FOCUS NOTE Consumer Protection in Digital Credit D igitally delivered credit is quickly expanding in emerging markets. “Digital credit” refers to credit products—including digital payments Digital credit is also promising from a financial inclusion perspective, given the low access to formal credit by low-income consumers in most products such as mobile money—that are delivered developing countries and the limitations of informal fully via digital channels, such as mobile phones and semi-formal options. Yet the very attributes of and the internet.1 CGAP research in Africa, Asia, digital credit—instant, automated, and remote— and Latin America counts 22 deployments with create consumer protection risks that are distinct an estimated total of 24 million subscribers, and from those of more traditional consumer and six deployments with a total of more than 1 million microenterprise credit models. These consumer users. Commercial Bank of Africa’s M-Shwari in protection risks include low-income consumers’ Kenya and M-Pawa in Tanzania lead the way with poor understanding of loan costs and the 13.5 million and 4.8 million users, respectively consequences of default, which can be exacerbated (Vidal and Hwang 2017 and GSMA 2017). To date by interface limitations, such as small screens and most loans are low in value (generally $10–50 to short menus; their lack of “intentionality” when start) and very short in tenor (typically 2–4 weeks). making borrowing decisions on the spot;3 and the Interest rates in digital credit commonly range opportunity to easily renew a series of high-cost between 6 percent to 10 percent monthly for a loans. These risks can result in consumers taking one-month loan, which is relatively expensive on expensive loans, borrowing when they do not compared to traditional formal loans in similar have a real need, facing challenges in on-time sectors such as microfinance,2 although possibly repayment, and suffering the consequences of a less expensive than informal moneylenders who negative listing in the credit bureau. Also, they may charge an interest fee equal to the amount may not benefit as much as they could from the borrowed (Ochieng 2016). “digital data trails” and positive credit history they create when the way in which these data are used These business models are driven by strong is not made clear, or when consumers have limited customer demand, lower operating costs, and control over who has access to these data and the greater reach of the instant, automated, and for what purpose. (For a more extensive look at remote lending methodology. Because of these potential behavioral shifts of borrowers in digital factors, they can scale more quickly than traditional environments, see Annex.) small-loan models (Chen and Mazer 2016). The convenience and speed of digital credit are well Risks to lenders are also important because they matched to urgent and unanticipated needs, such may have negative impacts on consumers and on as a late-night emergency visit to the hospital or the development of sustainable and competitive working capital for the quick-turnover, high-margin credit markets. These risks include the poor No. 108 economic activities common for microenterprises. disclosure of prices, terms, and conditions; weak August 2017 For example, in Kenya the leading lender’s loan client communications once the loan is executed; volume surges between 3 and 5 in the morning limited efforts to assess affordability and ensure Rafe Mazer and because that is when small-scale traders purchase suitability of their product offerings for specific Kate McKee their stock for the day (Omondi 2017). consumers or segments; and incentives to engage 1 In this publication, “digital credit” refers to unsecured cash loans in emerging markets that are obtained via digital channels without the involvement of a salesperson, that use digital channels for loan disbursement and collection, and that leverage digital data to make lending decisions via automated processes. 2 A CGAP review of interest rates in microfinance found average annual interest rates near 30 percent (Rosenberg et al. 2013). 3 “Intentionality” refers to the level of prior consideration of need, costs, and benefits consumers take when choosing to borrow as well as the extent to which they have thought of the purpose of the loan they take on. 2 in behavior that limits consumers’ ability to use their and analysis of account-level data of digital borrower history and other data sources to seek borrowers. The demonstrations identified cost- multiple digital credit offers and foster competition effective, win-win practices that can influence among providers. The results for lenders can 4 product design and delivery to better protect include underperforming loan portfolios, subpar borrowers while improving providers’ business profitability, and loss of trust from customers, viability (see Box 1). regulators, and the general public. Well-designed and enforced consumer protection At the same time, there are ways in which new digital rules based on emerging provider good practices credit models can reduce the risks and problems will likely be a necessary complement to the common in conventional lending to low-income efforts of individual lenders or industry initiatives. consumers. For example, the risk of pressure sales Specifically, rules of the road are needed to do the or aggressive collections is lower because people following: can obtain loans without the participation of a loan officer. Consumers could benefit from the ability of • Create a level playing field across the market lenders to better standardize the marketing, sales, through common standards for suitable product and post-loan servicing processes. Furthermore, design and responsible business conduct the lender has far greater opportunity to tailor for all digital lenders, whether regulated or communication to the borrower by using SMS or unregulated. a smartphone app that could include additional • Enable consumers to understand and leverage educational content on the product, tips on good their data and borrowing history, which in turn borrowing habits, reminders on when and how to can help drive consumer choice and provider pay and the importance of timely repayment, and competition. explanation of the consequences of failure to repay in full. Digital credit and consumer protection experiments This paper explores new approaches to address risks and problems in five areas: Since 2014, CGAP has partnered with diverse providers to identify key consumer protection • Disclosure of loan terms and conditions issues raised by digital delivery of small loans • Marketing approaches to promote responsible and to test potential solutions. This exploratory borrowing research was conducted at relatively low cost— • Appropriate and tailored products to meet the approximately $50,000 or less in research costs for needs of specific consumer segments each of the projects presented in Table 1. Findings • Repayment and collections from these digital credit experiments inform this • Credit reporting and information sharing publication. It draws from recent demonstrations by diverse Responsible digital East African lenders on how to identify potential credit throughout the solutions to common problems. The demonstrations customer journey used a range of methods, including lab testing with typical consumers that simulates the live borrowing For digital credit, each stage of the product lifecycle experience, live testing by lenders to modify is clearly demarcated, standardized, and unaffected digital credit products and communications and by financial services provider (FSP) staff behavior or then measure the resulting changes in consumer biases because the process is automated and remote. behavior, qualitative research, consumer surveys, This enables analysis of risks specific to each stage. 4 See, e.g., the recent discussions and concerns raised regarding digital credit and consumer protection from a range of actors and markets: Pande and Memon (2017), Robinson and Wright (2016), AFI (2015), and Owens (2017). 3 Box 1. Top 10 tips for digital lenders to build strong business models and customer relationships Disclosure of loan terms and conditions make sure consumers are making active and well- • Provide consumers the all-in price before they considered credit decisions. sign a loan agreement. Consumer understanding • Structure the loan process to collect—with clear of costs can improve intentionality and repayment data privacy protections in place—more customer performance. data upfront to better assess needs and avoid • Test and adopt measures so borrowers read the observed tendency toward “mono-product,” and understand the terms and conditions one-size-fits-all digital loans. (T&Cs) and their obligations. This includes Repayment and collections using cost-effective tweaks to the menu design, • Optimize effectiveness of payment reminder messages “opt-out” framing, and screens that summarize through framing content, timing, and tailoring to “Key Facts” in a clear and simple manner. different borrower segments and preferences. Marketing • Reward strong repayment performance by using • Consider whether push marketing (addressed incentives such as risk-based pricing, lower lending later in this paper) and unsolicited offers are costs, or longer terms to create incentives for your effective strategies in the long term because they “prime” customers. exacerbate the risk of encouraging borrowing • Consider whether your system allows for flexibility without a purpose. in repayment options, such as semi-automated loan • Design effectively framed loan offers to reduce restructuring. the likelihood that consumers will take the largest Credit reporting and information sharing amount available without thinking through their needs and repayment capacity. • Increase borrower awareness of their data trails and credit histories—including their credit reports— Suitability and product design and their ability to ensure accuracy, which in turn • Introduce measures to improve intentionality and