First published by NextBillion and re-posted with permission.
Technology has vast potential to improve lives, and financial inclusion is no exception. Yet with half a million digital borrowers blacklisted on Transunion Credit Reference Bureau in Kenya – many for defaulting on loans of just a few dollars – it’s clear that digital finance also has its downsides.
Is it possible to link digital finance with client protection? Yes, but it’s not an easy task, even if the incentives are clear. The process involves building client trust, improving providers’ reputations, attracting impact investors and mitigating any regulator concerns. The Smart Campaign (hereinafter Smart) and MFR are working to achieve this by adapting client protection standards to digital business models, detailing concrete guidelines to improve practices, and providing a credible way to prove adherence to them (i.e. Smart Certification).
Smart, housed at the Center for Financial Inclusion at Accion, is the self-regulation standard-setting body of reference for client protection in inclusive finance: It has promoted good practices in the industry for over 10 years. Smart maintains a rigorous certification program, elevates the client voice, convenes partners to effect change at the national level, and has led the effort to update industry-facing client protection standards to encompass digital credit models. MFR, an independent rating agency specializing in inclusive finance, has conducted 57 percent of Smart client protection certifications, and is the sole assessor for the GSMA Mobile Money Certification.
Smart and MFR joined forces in 2018 to draft client protection standards applicable to digital credit providers, and pilot tested them in the field with two different companies in Kenya. These companies included 4G Capital, using a hybrid model based on both technology and in-person interaction with clients, and Tala, using a low-touch model where the client interactions mainly happen through a dedicated app. We shared the first findings of the pilot at the European Microfinance Week 2018, organised by the European Microfinance Platform. The draft standards, revised in light of the pilot feedback, are now available for a six-week period of public comment, allowing experts and the industry to weigh in and shape the final version of the standards. The draft standards were defined by Smart and MFR building upon multiple experiences, including Smart’s Fintech Protects Community of Practice, the SPTF Social Investor Working Group, the Responsible Finance Forum, GSMA Mobile Money Programme, GOGLA, Better Than Cash Alliance, ITU, the World Bank, CGAP and others.
VIEWING CLIENT PROTECTION THROUGH A DIGITAL LENS
What changes when one looks at client protection through a digital lens? The objective clearly does not: Keeping clients first remains the ultimate goal, and the motivation for many of us to go to work each day. But other aspects do change. Here are some of the lessons learned from our research, and the client protection practices we recommend, based on these lessons. Please refer to the full list of draft standards for more details.
First of all, the time when a financial service used to be provided by one organization is over. Recognizing the fragmentation of functions among ecosystem players, the draft standards propose holding the DFS provider accountable for the functions which are under its direct or indirect control. When the DFS provider is partnering with an outsourced company (e.g. a call center), we encourage it to manage the external client protection implications. Conversely, when the DFS provider is a user of a service with negligible negotiation power vis-à-vis the external company (e.g. a mobile network operator), this expectation is not realistic. Hence, synergies with other initiatives, such as the GSMA Mobile Money Provider certification, should be explored. As the public comment period continues, industry feedback will be especially crucial for shaping the final standards in light of the complexity of these partnerships.
CLIENT PROTECTION IN DIGITAL LENDING
When it comes to client protection in digital finance, algorithms for loan underwriting cause the most excitement – and the most headaches. If external assessors cannot look inside the algorithm for proprietary reasons, then we encourage them to review its governance (e.g. systems to manage the risk of over-indebtedness and discrimination) and its results (e.g. loan portfolio quality). For instance, one good practice we’ve observed is to establish an underwriting policy that actively and deliberately excludes any variables that may be discriminatory (e.g. gender, race, ethnicity, religion, national origin, sexual orientation, disability, medical history) – even if these factors may correlate with repayment likelihood.
The common trial-and-error approach to product design and algorithm development can be an effective innovation technique, as long as the inevitable experimental mistakes are not paid for by clients. To that end, one positive client protection measure that emerged involved not reporting clients to the credit bureau if they failed to repay loans taken during the underwriting algorithm’s initial stage of development. And when it comes to the question of whether to use client cash-flow analysis or alternative data to estimate the repayment likelihood, we recommend a risk-based approach that uses different methods to assess the borrower’s financial information, depending on their risk of over-indebtedness.
The ease of access enabled by digital finance is exciting. At the same time, immediate access may lead to impulsive, unnecessary debt. Compared to traditional finance, this increases concerns about client harm – and it’s something providers should consider when clients walk in more than when they drop off. That’s why we’ve explicitly incorporated “time to think” and analysis of loan use into the draft standards methodology. One good example of this from the pilots involved giving clients two days after their loan approval to decide if they actually wanted to take the loan.
