are reported to credit bureaus; when it comes to seven-day loans, reporting every 30 days may not be often enough to help other lenders assess borrower risk.
  3. Mandate that consumers be allowed to access and share with other financial services providers all the financial history they have accumulated — not just loan history.
  4. Make it easier for consumers to view their credit histories and other key data so that they can dispute inaccuracies or missing information.
  5. Sanction providers who intentionally refuse to report borrowers’ data, as has already happened with the largest digital lender in Kenya.

Digital credit began as a relatively siloed ecosystem with limited product diversity, but it does not need to stay that way. A more open, consumer-friendly data environment should benefit consumer welfare, competition and providers’ risk management strategies.



CGAP’s experiments with a diverse range of financial service providers have found that innovative approaches to digital credit can improve the products and improve consumer protection, which often is a wise business decision.


27 August 2017 Submitted by Barbara Magnoni (not verified)

I would be interested in your opinion about whether these markets are "ready" to share credit bureau data. What I mean is that if repayment is low, many people may be shut out of more useful loan size borrowing just because they neglect to pay a small digital loan (perhaps because of forgetfulness, time, etc.), which would gradually lead to larger groups of excluded borrowers rather than included. It would be important to understand the repayment trends that you are seeing to complement the recommendation.

28 August 2017 Submitted by Kris Carter (not verified)

Hi Barbara, I assume your question was for Rafe, but I wanted to respond to it. I can see how poor repayment of small digital loans could certainly harm some people's credit rating, and might even impact their ability to get larger loans. But is that really a bad thing? Sure it may be bad for that individual, but I would always advocate for more transparency rather than less. In the specific case you mention, the fact that the loan is small might mitigate some of the damage to their credit. If they don't pay back a 100 shilling loan that is much less damaging than not paying back a 10,000 shilling loan. Also, the simple fact of credit reporting should improve repayment rates, which is what everyone wants. I do see the issue you raise, and I recognize that there could be some short term negative results for those who fail to repay loans, but the simple fact is, their failure to repay a small loan probably should be indicative of whether they are prepared for a larger loan, which once again argues for more transparency rather than less. There are many challenges to getting good credit reporting data, but I'm not sure I would argue that the harm to borrowers who default should be a primary concern. Besides, as we have all experienced, less transparency almost always ends up hurting rather than helping the most vulnerable, so for that reason I would say we should strongly advocate for quality credit reporting.

04 September 2017 Submitted by Rafe Mazer (not verified)

Barbara and Kris,

Very interesting point and debate. My belief is that more problems arise from hidden or inqueal treatment of credit data than from making a judgement on what data is "too small," "too new" or other criteria to restrict information. The issues and concerns Barbara raise are real and important, but I would address them instead by looking at the following possible root causes:
1. Inconsistent rules for sharing data across digital lenders. Not all must report the same data, leading to incomplete pictures where a small negative listing for one bank coudl show up, but not a positive larger repayment history for a competing FinTech.
2. Binary approaches to credit scoring by lenders. Many lenders don't use a score from the bureau and instead just look at whether there is an entry in arrears or default or not, and then exclude all who fit the former criteria, so they treat a $1 defaulter same as a $100 defaulter.
3. Consumers cannot check and correct their data easily, which is why we piloted this approach with TransUnion, CIS Kenya and M-Kopa.
I would say that the solution here is requiring more providers to report full file data, and have this open the doors for those with more sophisticated scoring models to figure out a way to lend to those who have tiny value loan defaults in their history. Hopefully if the small fish get the same data as the big fish, and start taking market share from them, even the bigger lenders who use the binary scoring approach will adapt to not lose out on possible borrowers.

01 September 2017 Submitted by John McKechnie (not verified)

Rural communications, whether in Sub Saharan Africa, South Asia, Europe or USA are under development and will facilitate not only financial services but also small loans for the poor. We cannot seem to get through one set of interests to reach another but we know we can deliver not only small loans but also distance education, third party commerce, healthcare and directly to globalisation.

Add new comment