incentivizes strong performance and strengthens increase the “friction” in the borrowing process to loyalty. For example, the lender sends out preprogrammed the same repayment status get the same repayment identical sales and marketing SMS invitations to new reminders and are treated the same for collections borrowers; standard disclosure screens appear at purposes. In contrast, with in-person lending it may the same time and sequence for all consumers; loan be hard to demarcate where marketing ends and a assessment criteria are formulaic; and borrowers with sale begins, ensure that a pre-agreement form was Table 1. CGAP exploratory research Partners (Year Research Was Conducted) Product Research Methods Topics FirstAccess (2014) Alternative data Qualitative consumer Disclosure; credit scoring service research reporting and information sharing Jumo, Busara Center for KopaCash Lab testing; field testing Disclosure; suitability; Behavioral Economics (2015) repayment Commercial Bank of Africa, M-Pawa Interactive SMS; data Marketing; suitability; Vodacom, TechnoServe, Arifu, analytics repayment Busara Center (2015) Kopo Kopo, Busara Center (2016) Grow Consumer interviews; Disclosure; suitability data analytics Pesa Zetu, Busara Center Peer-to-peer loans Consumer interviews; lab Disclosure; repayment (2016–2017) testing; data analytics M-Kopa, Credit Information Sharing Solar energy Interactive SMS; data Credit reporting and Kenya, TransUnion, Flag42 (2017) devices analytics information sharing 4 truly presented before a loan agreement was signed, • Complex “key facts” information. T&C documents or ensure that collections practices of individual tend to be long and complex. Most borrowers do employees are consistent with FSP policies. not take the time to review a multipage standard form contract or T&Cs presentation, and most This process standardization means that we can lenders do not provide a summary of the most assess consumers’ purchase experience as a series important T&Cs. of engagements with digital interfaces (typically • Unclear disclosure of data handling practices. the screen on their phone), and therefore measure Providers do not clearly communicate the types the quality of this experience with considerable of personal information or data they are collecting certainty across the entire customer base of from consumers, how they or their partners are a lender or a product. We analyze consumer using these data to make digital credit offers, and protection challenges in the five key areas of the whether consumers are able to access, correct, and lending and borrowing process: disclosure of loan restrict the sharing of their data. terms and conditions, marketing, suitability and • Lack of timely disclosure. For consumers to factor product design, repayment and collections, and price and other key T&Cs into their decision- credit reporting and information sharing. Where making, they must receive the information before relevant, we suggest possible good practices they click “I accept” and enter into the loan for digital lenders at each stage, based on our agreement. Some digital lenders disclose the experiments and others’ research. price only after the loan is executed (Mazer and Rowan 2016). Disclosure of loan terms and conditions Since disclosure is digital, it is particularly useful Most digital lenders’ disclosure practices fall and feasible to test different approaches. Providers short on sufficiency, consistency, or timeliness of can quickly and precisely measure the impact of information provided to consumers. Our review a specific approach on borrower behavior, for of products on offer in Kenya and Tanzania example, and with sufficient sample size, they reveals common gaps in disclosure and client can be highly certain of causality when aggregate communications: borrower behavior shifts in response to one type of message, disclosure, or menu format versus • Inaccurate presentation of costs of the product. another. This also strongly suggests that lenders This includes (i) failure to clearly state the actual should take an iterative approach to disclosure sum of finance charges the consumer will pay; and consumer understanding, by frequently testing (ii) use of monthly or weekly interest rate figures and refining their interfaces to positively affect instead of a standardized calculation such as consumer choice of loan size, repayment rates, annual percentage rate (APR); (iii) inconsistent and other behaviors that will be a win-win for both disclosure of finance charges across lenders ; 5 lenders and consumers. and (iv) nondisclosure of costs (and benefits) of other products that are bundled with the digital How to increase understanding of costs loan.6 and transparency in digital credit • Inaccessible terms and conditions. Many digital lenders offer access to the product T&Cs via only a In Kenya, digital credit provider Jumo used a series weblink, which cannot be viewed directly through of lab and field experiments to test different ways to the channel the consumer uses to borrow (i.e., on help consumers understand terms for its KopaCash the handset, unless the borrower has a smartphone product, and to observe borrower decision-making and data plan). during the loan application process (Mazer, Vancel, 5 See, e.g., the wide range in digital lenders’ APRs when charges are fully accounted for and standardized (Chege and Kaffenberger 2016). 6 E.g., some providers offer credit life insurance on digital credit, which raises the question of whether insurance is good value-for-money when loans are so small and short term. 5 Figure 1. Testing of formats for disclosure of loan cost Choose your repayment plan: Choose your repayment plan: 1. Repay 228 in 45 sec 1. Repay 200 + 28 in 45 sec VS. 2. Repay 236 in 1min and 30sec 2. Repay 200 + 36 in 1min and 30sec 3. Repay 244 in 2min and 25sec 3. Repay 200 + 44 in 2min and 25sec and Keyman 2016). This included a lab-based who viewed the content had a 7 percent lower experiment in Nairobi, where participants borrowed absolute delinquency rate. (See Figure 2.) money to participate in income-earning activities Summary T&Cs is possible on USSD or SMS 3.  and had to choose among various time periods channels. An improved, short summary of selected and costs when making their borrowing decision.7 key T&Cs that would fit on the USSD channel Three of the most promising findings from the lab Jumo uses for its KopaCash loan was developed. experiment were the following: Since this 2015 experiment, Jumo has updated Borrowers make better loan decisions when 1.  and expanded its disclosure of terms before loan costs are made salient . In this simulation, approval. It integrated several elements of the lab separating loan principal payments from finance experiment and other pro-consumer approaches, charges reduced defaults in income-earning as seen in Table 2. activities from 29.1 percent to 20 percent. (See Figure 1.) Other digital credit providers have tested Good design of the purchase process increases 2.  innovations in the loan enrollment process for consumer attention to T&Cs. “View T&Cs” is different customer segments. For example, Kopo often the last option on a loan product main menu. Kopo offers its Grow business loan product to By simply moving it to the next screen, followed merchants who use its digital payments services: by a short summary of the key facts from the each merchant can set its loan repayments as a T&Cs, viewing of T&Cs increased from 9.5 percent percentage of the value of payments received to 23.8 percent of consumers. Significantly, those from its customers via the Kopo Kopo payment Figure 2. Use of active choice to increase viewing of summary terms and conditions Welcome to Choose your Welcome to Kindly take a TOPCASH: loan amount: TOPCASH: minute to view Terms and 1. Request a 1. KES 200 1. Request a Conditions of loan 2. KES 400 VS. loan taking out a 2. About 3. Exit Loan 2. About loan: TOPCASH TOPCASH 3. View T&C’s 1. View Ts&Cs 2. Proceed to loan request 7 The experiment was conducted with lower-to-middle-income consumers in Nairobi at Busara Center’s lab testing facilities using computer screens that replicated the USSD menus used by Jumo. Participants went through several rounds of borrowing decisions and income-earning activities each, with their decisions and earnings tracked for each round. Because this was a lab test, the time periods were only a few minutes, versus a period of days and weeks as in actual digital credit deployments. The impact findings are therefore most useful for digital credit product design when considered on a comparative basis, not on an absolute basis (e.g., which approaches to messaging have more positive effect on borrower behavior). 6 Table 2. Updates to Jumo disclosure screens, October 2016 Screen 1: Separation of finance Choose your repayment plan: charges and principal   1.  Repay 1000 + 135 in 7 days   2.  Repay 1000 + 170 in 14 days   3.  Repay 1000 + 205 in 21 days *Back Screen 2: Separate line detailing Loan: 1000 loan fees and loan repayment details Loan Fees: 135 (13.5%) Loan term: 7 days Repayment: 1135 to be deducted from Airtel Money Wallet on   1. Confirm *Back Screen 3: Late payment penalty Failure to repay your loan by the due date will result in a late details payment fee of being added. You may also lose access to KopaCash.   1. Next *Back Screen 4: Active choice approach to Agree to the T&Cs below in order to proceed with your loan view T&Cs application. tc.jumo.world/akec   1. Agree   2.  