THE QUESTION OF RISK VS. RETURN IN LOAN PRICING
Solving the risk-return and client value equation becomes even more paramount in digital finance. But though the link between growth, portfolio quality and prices remains valid, what changes is the higher risk appetite – especially in low-touch models – and the associated expectation of higher returns. Moving from high to low-touch models, the predominant cost to run the business generally changes from operating expenses to loan loss expenses. A 2018 CGAP research report in Tanzania calculated loan default rates over 90 days ranging from 35 percent for borrowers taking their first loan to 12 percent for borrowers taking their 7th – 10th subsequent loans – significantly higher than the single digit indicators of portfolio quality (e.g. Portfolio At Risk more than 90 days) in traditional microfinance.
In light of this increased risk, our standards will incorporate benchmarks of portfolio quality and pricing, among others, to digital models as the data become available. However, the standards maintain the notion that poor portfolio quality costs shall not be passed on to clients (who repay) through high prices. We recommend that the pricing of very short-term loans be described as Monthly Percentage Rate – still easily comparable to the Annual Percentage Rate – and that it be compared with the Monthly Percentage Rate of loans of similar amount and tenure. The draft standards explicitly consider price trends and risk-based pricing as evidence of a lending algorithm’s capacity to learn.
OTHER ISSUES TO CONSIDER
Human interaction in financial services may be changing in light of the digital revolution, but it remains important. Live interaction should be available for clients to understand what they are buying and to solve complaints. In-person interaction is key to building trust with the most vulnerable populations, especially during some critical phases of financial service use (Uniting Tech and Touch, ACCION). However, not all digital lenders target the low-income population, and effective live interaction may take the form of in-person assistance, voice or live chat support, depending on what target clients are comfortable with. As the pilots demonstrated, call centers offer the opportunity to monitor the recorded calls and response time, rewarding good customer support with bonuses.
We’ve also strengthened our client protection methodology in the areas of data privacy and security. While it may be too early for the majority of providers to implement the General Data Protection Regulation in full, many of its concepts have been included in the standards in spirit – for example the notion of minimizing and justifying the personal data collected, kept and shared.
As you can see in the above examples, linking digital finance and client protection is not always easy. We still have much to learn, and we will need to constantly interpret the standards in ways that make sense in the variety of cases that exist, whether that involves a provider self-assessing and self-improving, an investor doing due diligence, or Smart certifying a provider. Digital finance is a complex industry, and defining appropriate client protection is an equally complex task. However, we all have the responsibility to contribute to this important process, starting now. We believe that client protection will be a key factor in the ongoing success of digital finance providers – and it remains an essential element of the financial inclusion movement. Let’s join forces in this journey: Share your views on our draft standards during the public comment period, which runs until March 22, and help Smart improve upon the standards based on the insights and real experience of the industry.
by Lucia Spaggiari, Business Development Director at MFR & Isabelle Barrès, Global Director of the Smart Campaign at Accion.
Image courtesy of WorldRemit Comms.
First published by NextBillion and re-posted with permission.
Is it possible to link digital finance with client protection? Yes, but it’s not an easy task, even if the incentives are clear. The process involves building client trust, improving providers’ reputations, attracting impact investors and mitigating any regulator concerns. The Smart Campaign (hereinafter Smart) and MFR are working to achieve this by adapting client protection standards to digital business models, detailing concrete guidelines to improve practices, and providing a credible way to prove adherence to them (i.e. Smart Certification).
Smart, housed at the Center for Financial Inclusion at Accion, is the self-regulation standard-setting body of reference for client protection in inclusive finance: It has promoted good practices in the industry for over 10 years. Smart maintains a rigorous certification program, elevates the client voice, convenes partners to effect change at the national level, and has led the effort to update industry-facing client protection standards to encompass digital credit models. MFR, an independent rating agency specializing in inclusive finance, has conducted 57 percent of Smart client protection certifications, and is the sole assessor for the GSMA Mobile Money Certification.
Smart and MFR joined forces in 2018 to draft client protection standards applicable to digital credit providers, and pilot tested them in the field with two different companies in Kenya. These companies included 4G Capital, using a hybrid model based on both technology and in-person interaction with clients, and Tala, using a low-touch model where the client interactions mainly happen through a dedicated app. We shared the first findings of the pilot at the European Microfinance Week 2018, organised by the European Microfinance Platform. The draft standards, revised in light of the pilot feedback, are now available for a six-week period of public comment, allowing experts and the industry to weigh in and shape the final version of the standards. The draft standards were defined by Smart and MFR building upon multiple experiences, including Smart’s Fintech Protects Community of Practice, the SPTF Social Investor Working Group, the Responsible Finance Forum, GSMA Mobile Money Programme, GOGLA, Better Than Cash Alliance, ITU, the World Bank, CGAP and others.
VIEWING CLIENT PROTECTION THROUGH A DIGITAL LENS
What changes when one looks at client protection through a digital lens? The objective clearly does not: Keeping clients first remains the ultimate goal, and the motivation for many of us to go to work each day. But other aspects do change. Here are some of the lessons learned from our research, and the client protection practices we recommend, based on these lessons. Please refer to the full list of draft standards for more details.
