The theme for Financial Inclusion Week (FIW) 2022 was “Inclusive Growth in a Digital Era.” For years, financial products and services have been moving digital, and the pandemic further accelerated the trend toward digitalization. As digital becomes the new norm, we must ensure that these services are safe and inclusive of all people and do not exacerbate the exclusion of marginalized and vulnerable groups. FIW 2022 explored how the global financial inclusion community can harness digital innovation to expand inclusive financial services for all people across the globe, ensuring that products meet customer needs, have appropriate safeguards in place, and contribute to inclusive economic growth.

FIW 2022 hosted 125 live and on-demand sessions, featuring nearly 400 speakers from around the world. With over 3,330 registered participants from 2,000 organizations and 146 countries, FIW 2022 was the largest event to date.

To capture the learnings and key takeaways from the many Financial Inclusion Week 2022 sessions, CFI is compiling short summaries that highlight content and ideas from FIW 2022 sessions on select topics.

FIW 2022 Recap

This report and infographic examine data from the 250+ eligible fintech applicants from the fourth Inclusive Fintech 50 competition to better understand the state of early-stage fintechs. CFI analyzes all applications to identify trends and to find innovations with disruptive potential. The massive scale that is only possible with digitalization matters greatly to the financial inclusion agenda, and this year’s IF50 applicants are already collectively reaching 54 million people, many of whom were previously underserved, across 101 countries of operation. The fintechs are young and small now, but as we’ve seen in past years, they grow to reach millions more people and raise millions more in investment. However, despite this tremendous outreach and potential, there are also areas that require attention: funding and gender gaps, inconsistent practices around data privacy and consent, and limited data collection for outcomes management.

Insights from the 2022 Inclusive Fintech 50 Applicant Pool

Consumer privacy risks have proliferated in the last decade in parallel with the increased adoption of digital technologies and the availability of consumer data. The great onboarding onto digital finance necessitated by COVID-19 also opened up vulnerable consumers to data misuse through scams in markets like the Philippines, Kenya, and India, among many others. Privacy violations can also have serious long-term consequences for financial inclusion and, even as access has improved, trust remains a critical ingredient in translating that progress into sustainable gains for consumers.

Evidence shows that when consumers fall prey to privacy-related harms, the loss of trust dissuades them from the opportunities afforded by digital finance. Privacy implications not only affect consumers but also impact businesses. While compliance with new data protection regulations has brought stronger safeguards, the rapid speed at which consumer data is collected, used, and shared in digital finance has far outpaced consumers’ understanding. For low-income digital financial consumers, the limits of informed consent and choice are even starker.

This report is intended to be an introduction to Privacy by Design (PbD) and to review the main PbD approaches in literature and wider market practice for their potential applications in the inclusive finance sector. The analysis is based on a literature review and key informant interviews with privacy and inclusive finance experts conducted by the Center for Financial Inclusion (CFI) between May and August 2022. The paper groups findings into four approaches to PbD, each of which attempts to embed privacy through different means, often by combining various tools, methods, and areas of application. The paper also identifies a series of gaps and barriers that limit the potential of PbD, particularly in emerging markets and developing economies. The report closes with considerations for the adaption of PbD for fintechs working with low-income and vulnerable consumers. CFI aims to build off this initial report to develop practical guidance on PbD for fintech product teams.

Embedding Trust: The Potential of Privacy by Design for Inclusive Finance

<span data-contrast="auto">“</span><i><span data-contrast="auto">Essentially all models are wrong, but some are useful,</span></i><span data-contrast="auto">” British statistician George E. P. Box declared. Despite progress made in machine learning and artificial intelligence, digital credit models risk being considered “wrong, but useful” – useful for certain groups such as urban salaried workers, but wrong as they leave behind large segments of the economy whose data trails are too faint or complex for a credit scoring algorithm. </span><span data-ccp-props="{&quot;201341983&quot;:0,&quot;335559739&quot;:160,&quot;335559740&quot;:259}"> </span>