View T&Cs *Back till (i.e., payments made in digital form rather these features are hard to implement when using than cash). As a result, the merchant’s repayment USSD. The elements include loan terms in large, period for the Grow advance will vary, depending bold fonts and fields consumers must complete on how much the merchant borrows, the number themselves (making it more likely that they of customers who pay digitally, and the share of understand their repayment obligations). digital revenue the merchant chooses to allocate for loan repayment. To help make this decision point Nonetheless, providers that use USSD, SMS, or SIM and the finance charges clearer to the borrower, Toolkit are not exempt from disclosing costs and Kopo Kopo uses “sliders” on its website tool so key terms properly and transparently to consumers. the borrower can test out different combinations of Some lenders, such as Jumo, that use USSD or loan size, loan term, and the share of transactions other more basic communication channels for loan allocated for repayment, and immediately see the enrollment disclose costs clearly via USSD, thus fee and estimated time to pay off the loan changes. proving that doing so is technically feasible, although not yet common.8 In fact, in 2016 the Competition Since Kopo Kopo’s borrower interface is a web Authority of Kenya issued a notice that all digital page, there are more options for the design and lenders must disclose costs of loans on the mobile user experience than with mobile screens and USSD handset before loan origination, thus establishing menus. CGAP’s review of loan application screens a minimum standard that all digital credit markets in Kenya found this to be the case with lenders that should strive for (Mazer 2016a). were using USSD or SIM Toolkit channels. These channels offer lenders fewer options to improve Disclosure of data handling disclosure than app-based channels do. They limit practices and consumer consent lenders’ ability to customize the interface and to data collection and use the menus’ character limitations. For example, Figure 3 shows the enrollment screen for an app- Some of the more successful digital credit deployments based lender compared to that of a lender that are partnerships between mobile network operators uses USSD. The app version has several nice design (MNOs) and lenders, where the lender leverages the elements to increase consumer understanding, but MNO’s distribution channels and customer data to 8 See also Martin and Mauree (2016). 7 Figure 3. App-based versus USSD-based loan enrollment screens (illustration) 67% 13:33 39% 8:54 TALA Select account 0770197305422 0770197305422 Congrats! You wish to request for Eazzy loan of Kshs 1000 on account YOU QUALIFY FOR “0770197305422” and have read, KSh 2000 understood and accepted the Terms and Conditions on link: http:// www.ke.equitybankgroup.com/ 30 days, 15% fee, due at the end Eazzy§loan ? Due Date Amount Due Cancel OK 29 Jul KSh 2300 1 TOTAL KSh 2300 I agree to repay this loan according to the schedule above. score and provide loan offers to large numbers of Figure 4. Consent for data use screen potential borrowers. While consumers have generally (illustration) consented to collection, use, and sharing of their personal information and transaction data in some fashion, it is important to note that in many cases this 100% “consent” process consisted solely of the customer’s acceptance of standard form contracts or services [LENDER REDACTED] needs agreements for mobile phone or mobile money access to accounts. Furthermore, as noted, T&Cs are typically Identity accessible only via the web. Thus, it is unlikely that Contacts the consumer has read and understood provisions related to which of their data were to be collected Location or shared, when, with whom, for which purpose, and SMS with which associated risks. There are some lenders, however, who are upfront about their data practices Phone in preloan disclosures, despite collecting extensive Photos/Media/Files data and personal information. (See Figure 4 for an Wi-Fi connection information example of the text used by one Kenyan lender to disclose its data-use practices. The message is clear ACCEPT and well-timed and does not require the consumer to open a link to view data policies or data provisions in lengthy T&Cs.) Strengthening this aspect of disclosure is increasingly important, given the strong likelihood that more customer-related data will be used in more ways to underwrite digital credit and for other purposes going forward. 8 Box 2. Developing informed consent approaches for digital data trails FirstAccess offers alternative data analytics and credit your phone, including phone calls, SMS, airtime scoring for FSPs in emerging markets. Its model top-ups, or a mobile money account. Questions? leverages data such as mobile phone call and mobile Call First Access 12345678 money transactions records to score loans, including This is a message from FirstAccess: FirstAccess for lenders such as microfinance institutions that ONLY uses your mobile phone records to make serve borrowers with lower levels of income, literacy, a loan recommendation to lenders. We NEVER and familiarity with formal credit. When FirstAccess share your personal information with anyone. launched in Tanzania, it conducted qualitative research Questions? Call FirstAccess 12345678 to understand how to help consumers with limited understanding of digital data trails meaningfully inform FirstAccess has used these messages as a template themselves of how FirstAccess would use their data, to develop consent messages for its partnerships and to uncover any concerns they might have about with telecommunications firms that are seeking data use. FirstAccess used the insights to design and authorization to share its customers’ digital data, as test a series of SMSs with information beyond that well as verbal explanations by their customer care staff already included in the SMS in which consumers were should consumers need further information. Beyond asked to authorize use of their data for FirstAccess to making consumers’ consent to data handling practices generate a credit score and provide it to a lender. The more meaningful, FirstAccess’ model also integrates original consent SMS read as follows: important “privacy by design” principles to reduce the risk of unauthorized or improper data sharing. This is a message from FirstAccess: If you just For example, FirstAccess does not share the mobile applied for a loan at Microfinance Bank and phone records with the lender. This keeps digital data authorize your mobile phone records to be separate from lending data and prevents aggregation included in your loan application, Reply 1 for Yes. and sharing without consent. FirstAccess also solicits Reply 2 for More Information. Reply 3 to Deny. a new authorization for each access of a consumer’s Since research showed that consumers wanted to mobile phone records. Together these practices give understand what mobile phone records are and how the consumer more control over when and how their FirstAccess used and shared this information, two data are used.a supplemental messages covered these points: This is a message from FirstAccess: Mobile phone records are information captured when you use Source: Mazer, Carta, and Kaffenberger (2014). Toward better standards on • Present a summary of the key terms of the disclosure in digital credit product, as a complement to the common practice of listing a weblink to T&Cs. The research findings suggest that there are certain • Provide price and other key information to the minimum transparency and disclosure standards consumer before the loan is accepted. that digital credit providers can and should follow, • Develop simple and transparent rules for using regardless of whether they use USSD, SIM Toolkit, and handling consumer data, and develop apps, or other media for their operations: effective ways to convey the provider’s data use and handling practices to the consumer before • Present a full accounting of all regular costs of loan origination. the loan both in monetary amount and APR, as well as costs of any other products or services that Marketing approaches to are bundled with the loan (e.g., companion deposit promote responsible borrowing product, mandatory insurance policy).9 • Provide a clear presentation of repayment due Traditionally, loans have been marketed face-to- dates, amounts, and penalty fees and when face in emerging markets. In contrast, marketing they will be assessed. Where relevant, note other of most digital credit is remote, with no in- consequences of nonrepayment. person interaction during product enrollment and 9 Ideally these products should be optional, with a separate opt-in step taken by the consumer. 9 Figure 5. Unsolicited credit offer via The risks of push marketing mobile phone (illustration) The clever sales messaging of push SMS and other unsolicited marketing approaches, coupled with an easy and automatic enrollment process, encourage 100% some borrowers to take on loans without thinking through whether they need the loan and how they will repay it.11 Although the lender carries most of the risk of poor repayment rates, borrowers face Dear Customer, your consequences as well, including loss of airtime [LENDER REDACTED] benefits in MNO-linked models (e.g., bonus airtime loan limit is Kshs. 1,000. To access your credits or airtime advances), inability to borrow limit, present your ID from this lender if they have a future specific at a [LENDER borrowing need, or inability to borrow from other REDACTED] shop to providers because of the black mark on their credit update your details. history (Ngigi 2016). Common characteristics of http://bit.ly/2aVHAgJ unsolicited offers of digital credit raise four consumer protection concerns. First, when the person is solicited simply because he or she is an MNO customer, the offer is unrelated to the original purpose for which he or she opened the account with the MNO (telecommunications, mobile money, or reliance on SMS, advertisements on the internet, value-added service). Either the customer did or app stores for marketing communications (see not consent or consent was obtained from earlier Figure 5). agreements, such as a standard form contract for a mobile phone or mobile money account. For This section addresses key issues in marketing example, one Tanzanian MNO’s service has very digital credit products to low-income mass-market broad terms of service that state “You accept that consumers and explores solutions to produce we may disclose or receive personal information or better borrower and lender outcomes. A central documents about you. . . for reasonable commercial challenge is increasing borrower intentionality— purposes connected to your use of the mobile that is, making sure that the borrower has thought service or the M-PESA Services, such as marketing through “do I need this loan and how will I repay and research related purposes.”12 it?”