First of all, the time when a financial service used to be provided by one organization is over. Recognizing the fragmentation of functions among ecosystem players, the draft standards propose holding the DFS provider accountable for the functions which are under its direct or indirect control. When the DFS provider is partnering with an outsourced company (e.g. a call center), we encourage it to manage the external client protection implications. Conversely, when the DFS provider is a user of a service with negligible negotiation power vis-à-vis the external company (e.g. a mobile network operator), this expectation is not realistic. Hence, synergies with other initiatives, such as the GSMA Mobile Money Provider certification, should be explored. As the public comment period continues, industry feedback will be especially crucial for shaping the final standards in light of the complexity of these partnerships.
CLIENT PROTECTION IN DIGITAL LENDING
When it comes to client protection in digital finance, algorithms for loan underwriting cause the most excitement – and the most headaches. If external assessors cannot look inside the algorithm for proprietary reasons, then we encourage them to review its governance (e.g. systems to manage the risk of over-indebtedness and discrimination) and its results (e.g. loan portfolio quality). For instance, one good practice we’ve observed is to establish an underwriting policy that actively and deliberately excludes any variables that may be discriminatory (e.g. gender, race, ethnicity, religion, national origin, sexual orientation, disability, medical history) – even if these factors may correlate with repayment likelihood.
The common trial-and-error approach to product design and algorithm development can be an effective innovation technique, as long as the inevitable experimental mistakes are not paid for by clients. To that end, one positive client protection measure that emerged involved not reporting clients to the credit bureau if they failed to repay loans taken during the underwriting algorithm’s initial stage of development. And when it comes to the question of whether to use client cash-flow analysis or alternative data to estimate the repayment likelihood, we recommend a risk-based approach that uses different methods to assess the borrower’s financial information, depending on their risk of over-indebtedness.
The ease of access enabled by digital finance is exciting. At the same time, immediate access may lead to impulsive, unnecessary debt. Compared to traditional finance, this increases concerns about client harm – and it’s something providers should consider when clients walk in more than when they drop off. That’s why we’ve explicitly incorporated “time to think” and analysis of loan use into the draft standards methodology. One good example of this from the pilots involved giving clients two days after their loan approval to decide if they actually wanted to take the loan.
THE QUESTION OF RISK VS. RETURN IN LOAN PRICING
Solving the risk-return and client value equation becomes even more paramount in digital finance. But though the link between growth, portfolio quality and prices remains valid, what changes is the higher risk appetite – especially in low-touch models – and the associated expectation of higher returns. Moving from high to low-touch models, the predominant cost to run the business generally changes from operating expenses to loan loss expenses. A 2018 CGAP research report in Tanzania calculated loan default rates over 90 days ranging from 35 percent for borrowers taking their first loan to 12 percent for borrowers taking their 7th – 10th subsequent loans – significantly higher than the single digit indicators of portfolio quality (e.g. Portfolio At Risk more than 90 days) in traditional microfinance.
In light of this increased risk, our standards will incorporate benchmarks of portfolio quality and pricing, among others, to digital models as the data become available. However, the standards maintain the notion that poor portfolio quality costs shall not be passed on to clients (who repay) through high prices. We recommend that the pricing of very short-term loans be described as Monthly Percentage Rate – still easily comparable to the Annual Percentage Rate – and that it be compared with the Monthly Percentage Rate of loans of similar amount and tenure. The draft standards explicitly consider price trends and risk-based pricing as evidence of a lending algorithm’s capacity to learn.
OTHER ISSUES TO CONSIDER
Human interaction in financial services may be changing in light of the digital revolution, but it remains important. Live interaction should be available for clients to understand what they are buying and to solve complaints. In-person interaction is key to building trust with the most vulnerable populations, especially during some critical phases of financial service use (Uniting Tech and Touch, ACCION). However, not all digital lenders target the low-income population, and effective live interaction may take the form of in-person assistance, voice or live chat support, depending on what target clients are comfortable with. As the pilots demonstrated, call centers offer the opportunity to monitor the recorded calls and response time, rewarding good customer support with bonuses.
We’ve also strengthened our client protection methodology in the areas of data privacy and security. While it may be too early for the majority of providers to implement the General Data Protection Regulation in full, many of its concepts have been included in the standards in spirit – for example the notion of minimizing and justifying the personal data collected, kept and shared.
As you can see in the above examples, linking digital finance and client protection is not always easy. We still have much to learn, and we will need to constantly interpret the standards in ways that make sense in the variety of cases that exist, whether that involves a provider self-assessing and self-improving, an investor doing due diligence, or Smart certifying a provider. Digital finance is a complex industry, and defining appropriate client protection is an equally complex task. However, we all have the responsibility to contribute to this important process, starting now. We believe that client protection will be a key factor in the ongoing success of digital finance providers – and it remains an essential element of the financial inclusion movement. Let’s join forces in this journey: Share your views on our draft standards during the public comment period, which runs until March 22, and help Smart improve upon the standards based on the insights and real experience of the industry.
by Lucia Spaggiari, Business Development Director at MFR & Isabelle Barrès, Global Director of the Smart Campaign at Accion.
Image courtesy of WorldRemit Comms.