<span data-contrast="auto">CFI recently explored how automated decision-making tools for digital credit are playing out for one demographic in India – </span><a href="https://www.habitat.org/sites/default/files/documents/Leaving-the-City-Behind_Rapid-assessment-with-migrant-workers-Maharashtra.pdf" target="_blank" rel="noopener noreferrer"><span data-contrast="none">the more than 54 million</span></a><span data-contrast="auto"> interstate migrants – </span><span data-contrast="none">and what this might teach us about the ways data inputs and algorithms intersect with the digital lives of those at the margins.</span>
<h1><span class="TextRun MacChromeBold SCXW59723147 BCX0" lang="EN-US" xml:lang="EN-US" data-contrast="auto"><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">Digital </span><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">C</span><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">redit in India is </span><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">B</span><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">ooming, </span><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">Y</span><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">et </span><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">L</span><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">eaving </span><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">B</span><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">ehind </span><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">L</span><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">arge (and</span> <span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">P</span><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">ossibly</span> <span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">V</span><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">iable) </span><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">S</span><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">egments of the </span><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">E</span><span class="NormalTextRun SCXW59723147 BCX0" data-ccp-charstyle="Strong">conomy</span></span></h1>
The fintech ecosystem has ballooned in India in recent years, with an estimated 2,000 fintech startups and 300 digital lenders, including Buy-Now-Pay-Later (BNPL) players, operating by the end of 2020. Digital lending alone is expected to reach <a href="https://www.reuters.com/article/india-tech-financing-idUSL5N2NK4XL" target="_blank" rel="noopener noreferrer">$1 trillion by 2023</a>. While digital lenders existed prior to the pandemic, they have boomed over the last two years in the rush to conduct business and daily tasks remotely.

Earlier this year, CFI conducted desk research and interviews across the digital finance ecosystem in India and learned that most digital lenders are focused on serving urban, young, and employed consumers. With higher and more predictable income streams, this customer segment creates large volumes of digital data that can be mined to determine creditworthiness. And given the overall size of the Indian market, there is still room to grow before even this segment is saturated. But where does that leave the rest of consumers, including migrants?

While migrants have propelled the growth of vital industries such as <a href="https://cprindia.org/wp-content/uploads/2022/04/Push-the-policy-needle-forward-on-migrant-support.pdf" target="_blank" rel="noopener noreferrer">construction, hospitality, and commercial agriculture</a>, they historically have faced hurdles in accessing formal financial services and government benefits. And now with the increasing reliance on digital data to allocate benefits and make credit decisions, migrants potentially could be further marginalized.
<blockquote class="pullquote--right">A migrant's data trails do not fit the mold built for the urban, young, employed segment digital lenders were designed to serve.</blockquote>
Take the case of Mohan Kumar, an archetypal migrant whose data trails systematically exclude him from financial services. Mohan works in Surat, India in a garment factory, but is originally from a village near Ganjam, Odisha where his family still lives. Recently married, he visits his wife, who now lives with his parents back in Ganjam, for two weeks twice a year. On his last visit home, Mohan and his wife decided to open a stationary store. When seeking advice from friends, Mohan learned about digital loans available through the phone. Although small, these loans helped a few of his peers with fledgling, entrepreneurial ventures. Despite having a steady, full-time job in Surat for the last four years, Mohan is rejected again and again from numerous digital lenders. Simply put, his data trails do not fit the mold built for the urban, young, employed segment they were designed to serve.

Millions of migrants face similar challenges as Mohan. In general, migrants have fainter data trails due to lower levels of digital activity. Many low-income migrants have limited financial and digital capability and do not possess smartphones. And those who do have a smartphone typically only use mobile data when necessary due to the high cost. Additionally, migrants often rely on agents, such as kirana store owners, to assist them with digital transactions, and these transactions may not be directly associated with their identity.