—especially when loan marketing is “push” (unsolicited by the consumer) rather than “pull” Unsolicited credit offers may be framed to (the consumer takes the initiative to seek out exploit certain behavioral biases, which may loan products). 10 Also important is how best to entice consumers to borrow even when they introduce some “friction” into the consumer’s do not have a specific use in mind for the loan. decision-making process, including for borrowers Behavioral research has described “present bias,” who tend to borrow and renew loans whenever where a person overvalues short-term outcomes and they have the option to do so. undervalues long-term outcomes, as well as “loss 10 Sometimes—as with prequalification offers for credit card marketing—this entails a specific credit offer, such as “You are qualified to borrow up to $XXX,” especially in partnerships where a lender teams with an MNO to reach out to potential borrowers from the MNO’s customer base. 11 This may help explain why some digital lenders have noted default rates as high as 40 percent or 50 percent in their first round of loan offers, where they send out invitations to a wide swath of prospective borrowers (CGAP interviews). 12 See also Ombija (2017). 10 aversion,” where a person overvalues the worth of Figure 6. Example of framing of loan something he or she possesses (such as a loan offer) sizes available (illustration) and shows a strong aversion to giving up that item, irrespective of its actual value (Kahneman, Knetsch, and Thaler 1991).13 Consumer research suggests that 100% push marketing may also lead some consumers to My Loans borrow without a clear use for the loan (McCaffrey, Your Loan Offers Obiero, and Mugweru 2013). When questioned, You qualify for the following some consumers say they do not want to miss out on loans. Select from the options below. a chance to borrow when they have not been able to easily access loans in the past. Others report that Select Loan Amount they may want to test the product, even to the point KES 1,500 500 of taking on a high-cost loan they do not need just to 2,000 “see what it’s about.” This risk is exacerbated when combined with poor disclosure of costs by many SELECT TERMS digital lenders. 4 weekly payments of KES 566 Consumers’ choices can be influenced by how Interest KES 264 offers are framed when they are considering Total Payments KES 2,264 several options (The Economist 2009 and Nofsinger 2008). This is especially germane in digital credit because the lender often suggests directly or indirectly the size of the loan amount—for example, the “limit” or maximum sum that can be borrowed— rather than the consumer initiating the request. These and other efforts to build up loan limits Interviews with consumers and lenders confirmed quickly can be costly for borrowers, because typical that many customers borrow at the suggested loan loans are relatively expensive. limit rather than propose a lower sum that would be sufficient to meet their immediate needs. Thus, Opt-in versus opt-out marketing a borrower’s loan size and term decision can be heavily influenced by the lender’s offer and even There is a simple alternative to unsolicited credit how the choices are presented on the digital offers: making consumers actively request—or interface. (See, e.g., Figure 6.) “opt-in” to—a credit offer or product information, by sending the request to the lender or accessing Finally, marketing messages may encourage the lender’s app or menu on their phone. Instead repeat borrowing by emphasizing higher loan of consumers having to “opt-out” by asking the limits. Some borrowers engage in multiple high- lender to stop sending them marketing messages, cost borrowing cycles in mere days, with the sole “opt-in” marketing relies on consumers to self- aim of increasing their loan limit. Among the factors register via channels such as short codes. Lenders contributing to this behavior are the instant feature could attract consumers via advertisements in of digital credit requests or credit limit checks and media, on the internet, in app stores, or even in- marketing messages that encourage consumers to person promotion by sales staff, agents, or others.14 borrow more to grow their loan limits, which are In fact, most app-based digital credit providers use linked to the fact that digital lenders typically start this approach because they do not have access consumers off with very low loan limits as a risk to customers’ mobile phone accounts and contact management strategy (Mazer and Fiorillo 2015). information through partnerships with MNOs. 13 For a description of how these behavioral biases are applied in marketing materials for credit card solicitations see Birken (2014). 14 Deceptive advertising and in-person pressure sales are beyond the scope of this Focus Note. 11 An opt-in approach may reduce the overall volume Suitability principles applied of applications and uptake of the credit offer. It also to digital credit could have a positive outcome for both responsible borrowing and lending, because repayment, A growing number of jurisdictions such as India, portfolio quality, and credit history suffer when Ghana, and South Africa have identified suitability push marketing results in borrowers taking on loans as a principle that could apply to lenders, including without due consideration for the loan use and an obligation that lenders accurately assess repayment source. There is no conclusive evidence individual consumers’ needs and capacities and across lenders as to the impact of push marketing sell only those products that are appropriate to on uptake and repayment behavior. This is an area and meet the needs of that consumer (or that where lenders would be well-served to conduct segment). 15 IFMR Trust has conducted financial further testing. diaries research in India to construct a detailed suitability assessment approach for microfinance Appropriate and tailored providers (Prathap and Khaitan 2016). For credit, products to meet the needs of the heart of the suitability principle is that lenders specific consumer segments should target and sell only those products that have features and in-built incentives that fit the Many of the early digital credit models offered circumstances of the defined customer segments simple “monoproducts”—very small, short-term, or individual customers to which they are being unsecured consumer credit with a fixed price for marketed. To meet this standard, digital lenders all customers. Many still do, even in markets with would need to better segment potential and several years of lending experience. This standard current customers, more carefully assess their product approach can make sense during the repayment capacity, and target appropriate use process of building an adequate scoring model cases—such as high-turnover microenterprises or and testing the proof of concept. However, as households that are experiencing health shocks. markets mature, consumers and the retail credit sector would benefit from diverse product offerings Lenders’ post-sale servicing costs are relatively that reflect the varying needs and risk profiles of low, because payments are collected remotely different customer segments. For example, one of and loan monitoring is automated, which should the earliest providers of digital credit in East Africa facilitate tailoring, innovations in product design, still offers all customers the same loan terms and and flexibility. This is beginning to happen in pricing several years after launch of the product East Africa as some newer entrants diversify and despite high loan volume, profitability, and strong customize their product types, tenor (amount of repayment performance. Instead, the lender could time to repay a loan), and pricing models. This have used its data to segment consumers as the includes charging daily interest (creating a form basis for risk-based—and generally lower cost— of “pay for what you use” approach to interest pricing and more diverse product offerings. Simply and fees), risk-based pricing both on initial and put, why should borrowers who have successfully recurring loans, and not charging penalties for late repaid 10 loans on time or early pay the same repayment—all of which are innovations that could interest rate on their 11th loan that they paid for on benefit from further testing and documentation of their first loan, despite having demonstrated strong impact (Ananth 2017). See Box 3 for insights from repayment behavior over and over? This is one the Grow product offered by Kopo Kopo. example of the need to shift toward better needs assessments and to apply “suitability” principles in Access to more and better data and information about digital credit, to encourage better choice, pricing, consumers, such as consumer characteristics and and tailoring of offerings to customer circumstances. needs and the purposes for the loan, can help lenders 15 For a discussion of suitability and its application to base-of-the-pyramid consumers, see Mazer, McKee, and Fiorillo (2014). 