Furthermore, digital lending algorithms do not understand the data trails of migrants very well. Migrants’ mobility causes multiple barriers. Some migrants change phones and phone numbers between their destination and home, which can impede the verification of transactions or authorizing identity via one-time-passwords (OTP), leading to KYC failures. Additionally, a mismatch between the location of the cell tower where a potential borrower applies for a loan and the location where the phone’s SIM is registered could cause the applicant to be rejected. While providers rationalize that the geographic mismatch might imply that a customer would be hard to reliably reach, unfortunately, this can unintentionally screen out viable borrowers who are inherently on the move, such as migrants.
<blockquote class="pullquote--left">Migrants' payment patterns are not well-understood or formalized, organized, or digitized.</blockquote>
When it comes to assessing cashflows to determine repayment capacity, migrants’ digital data trails make it difficult to see the full picture of their economic lives. For example, even if a migrant has established a stable income over multiple years, as is the case with Mohan, often their payment patterns are not well-understood or formalized, organized, or digitized. For example, some contractors provide an advance payment while others pay weekly or monthly, and some cover living expenses for migrants but deduct the costs from their salaries. In addition, migrants’ data trails might show varied consumption patterns in their two locations – for instance, in Surat, Mohan spends his money on accommodations and food, while back in Odisha he spends on household expenses and children’s education. Such variations in earning and spending tend not to fit within lending models’ range of what makes a viable borrower.

Finally, there is the issue of accessibility. Examining the top 30 downloaded digital credit apps in India (15 on the App Store and 15 on the Google Play Store), we found that 28 out of 30 worked only in English, with just 2 offering Hindi. Such limitations favor highly educated segments and inherently exclude vast numbers of potential clients.
<h1>More Effort is Needed to Ensure Digital Credit Does Not Further Exclude Marginalized Groups</h1>
The Reserve Bank of India (RBI) has taken an interest in digital lending and consumer protection issues over the last two years, asking Google to <a href="https://economictimes.indiatimes.com/tech/technology/google-removes-personal-loan-apps-violating-user-safety-policies-from-play-store/articleshow/80267043.cms?from=mdr" target="_blank" rel="noopener noreferrer">remove at least 30 lending apps from the Play store</a> that did not comply with local laws, <a href="https://www.rbi.org.in/Scripts/BS_PressReleaseDisplay.aspx?prid=52589" target="_blank" rel="noopener noreferrer">convening a working group</a> under the chairmanship of executive director Jayant Kumar Dash, and empowering <a href="https://faceofindia.org/index.html" target="_blank" rel="noopener noreferrer">FACE</a> to become a self-regulatory organization (SRO). The working group focused on a range of consumer protection issues including transparency and disclosure, aggressive sales, consent, and algorithmic design. For instance, one of the group’s recommendations suggests that providers have verifiable audit trails for their algorithms and models should be documented to ensure transparency. While a promising direction, little is understood about what this will look like in practice and how it will balance accountability with the proprietary nature of the models.

Moving forward, digital lenders should find ways to understand and assess the unique activities of groups like migrants through more nuanced models and data inputs that give a more complete picture of a migrant’s economic life. For example, do workers travel back to the same destination each working cycle? How many trips home do they make? How have income patterns shifted within their employment – is it consistent or is it varied?
<blockquote>Digital lenders should find ways to understand and assess the unique activities of groups like migrants through more nuanced models and data inputs that give a more complete picture of a migrant’s economic life.</blockquote>
Providers could draw inspiration from the work of <a href="https://indiamigrationnow.org/" target="_blank" rel="noopener noreferrer">India Migration Now</a> (IMN), which is working to understand the nuances of the data trails of migrants. They are preparing to launch a pilot for a customized, digital lending product for migrants that will leverage both traditional and non-traditional indicators to determine creditworthiness. IMN’s scoring model will leverage data inputs such as frequency of travel, place of origin, destination for migration, frequency of remittance transfer, length of stay at migration location, GPS, and mobile top-up amounts. IMN’s focus on migrants requires commitment, research, time, and resources, which may be less attractive to other digital lenders when there are still potential customers with more robust data trails who are easier to reach.
<h1>What Can Be Done to Maximize the Inclusivity of Algorithms?</h1>
<strong>Research</strong>: We must do more to understand <a href="https://escholarship.org/content/qt3gt9x28k/qt3gt9x28k.pdf?t=qx7nai" target="_blank" rel="noopener noreferrer">whether specific segments that are already at risk for exclusion</a> are further included or excluded by algorithmic decisions. Researchers could help by leveraging context-specific research about specific segments – for instance, migrants or <a href="https://ncst.nic.in/" target="_blank" rel="noopener noreferrer">scheduled tribes</a> in India – to help data scientists better understand the data trails of these groups and incorporate insights into models.

<strong>Guidance and Incentives for Providers</strong>: While advancements in auditing and disparate impact testing can measure models' results across a set of sensitive characteristics (e.g. gender, race, religion, ethnic group), providers may not prioritize the inclusivity of their algorithms. In markets like India, for instance, financial service providers continue to focus on growth among customer segments with more “intuitive” data trails and on keeping customer acquisition costs low. In our view, stakeholders such as impact investors and industry associations can engage with providers to provide guidance and/or incentives for inclusive algorithms. CFI <a href="https://dai-global-digital.com/how-four-awardees-are-paving-the-way-for-a-more-equitable-ai-powered-future.html" target="_blank" rel="noopener noreferrer">recently launched a project</a> to help impact investors engage with fintechs on their algorithms during their due diligence processes.

<strong>Government Engagement</strong>: Because this is a very nascent field, more work is needed to test regulatory approaches and efforts to build the public sector’s capacity for supervision. Government entities, such as the Reserve Bank of India, have begun to think <a href="https://www.centerforfinancialinclusion.org/accountability-for-algorithms-a-balancing-act-for-governments">more strategically about the supervision</a> of high-stakes algorithmic decisions. Ongoing work is needed to understand and advance data protection regulation, national AI strategies, and sector-specific initiatives, such as the RBI working group or Project Veritas in Singapore, to name a few.

Digital Credit Algorithms in India: Expanding Access or Cementing Exclusion?

<span class="NormalTextRun SCXW205969893 BCX0">Platforms </span><span class="NormalTextRun SCXW205969893 BCX0">provide </span><span class="NormalTextRun SCXW205969893 BCX0">critical employment opportunities for many underserved populations and are increasingly interconnected with financial services. To advance financial inclusion and support development goals, </span><span class="NormalTextRun SCXW205969893 BCX0">however, </span><span class="NormalTextRun SCXW205969893 BCX0">platforms must focus on the unique digital, financial, and livelihood needs of micro and small enterprises. This </span><span class="NormalTextRun SCXW205969893 BCX0">publication</span> <span class="NormalTextRun SCXW205969893 BCX0">provides an overview</span><span class="NormalTextRun SCXW205969893 BCX0"> of the increasing interconnectedness of</span><span class="NormalTextRun SCXW205969893 BCX0"> platforms and financial service offerings, how platforms can push the frontiers of inclusive finance, how platforms support women, and what is needed from policymakers for platforms to serve broader development goals. </span><span class="NormalTextRun SCXW205969893 BCX0">The audience of this report includes</span><span class="NormalTextRun SCXW205969893 BCX0"> policymakers, platform leaders, donors, international development actors</span><span class="NormalTextRun SCXW205969893 BCX0">,</span><span class="NormalTextRun SCXW205969893 BCX0"> and anyone interested in the potential for financial services and/or platforms to improve livelihoods.</span>

Platforms, Livelihoods, and Inclusive Finance

Imagine you’re a young mother living on the outskirts of Kampala, Uganda. Due to unexpected medical expenses for your son, you decide to apply for a digital loan. You download a credit app that you heard about from a neighbor, answer a few questions about yourself, consent to have data on your phone shared, and wait 30 minutes. Then, you receive a rejection message: “Your credit limit is 0 UGX.”