12 Box 3. Insights from Kopo Kopo’s merchants on loan use and borrowing patterns Some consumers may be borrowing on a recurring of just three days between paying one off and taking basis out of habit or because they do not want to another. miss out on a credit opportunity, rather than for a This seems to be more consistent with a pattern of specific purpose. For example, CGAP research with repeat borrowing than aligning with their stated Kopo Kopo customers found that many believed “a preference for “emergency only” use. Consumer smart businessperson should take credit whenever it lenders that use digital channels report similar is available, as a need will always arise” (Kaffenberger behavior: borrowers tend to repay early and borrow 2017). Nearly all had multiple loan sources beyond again quickly. These examples indicate that for lenders Kopo Kopo’s Grow. Customers reported that they to properly administer suitability principles, they may viewed Grow loans as being for use “in cases of need to not only address upfront assessment of emergency, when you need a loan quickly and are consumer needs, but also monitor borrower behavior willing to pay the higher cost.” However, a vast over multiple loan cycles to ensure product features majority actually repaid loans faster than expected, align with evolving consumer needs. and most took out new loans quickly, waiting a median determine customer suitability. Data could be self- Repayment and collections reported as well as verified data, which is already the foundation of most digital credit scorecards. The type Remote collection of digital loans can enable of data used often depends on data available to the flexibility in repayment frequency and amount. lender (e.g., voice, SMS, and mobile money transaction Borrowers can easily make small installment information provided to a lender by an MNO partner). payments via phone or the internet as they are Ideally, such data collection should be done only on able. When borrowers pay late, however, digital an opt-in basis, with meaningful consumer consent. lenders have more limited collections options While these approaches raise privacy and data security than their “in-person” peers who can send loan concerns (to be discussed later), if proper precautions officers to the door or, in the case of group are taken, they also could allow for more granular lending methodologies, rely on the microfinance understanding of consumer segments, occupations, group to enforce repayment. Digital lenders and other activities, which in turn could help improve face another distinct collections challenge: Their lenders’ assessments of borrower needs, repayment primary communication channel—SMS and other capacity, and the suitability of the loan offer. mobile-based messaging—is also the channel of choice for many different marketing messages While not fully verifiable, self-reported data willingly from MNOs and other firms. Borrowers who are shared by thousands of consumers might start to inundated with SMSs are less likely to pay attention reveal behavioral trends and segments relevant for to repayment reminders in their crowded SMS enabling suitable product design and sale. As Figure 7 inbox. For example, for a 2016 CGAP survey on shows, some lenders are gathering self-reported data DFS pricing transparency, a respondent described by integrating questions such as the purpose for downloading a lender’s app to borrow, then the loan (e.g., business versus personal expense) immediately removing the app to avoid receiving or income estimates into processes for customer reminder messages, only to reload the app when onboarding and loan origination. More interesting still he was ready to pay off the loan.16 At the same from a suitability perspective, tools such as interactive time, well-designed messaging may make the SMS are making it possible for lenders to have robust benefits of repayment and the consequences of conversations with customers before, during, and nonrepayment more salient for the borrower. after borrowing cycles, which can improve customer Our research with lenders shows that there are understanding and use of the product and enable several ways digital lenders can leverage the digital better segmentation and product diversification (see communication channel to increase probability of Box 4). repayment. 16 CGAP pricing transparency awareness survey, Kenya. 13 Figure 7. Example of customer self-reported data in app-based digital credit offer 71% 14:14 70% 14:17 70% 14:19 TALA TALA TALA What would you like to use your 56% 78% loan for? Please describe your main source Business Expense What would you like to use your of income What kind of business expense loan for? is this? Business Expense Start a business When did you start doing this? Please describe how you will use Personal Expense this loan in more detail start up CONTINUE On average, how much do you earn Check all that apply to you from this in KSh? Have a job: I work for someone who pays me Per Please describe your main source Select Answer.... of income job Do you always earn the same When did you start doing this? amount from this source? Jan 2015 Yes, I always earn the same amount. On average, how much do you earn from this in KSh? No, the amount I earn 45000 sometimes changes. SUBMIT Box 4. Using interactive SMS to support digital savings and borrowing in Tanzania In 2015, M-Pawa, a digital credit and savings product • Took Tsh 1,017 larger loans than they had before in Tanzania offered through a partnership between interaction with the learning content, repaid Commercial Bank of Africa and Vodacom, used the their loans 5.46 days sooner (both 95 percent digital learning platform Arifu to deliver learning confidence interval), and had Tsh 1,730 larger first content to Tanzanian farmers via interactive SMSs. The payments on their loans (99 percent confidence farmers opted in to receive the free content on how interval). to use M-Pawa and its savings and credit components. • Took Tsh 1,666 larger loans (99 percent Arifu used educational strategies like narrative-based confidence interval), had Tsh 2,654 lower amounts content, social norms, and interactive tools to help outstanding (90 percent confidence interval) at farmers register for M-Pawa, borrow, save, set savings 90 days, and made payments 3.42 days sooner goals, and calculate loan cost. Analysis of two years (90 percent confidence interval) than non-Arifu of pre- and post-treatment transaction data showed users. that the farmers who opted in and interacted with the learning content used both the savings and credit This case demonstrates how farmers’ ability to facilities more and to better effect than before and in save more on M-Pawa led them to obtain larger a complementary way. They also used M-Pawa more loan amounts and repay at higher rates when they and to better effect than those who did not opt in increased their M-Pawa savings. It shows how digital (Mazer Ravichandar, and Dyer 2017). credit products linked to a savings account can encourage not just borrowing, but also the use of The learners demonstrated the following financial savings to accumulate capital to use for household behaviors: or business needs. Similarly, since M-Pawa savings • More than doubled their savings account balances, behavior is part of the credit scoring model, increased from Tsh 2,673 to Tsh 7,120 after interaction savings activity helped these farmers get larger loans with the learning content (99 percent confidence when needed and in turn demonstrate improved interval). repayment behavior. 14 Optimizing the effect of Information Sharing Kenya to test an interactive repayment messages SMS platform where consumers could opt in to content that allowed them to learn about what As the research with Jumo in Kenya demonstrated “credit history” is, check theirs in real time, and for disclosure, increasing saliency of costs and correct any inaccurate information in their record. consequences can positively impact borrowing The content itself employed positive framing of behavior, including repayment. Similarly, lenders credit history as something that is linked to positive are experimenting with new approaches to repayment and helps consumers access future encourage repayment of digital credit that include loans. This approach aims to show borrowers their content, timing, and frequency of repayment actual credit history—which is overwhelmingly reminders; presentation of the reporting of positive for most borrowers—and link it to future borrowers’ on-time payments to credit bureaus in credit access more concretely. a positive light; use of behavioral design concepts, such as social norms, in SMS reminders; and the Initial evidence from the six-week (February to restructuring of delinquent loans to be flexible in March 2017) prepilot with M-Kopa customers response to borrowers’ circumstances. showed strong engagement—384 of the 1,632 invited customers opted in to receive learning Field research suggests that each lender should content; each consumed 5.7 messages on average. design its reminder messages with a clear Customers requested their credit history 225 times, concept of intended impact that balances carrots with 53 requests for follow-up from TransUnion and sticks, and then do iterative A/B testing regarding the data shared in their credit history. to see which content, timing, and frequency is There were also 601 uninvited people who accessed most effective.17 For example, Jumo’s Kenya field the learning content, primarily due to a radio host experiment on timing of repayment reminders learning of the technology and promoting it, found repayment responses were 8 percent higher showing a very strong word-of-mouth effect and for reminders sent in the evening than for those sent interest in this service. in the morning (Mazer, Vancel, and Keyman 2016). When analyzing the impact of the credit history Another common tactic to drive repayment in digital content on repayment, we saw that M-Kopa credit is SMS-based threats of negative listing in customers who opted into the SMS invites took a credit bureau. Some lenders send regular, even up more credit, were less likely to be blocked (by daily, messages to late borrowers, threatening the failing to make a payment in 90 days), and were borrower with being listed as a defaulter. Others more likely to have paid off in full (see Table 3.) A have discontinued negative listing messages because larger pilot is now underway to determine if the of consumer complaints and the lack of clear results. initial findings hold across a larger sample and to The generic threat of blacklisting may not work well address questions such as potential selection bias both because it is negatively framed—encouraging in consumers who chose to opt into the learning tunneling —and because the description of the 18 content. See Box 5 for more on repayment credit listing is not specific enough to conjure up the behavior. actual consequences in the borrower’s mind. Flexible payments and debt restructuring This is not to say credit history cannot be in an automated environment framed as an effective incentive for repayment. For example, off-grid energy provider M-Kopa Many digital lenders apply a fixed penalty— in Kenya partnered with TransUnion and Credit sometimes equivalent to the original interest 17 A/B testing refers to the randomized assignment and testing of different designs or approaches—in this case, different versions of USSD screens—to measure comparative impact and to determine the more effective approach. See, e.g., Dibner-Dunlap and Rathore (2016). 18 When people are faced with high stress or negative information, they tend to “tunnel” and ignore the stressful information points. 15 Table 3. Repayment of M-Kopa customers during credit history pre-pilot (February-March 2017) Average Blocked Finished Loan Credit Days Customers (%) Payment (%) Opted in: SMS invitation only 45 6.07 22.67 Opted in: Phone invitation then SMS 48 6.57 17.52 Did not opt in 35 25 11 Control 36 32.10 5.05 charge—and then allow a period of time (e.g., 90 processing is tiny in the absence of loan officers or 180 days) within which the borrower can make or physical infrastructure, could providers instead payment before the loan is written off (and in build in more flexibility for late-paying customers? some cases reported to the credit bureau). Since For example, doing so by offering (for an the marginal cost to digital lenders of payments additional fee) the option of extending their loan Box 5. Testing the “peer effect” in peer-to-peer lending with Pesa Zetu Pesa Zetu, a peer-to-peer digital lender, tested the included use of the name of a person lending money impact on repayment of content variations that were via the platform, reference to multiple lenders having designed to trigger behaviors by the borrower. The contributed to the borrower’s loan, characterization business model is based on individuals providing funds of the lenders as being similar to the borrower, and that are then on-lent to other individuals, all via mobile language that focuses on the lending platform rather money channels. One test varied the messages received than the individual peer lenders. by borrowers to see if different types of framing of The field testing showed borrowers had higher the “peer” element could improve borrowers’ on- repayment rates when receiving messages that time payment by creating a greater sense of social included either the name of an individual lender or obligation to the persons lending them the money. the name of the lending platform. Based on these Phase one of the research tested four different types indicative results, the peer-to-peer lending platform of repayment reminders plus a control message is now being tested to measure actual effects of the (see Figure B5-1). The different types of reminders different treatments on repayment behavior. Figure B5-1. Treatments tested for framing of repayment obligation of peer-to-peer lending platform 100% 100% 100% 100% Your loan of Your loan of Pesa Zetu lenders Your loan of KES[XX] is from KES[XX] is are just like you KES[XX] is being [Name], a Pesa provided by 5 and me. Your loan provided through Zetu lender. Pesa Zetu of KES[XX] is Pesa Zetu. Pesa They are lenders. They are provided by Zetu expects you counting on you all counting on someone like you. to repay by [date]. to repay by you to repay by They are counting Zangu. Zako. Zetu. [date]. Zangu. [date]. Zangu, on you to repay Zako. Zetu. Zako, Zetu. by [date]. Zangu. Zako. Zetu. 1. Lender Name 2. Multiple Lenders 3. Lender Similarity 4. Institutional Lender 16 term (provided the rollover is not too expensive borrowers by offering them better prices or options or permitted multiple times, which is a common (Mazer and Rowan 2016). problem cited in payday lending models) (Burke et al. 2014). This shift in approach could reframe This may explain—but does not excuse—the less the late payment in the borrower’s mind from being than full compliance with credit reporting rules, a penalty to being an opportunity to work his or her which has been observed in several active digital way out of debt or to buy more time to repay. This credit markets that are in the early stages of in turn may reduce write-offs, increase loyalty, and developing their credit information systems. For improve the borrower’s feeling of reciprocity with example, the largest digital lender in Kenya reported the lender. only negative repayment data on its customers to the credit bureau until May 2016, despite the CGAP is partnering with a lender that offers Kenya Credit Reference Bureau Regulations (2013) borrowers who miss a payment the ability to choose requiring full-file reporting. At the same time, a new loan tenor from several options—each TransUnion Kenya in 2016 noted more than 400,000 of which includes a penalty charge so as not to consumers were listed as defaulters for loans of encourage strategic delinquency. While the Ksh 200 (approximately US$2) in its credit bureau— lender’s call center is currently managing this, as these are nearly certain to be digital loan borrowers the portfolio scales, the lender may need to use (Ngigi 2016). This raises concerns about whether automated SMS-based options, which are being credit reporting practices are proportionate and fair tested now. The potential consumer welfare gains for consumers, as delinquency and default on very from debt restructuring options, combined with small loans could have significant consequences, the lender’s ability to automate this process and whereas consumers’ positive repayment history may use low-cost collections channels, makes this an not always be reported by lenders. (Box 6 addresses exciting frontier in consumer protection that digital the accuracy of digital credit histories.) lenders should explore and test. Unlike large, incumbent telecommunications and Credit reporting and FSPs, digital lenders that do not have access to information sharing information on applicants’ bank or mobile money transactions often must rely on more intrusive Digital data trails drive the scoring models of alternative data to build their scoring models. digital credit deployments and enable lenders to App-based FinTechs, for example, often ask assess and manage the risk of lending to people consumers to authorize access to a wide range of with whom they have had no prior interaction data stored on the handset, including from social or credit relationship. Yet mobile money or media, mobile wallets, or e-mails (which they scan telecommunications services providers often treat for references to past due loans, mobile money consumers’ mobile money transactions records as transaction receipts, and other potentially relevant proprietary—lenders could otherwise use these indicators of creditworthiness). records to estimate borrower cash flow and to build credit scores. The lack of control of their own data Should consumers have to give up their privacy prevents customers from maximizing the utility to this extent? Some consumers must authorize of the data trail they generate, for example, to access to more sensitive and extensive personal receive competing credit offers. Lack of consumer information and data because their preferred control also helps dominant telecommunications lender cannot obtain arguably more relevant credit companies or FSPs leverage their market-leader and financial digital history if MNOs and other status in telecommunications, banking, or mobile data holders do not provide it access. Is there a financial services to suppress competition by scenario where consumers could control and use restricting new entrants and disadvantaging their digital data trail to receive competing lending other lenders that could otherwise benefit digital offers safely and securely? If so, perhaps lenders 17 Box 6. How accurate is digital credit history? Evidence from a consumer survey in Kenya It is important to enable borrowers to actively review and of consumers who checked their credit history reported check their information. For many, this may be their first that it contained incorrect information in their report ever loan reported to the credit bureau. This becomes (see Table B6-1). A deeper look into the inaccurate even more important given concerns that have been information these consumers reported (by the lender raised regarding accuracy of credit history, particularly involved) revealed a need to improve the accuracy of digital credit records. A pilot survey of digital credit digital credit histories in credit bureaus, because many users in Nairobi found that 92 out of 420 (22 percent) of the errors came from digital lenders (see Table B6-2). Table B6-1. Types of incorrect information consumers reported in their credit history Discrepancy reported % of Responses Incorrect loan balance 32.6 Loan already paid listed as unpaid 22.8 No credit history available 10.9 Not all loans are included 31.5 The respondent details do not match  2.2 Table B6-2. Lenders consumers reported having incorrect information in their credit history Lender or loan type % of Responses KCB Mpesa 20.7 Mshwari 17.4 Multiple mobile loans 7.6 Tala 4.3 M-Coop 2.2 Equitel 1.1 All nondigital loans 14.1 Credit history is entirely missing 32.6 would not need to capture so much personal, or sale of consumer transactional data by those social, and handset data. who collect it without express and restricted consumer consent. Toward a consumer-led data • Easy, secure processes for consumers to share sharing environment their own data. In some markets, there are private- sector scorecard providers that gather and structure To address data-related concerns, providers and alternative consumer data. A constructive next step policy makers should work together to develop new could be to encourage or mandate a neutral channel data use standards for digital credit that are pro- through which consumers could export data from consumer and pro-competition. They should take into their transactional accounts in a standardized account three principles (Mazer and McKee 2016): format,19 such as an open applied program interface (Hanouch and Moracyzinski 2016). • Consent and use restrictions. For example, • Standards on what types of data can be shared restricting use of data to a per-transaction basis, versus what should be kept private. While having clear user consent, and prohibiting sharing consumers should be able to control and share 19 Similar efforts are underway in the United Kingdom (Jones 2016 and EC 2016). 