You heard that having an active mobile wallet would increase your chances of being approved for a loan and made sure your MTN mobile wallet was active before applying. You’re confused about why your loan application was rejected and wonder how to improve your future chances. Most importantly, you’re still unsure how you will cover your son’s medical expenses.

Understanding the reasons behind credit decisions is increasingly more difficult as algorithms, rather than front-line staff, make and communicate outcomes. <a href="https://centerforfinancialinclusion.org/accountability-for-algorithms-a-balancing-act-for-governments/">Complex algorithms</a>, fueled by alternative data, are driving finance decisions about and for customers -- from underwriting to insurance to <a href="https://centerforfinancialinclusion.org/responsible-social-protection-lessons-from-covid-19-digital-cash-transfers/">eligibility for social protection</a>. As algorithms become more commonplace, we must look for ways to increase transparency around decisions and inform low-income customers about their options for rectification.
<h1>In the Dark on Automated Decisions</h1>
In honor of <a href="https://www.consumersinternational.org/what-we-do/world-consumer-rights-day/" target="_blank" rel="noopener noreferrer">World Consumer Rights Day</a>, CFI is researching how credit decisions are made and communicated to low-income consumers, the rights consumers deserve, and the tools they need to understand – and be able to challenge – those decisions.

In the not-so-distant past, most financial institutions relied on front-line staff to make and communicate decisions to customers. If a customer was rejected for a loan or insurance product, she could ask a loan officer to explain how her history was evaluated and could work with the officer to correct data errors. However, as digital models leveraging algorithms gain traction, the transparency that customers used to rely on from their in-person engagements with loan officers or agents has disappeared.
<blockquote class="pullquote--right">As digital models leveraging algorithms gain traction, the transparency that customers used to rely on from their in-person engagements with loan officers or agents has disappeared.</blockquote>
Because most customers are unaware of how an underwriting algorithm works or what data inputs are used, many customers are also unclear about how a decision is made or whom to ask for more information. This lack of clarity was evident from <a href="https://centerforfinancialinclusion.org/trust-of-data-usage-sources-and-decisioning-perspectives-from-rwandan-mobile-money-users/">CFI’s survey-based research</a> in Rwanda. Out of a sample of 30 digital borrowers in Rwanda, 16 respondents had been rejected for a digital loan but only 10 of those recalled an explanation for the denial. Of those 10 who received an explanation, six were dissatisfied with the provider’s communication; some received only a basic explanation such as, “Your credit limit is zero.”

When the full survey sample was asked to describe what an acceptable explanation of loan denial might look like, there was a clear ask for more specificity and more communication. A 52-year-old woman shared her experience: “They denied me a loan because I delayed to repay when I was in a hospital and very sick. If they had called to explain the reason for the denial, I would have explained my condition.” A 38-year-old woman reasoned: “When you delay to repay, they call to remind you about the loan. They should then do the same when you are denied a loan.”
<h1>Help for Consumers: Legislation, Awareness, Support</h1>
So, what can consumers do in this increasingly digital era? Most data protection frameworks give consumers the right to access and rectify their data. However, our Rwanda research suggests that consumers are largely unaware of what is being inputted in the first place. While data protection efforts – like the 2019 Kenyan Data Protection Act that gives individuals the right to request the rectification of personal data that is “inaccurate, out-of-date, incomplete or misleading” – are a good first step, more needs to be done to communicate to customers about the inputs, decisions, and digital rights.
<blockquote class="pullquote--left">More needs to be done to communicate to customers about the inputs, decisions, and digital rights.</blockquote>
Many nascent data protection frameworks offer consumers the right to be informed if they’ve been subjected to an automated decision by an algorithm, for example. For instance, the <a href="https://www.minijust.gov.rw/fileadmin/user_upload/Minijust/Publications/Official_Gazette/_2021_Official_Gazettes/October/OG_Special_of_15.10.2021_Amakuru_bwite.pdf" target="_blank" rel="noopener noreferrer">Rwandan Data Privacy Law</a> mandates that individuals should be informed about the logic involved in their automated decision at the time of personal data collection. <a href="https://www.dataguidance.com/notes/brazil-data-protection-overview" target="_blank" rel="noopener noreferrer">Brazil’s Data Protection Act</a> gives consumers the ability to ask for a review of a decision made with their personal data taken solely through automated processing; included in the review should be the criteria and procedures used for the decision.