18 their financial transaction data, digital lenders practices could then be adopted by providers might reasonably argue that there is other through collective action, such as industry codes of information that they collect or generate that conduct or regulation. Authorities can also adopt should be considered proprietary (e.g., data on an encouraging stance toward ongoing consumer the pages consumers visit on the lender’s website). research and testing by firms to further strengthen Lenders could work with policy makers and responsible lending and borrowing. consumer organizations to define this line between consumer-controlled and proprietary information Specific regulatory implications and the associated sharing rules. from findings in each area In addition to using collective action to apply Disclosure. Transparency is essential to developing these principles for gathering and reporting data, healthy and competitive retail credit markets. individual providers should continue to test new Our demonstrations with lenders point to better ways to increase consumers’ understanding and disclosure practices that can be adopted at low sense of ownership of their digital history. Most cost. At the same time, regulation is likely to be borrowers are building a strong positive repayment required to ensure that all providers implement history that they can leverage in many ways. adequate and consistent disclosure practices across the market. Implications of the findings for digital credit Marketing. Lenders’ use of push marketing policy and regulation practices to provide consumers with individual digital credit offers on their mobile handsets Much of CGAP’s digital credit research has raise privacy and consumer protection concerns. focused on individual providers, and specifically In addition, the way loan offers are framed may on modifying their communications and product encourage borrowing with no clear purpose or design. The emerging good practices reported intentionality. Regulators should consult with in this paper should provide the foundation for MNOs, FSPs, and alternative data credit scoring standards of responsible lending that could begin firms to better understand their policies for using ensuring minimum consumer protections for the consumer data and contact information when entire digital credit sector. It is encouraging that marketing digital credit products. Rules limiting or so many of these practices are both a win-win— prohibiting push marketing should be considered they bring benefits to lenders as well as their to address any issues. customers—and relatively low-cost to implement. This could encourage other lenders to voluntarily Suitability and product design. Where suitability adopt these practices and be inspired to further principles are deemed appropriate for a jurisdiction, explore their own solutions that increase value for regulators may need to develop further guidance their customers and their business. as to how they may be applied by different types of lenders, and for different customer segments Lenders should document and promote good and types of products, including digital credit. practices that help to balance provider and To achieve this, regulators can encourage digital consumer interests and healthy competition in lenders to strengthen their customer segmentation the market. Progress in these areas also requires and product diversification efforts as the basis for collective action, including industry-regulator suitability-based lending, and then periodically consultation, formal industry standards, and policy review the lenders’ portfolios and policies and or regulatory measures. Given the diverse range of gather information through consumer surveys providers and products and the channels they use (Mazer 2016a). to disclose costs and T&Cs to consumers, regulators and lenders should work together to explore further Repayment and collections. Supervisors should good practices in each of these areas. These good monitor how digital lenders determine penalty 19 charges, the cost of these charges to consumers, how In the case of credit, the digital delivery—from these charges are communicated to consumers, and marketing to onboarding to repayment and data lenders’ policy for writing off delinquent payments. sharing—shapes consumer behavior and alters how Supervisors may want to go beyond monitoring, for certain consumer protection risks that are common example, by reviewing the standardized messaging to all consumer lending play out for both the scripts and call center protocols digital lenders borrower and the lender. Digital credit markets are use to communicate with delinquent borrowers still nascent in developing countries and emerging to ensure clear and responsible communication of markets. It is not yet clear which market-level penalty charges and collections practices. issues will arise regarding consumer protection and sustainable development of the sector. Will these Credit reporting and information sharing. Policy risks translate into serious problems or will their makers should, at a minimum, address gaps in impacts be fairly small and commensurate with coverage of and compliance with credit reporting loan sizes? Even if the absolute value of problem regimes across the various types of digital lenders in loans is small, or if the harms and risks are limited the market. They may also want to work with credit to certain segments, there may be a case for registers and lenders to explore the inclusion of new proactive standards and rules if relatively large and valuable customer data (e.g., mobile money numbers of low-income consumers could suffer the and payments data) in credit reporting systems. consequences of problems such as irresponsible Finally, policy makers may want to consider whether lending, erroneous listing, or data breaches. The the consequences of delinquency and default are “wait and see” approach might carry more consumer disproportionate for very small loans, especially and market-level risks for this product type than for when so many consumers are new to the product and other types. to the formal financial system. Policy makers could consider a range of options, such as ensuring that Providers and policy makers are already testing and existing rules for sharing both negative and positive developing solutions to these risks, as shown by credit information are enforced, enacting new rules the examples in this paper of improved disclosure, on reporting requirements for a specific subset of marketing, suitability, repayment and collections credit products, using moral suasion for noncompliant procedures, and credit reporting. Demonstrations lenders, disclosing consequences of nonpayment with diverse digital credit models have shown cost- more clearly to consumers, and undertaking financial effective ways to offer market-ready products that capability and consumer awareness efforts. better balance lender and borrower interests and embed consumer protection principles in product The tempest, the design and delivery. We hope the evidence and teapot, or both? examples will be an impetus for key stakeholders— lenders and other industry actors, regulators, and The “instant-automated-remote” model of digital consumer advocates—to take action. Practical credit offers significant potential to advance and enforceable good-practice standards will be financial inclusion by making possible much smaller needed to optimize the potential and manage the loan sizes, larger scale, and innovative business inherent risks of this important innovation. models than would otherwise be possible. At the same time, most digital credit today is relatively References high-cost consumer lending, and we should not forget the credit bubbles and repayment crises of AFI (Alliance for Financial Inclusion). 2015. “Digitally the past. The scalability of digital credit models, the Delivered Credit Policy Guidance Note and Results generally high cost of credit, and the typical target from Regulators Survey.” Guideline Note No. 17. populations—lower-middle class and working class Kuala Lumpur, Malaysia: Alliance for Financial consumers in emerging markets—argue strongly Inclusion. September. http://www.afi-global.org/ for standards on appropriate product design and sites/default/files/publications/guidelinenote-17_ consumer protection measures. cemc_digitally_delivered.pdf 20 Ananth, Bindu. 2017. “Leveraging Fintech for Hanouch, Michel, and Olga Moracyzinski. 2016. Risk-based Pricing & Personalization.” Blog post, “Riding the ‘Rails’: Unlocking Innovation with Open 4 May. http://www.ifmr.co.in/blog/2017/05/04/ APIs.” Blog post, 7 December. http://www.cgap.org/ leveraging-fintech-for-risk-based-pricing- blog/riding-%E2%80%9Crails%E2%80%9D-unlocking- personalisation/?utm_source=feedburner&utm_ innovation-open-apis medium=email&utm_campaign=Feed%3A+ifmr+ %28IFMR+Blog%29 Jones, Huw. 2016. “UK Plans New Data Rules to Boost Banking Competition.” Reuters, 9 February. http:// Birken, Emily Guy. 2014. “4 Ways Credit Cards uk.reuters.com/article/uk-britain-banks-regulations- Manipulate You into More Debt.” Wisebread, idUKKCN0VI009 7 July. http://www.wisebread.com/4-ways-credit- cards-manipulate-you-into-more-debt Kaffenberger, Michelle, and Quang Nguyen. 2017. “Responsible Digital Credit for Merchants: Insights from Burke, Kathleen, Jonathan Lanning, Jesse Leary, Kenya.” Blog post, 8 February. www.cgap.org/blog/ and Jialan Wang. 2014. “CFPB Data Point: responsible-digital-credit-merchants-insights-kenya Payday Lending.” Washington, D.C.: Consumer Kahneman, Daniel, Jack L. Knetsch, and Richard H. Financial Protection Bureau, March. http://files. Thaler. 1991. “Anomalies: The Endowment Effect, consumerfinance.gov/f/201403_cfpb_report_ Loss Aversion, and Status Quo Bias.” The Journal of payday-lending.pdf Economic Perspectives, 5(1), pp. 193–206. https:// www.princeton.edu/~kahneman/docs/Publications/ Chege, Patrick, and Michelle Kaffenberger. 2016. Anomalies_DK_JLK_RHT_1991.pdf “Digital Credit in Kenya: Time for Celebration or Concern?” Blog post, 3 October. http://www.cgap. Martin, Rebecca, and Vijay Mauree. 2016. org/blog/digital-credit-kenya-time-celebration-or- “Commonly Identified Consumer Protection Themes concern for Digital Financial Services.” Geneva: International Telecommunications Union, May. http://www.itu. Chen, Greg, and Rafe Mazer. 2016. “Instant, int/en/ITU-T/focusgroups/dfs/Documents/09_2016/ Automated, Remote: The Key Attributes of Digital ConsumerProtectionThemesForBestPractices.pdf Credit.” Blog post, 8 February. http://www.cgap.org/ blog/instant-automated-remote-key-attributes-digital- Mazer, Rafe. 2016a. “3 Steps Policy-Makers credit Can Take Now on Digital Credit.” Blog post, 6 December. http://www.cgap.org/blog/3-steps- Dibner-Dunlap, Aaron, and Yumna Rathore. 2016. policy-makers-can-take-now-digital-credit “Beyond RCTs: How Rapid Fire Testing Can Build Better Financial Products.” Blog post, 1 August. http:// ———. 2016b. “Kenya Ends Hidden Costs for www.poverty-action.org/blog/beyond-rcts-how-rapid- Digital Financial Services.” Blog post, 2 November. fire-testing-can-build-better-financial-products. http://www.cgap.org/blog/kenya-ends-hidden- costs-digital-financial-services EC (European Commission). 2016. “Guidelines on the Right to Data Portability.” Brussels: EC, 13 December. Mazer, Rafe, and Alexandra Fiorillo. 2015. “Digital http://ec.europa.eu/newsroom/document.cfm?doc_ Credit: Consumer Protection for M-Shwari and id=43822 M-Pawa Users.” Blog post, 21 April. http://www. cgap.org/blog/digital-credit-consumer-protection- GSMA. 2017. “State of the Industry Report on m-shwari-and-m-pawa-users Mobile Money 2016.” London: GSMA. https://www. gsma.com/mobilefordevelopment/wp-content/ Mazer, Rafe, and Kate McKee. 2016. “‘Co-Ownership’ uploads/2017/03/GSMA_State-of-the-Industry- of Mobile Money Data: Building a Kenyan Credit Report-on-Mobile-Money_2016.pdf Bureau for the Digital Age.” CIS Kenya Digest, April- May: pp. 4–7. 21 Mazer, Rafe, and Philip Rowan. 2016. “Competition Ombija, Sarah. 2017. “Review of DFS in Mobile Financial Services: Lessons from Kenya User Agreements in Africa: A Consumer and Tanzania.” Working Paper. Washington, D.C.: Protection Perspective.” Geneva: International CGAP, January. http://www.cgap.org/sites/default/ Telecommunications Union, January. http:// files/Working-Paper-Competition-in-MFS-Kenya- www.itu.int/en/ITU-T/focusgroups/dfs/ Tanzania-Jan-2016.pdf Documents/01_2017/ITU_FGDFS_Report-on- Review-of-DFS-User-Agreements-in-Africa.pdf Mazer, Rafe, James Vancel, and Ariana Keyman. 2016. “Finding ‘Win-Win’ in Digitally Delivered Omondi, Dominic. 2017. “A Glimpse into Consumer Credit.” Blog post, 13 January. http:// Kenya’s Billion Shillings Wee Hour-Business www.cgap.org/blog/finding-“win-win”-digitally- Oiling the Economy.” The Standard , 21 March. delivered-consumer-credit https://www.standardmedia.co.ke/business/ article/2001233461/a-glimpse-into-kenya-s-wee- Mazer, Rafe, Kate McKee, and Alexandra hour-business-oiling-the-economy Fiorillo. 2014. “Applying Behavioral Insights in Consumer Protection Policy.” Focus Note 95. Owens, John. 2017. “What Does Responsible Washington, D.C.: CGAP, June. http://www.cgap. Online and Digital Credit Look Like?” Blog post, org/publications/applying-behavioral-insights- 4 January. https://cfi-blog.org/2017/01/04/what- consumer-protection-policy does-responsible-online-and-digital-credit-look- like/ Mazer, Rafe, Nikhil Ravichandar, and Julian Dyer. 2017. “Can Digital Savings Reduce Risks in Digital Pande, Varad, and Niloufer Memon. 2017. Credit?” Blog post, 10 May. http://www.cgap.org/ “Protecting Borrowers in the Digital Era.” LiveMint, blog/can-digital-savings-reduce-risks-digital-credit 31 January. http://www.livemint.com/Opinion/ th3C4wuS4Qh6zzTYpoBrML/Protecting-borrowers- McCaffrey, Mike, Olivia Obiero, and George in-the-digital-era.html Mugweru. 2013. “M-Shwari: Market Reactions and Potential Improvements.” Briefing Note 139. Prathap, Vaishnavi, and Rachit Khaitan. 2016. MicroSave, February. http://www.microsave.net/ “When Is Microcredit Unsuitable?” IFF Working files/pdf/BN_139_M_Shwari.pdf Paper Series No. 2016-02. IFMR Finance Foundation, December. http://foundation. Ngigi, George. 2016. “Pain of Kenyans blacklisted ifmr.co.in/wp-content/uploads/2017/01/When- for Amounts as Small as Sh100 in Mobile Loans, is-Microcredit-Unsuitable-Guidelines-for- Bank Fees.” Daily Nation , 9 September. http:// Lending.pdf www.nation.co.ke/business/Pain-of-Kenyans- blacklisted-for-amounts-as-small-as-Sh100/996- Robinson, Ian, and Graham Wright. 2016. “Digital 3374952-k2dkdvz/ Credit Is Not Always Good for Financial Inclusion.” Blog post, 13 December. http://www.fsduganda. Nofsinger, John. 2008. “Surprised Again? The or.ug/data/dnews/18/Digital-Credit-is-not-always- Anchoring Bias of Investors.” Psychology Today, good-for-financial-inclusion.html 21 July. https://www.psychologytoday.com/blog/ mind-my-money/200807/surprised-again-the- Rosenberg, Richard, Scott Gaul, William Ford, anchoring-bias-investors and Olga Tomilova. 2013. “Microcredit Interest Rates and Their Determinants 2004-2011.” Forum. Ochieng, Moses. 2016. “The New Moneylenders of Washington, D.C.: CGAP, June. http://www.cgap. Nairobi.” Blog post, 11 April. https://www.fsdafrica. org/sites/default/files/Forum-Microcredit%20 org/knowledge-hub/blog/the-new-money-lenders- Interest%20Rates%20and%20Their%20 of-nairobi/ Determinants-June-2013.pdf 22 The Economist. 2009. “The Importance of Irrelevant Vidal, Maria Fernandez, and Byoung-Hwa Hwang. Alternatives.” The Economist, 22 May. http://www. 2017. “Digital Credit’s Evolving Landscape: 3 Things economist.com/blogs/democracyinamerica/2009/05/ You Need to Know.” Blog post, 20 April. http:// the_independence_of_irrelevant www.cgap.org/blog/digital-credit%E2%80%99s- evolving-landscape-3-things-you-need-know 23 Annex. Behavioral biases Saliency. The mobile user interface may make it in digital delivery of more difficult than in conventional credit processes credit products for low-income consumers to identify costs and key T&Cs that should affect their borrowing decision. Several insights from the field of behavioral Also, the digital nature of the borrowing experience economics may be particularly relevant when may feel less “real” for the borrower. In turn, the considering consumer behavior in the “instant- borrower may be unclear on how to pay, forget to automated-remote” world of digital credit. pay on time because the payment is so small and is digital, or prioritize paying other nondigital debts Hyperbolic discounting. Consumers often because the sum is larger or the consequences of overvalue and prioritize short-term gains over default seem more severe. longer-term benefits. Similarly, consumers may focus on the immediate appeal of money in their Default settings. Consumers often accept default mobile wallet that digital credit offers, while conditions of products irrespective of what their true ignoring the expensive nature of this debt. preferences may be. For example, the T&Cs to which the borrower concurs by clicking “I Agree” may not Anchoring. Consumers’ choices are often based give the borrower any choice on how their data on reference points in the information that is are collected, used, stored, shared, or reported to available to them as they make decisions. Digital the credit bureau. Borrowers may accept conditions credit marketing messages that highlight the they may not feel comfortable with simply because maximum loan amount available to a consumer the default setting is to accept data sharing. may cause them to borrow more than they otherwise would. Status quo. The borrower may take frequent repeat loans—borrowing more out of habit or to hedge Loss aversion. People often overvalue something their bets and to make sure they have the funds on they have (or have been offered). For example, hand, rather than taking the next loan only to meet credit offers that include “you are already qualified a specific need when it arises. This could also impact for . . .” or “don’t miss out . . .” may drive more the probability that borrowers will shop around consumers to take loans, for fear of missing out on and compare new credit offers or remain with their the opportunity. existing digital lender, even if it is more expensive. No. 108 August 2017 Please share this Focus Note with your colleagues or request extra copies of this paper or others in this series. CGAP welcomes your comments on this paper. All CGAP publications are available on the CGAP Web site at www.cgap.org. CGAP 1818 H Street, NW MSN IS7-700 Washington, DC 20433 USA Tel: 202-473-9594 Fax: 202-522-3744 Email: cgap@worldbank.org The authors of this Focus Note are Rafe Mazer and Kate McKee. and vulnerable people. Previously, McKee was CGAP’s senior © CGAP, 2017 Mazer, an independent consultant, previously served as CGAP’s policy adviser for consumer protection and responsible finance. financial sector specialist on digital credit and behaviorally The authors thank the teams at FirstAccess, Jumo, Kopo informed consumer protection policy. McKee is the head of a Kopo, M-Kopa, Pesa Zetu, and Vodacom Tanzania for their start-up global partnership that promotes livelihoods programs collaboration in developing innovative approaches to consumer that use the graduation approach to help the extreme poor protection in digital credit. Suggested citation: Mazer, Rafe, and Kate McKee. 2017. “Consumer Protection in Digital Credit.” Focus Note 108. Washington, D.C.: CGAP, August. ISBN: 978-1-62696-080-0