But what this will look like in practice, and how it will be enforced, remains to be seen. And given what we know about the challenges facing low-income consumers in accessing and feeling empowered to use grievance redressal mechanisms, there is concern that these data rights will not be fully exercised in a way that keeps companies accountable.

Consumer advocacy organizations are a potential avenue to help raise the voice of consumers vis-à-vis their digital rights. For instance, <a href="https://centerforfinancialinclusion.org/responsible-social-protection-lessons-from-covid-19-digital-cash-transfers/">our research</a> on government-to-person digital payments during COVID-19 found that civil society organizations often were the first to sound the alarm about digital systems failing or customers poorly served.

Whether it’s a mother in Kenya applying for a credit product to pay for medical expenses, or a farmer looking to insure his crops in the face of an increasingly unpredictable climate, we must support consumers’ voices and increase transparency around how digital decisioning is made.

<em>Along with other leading consumer advocates around the world and in partnership with <a href="https://www.consumersinternational.org/" target="_blank" rel="noopener noreferrer">Consumers International</a>, we’re excited to participate at the <a href="https://www.consumersinternational.org/fair-digital-finance-forum-2022/" target="_blank" rel="noopener noreferrer">Fair Digital Finance Forum</a> this week. Take a look at their <a href="https://www.consumersinternational.org/fair-digital-finance-forum-2022/" target="_blank" rel="noopener noreferrer">program</a> to get a lay of the land and stay up to date with CFI’s <a href="https://www.centerforfinancialinclusion.org/research/consumer-protection">consumer protection</a> workstream.</em>

Understanding the Decision: Consumer Rights and Algorithmic Decision-Making

Consumers all over the world may not know it, but many decisions about and for them are increasingly made by algorithms. Eighty percent of viewing activity on Netflix is driven by recommendations from algorithms; a doctor might recommend a course of treatment based on an algorithm that uses patient DNA; <a href="https://www.poverty-action.org/study/using-mobile-phone-and-satellite-data-target-emergency-cash-transfers-togo" target="_blank" rel="noopener noreferrer">social protection programs</a> might determine eligibility for an assistance payment using an algorithm that crunches your mobile phone transactions; and financial providers can determine your creditworthiness or insurability based on algorithms that leverage alternative data.

A persistent challenge in reaching low-income and vulnerable populations with financial services has been a limited availability of data. In recent years the growth of consumer data trails in emerging markets, largely generated by mobile phones, has helped change the equation alongside powerful new analytical tools. But what data goes into consequential algorithms, how it is used, and whether it leads to inclusion — or bias and exclusion — is largely a black box to consumers, as well as governments. While the expansion in consumer data and companies leveraging artificial intelligence (AI) technology can advance financial inclusion, these technologies introduce new risks and may reinforce historical patterns of bias and exclusion, with potentially harmful consequences for consumers.
<blockquote class="pullquote--right">These technologies introduce new risks and may reinforce historical patterns of bias and exclusion, with potentially harmful consequences for consumers.</blockquote>
Central banks and policymakers have long faced the challenge of balancing financial innovation, market growth, and inclusion alongside competition, stability, and consumer safety. That balancing act has been made more difficult by the speed and scope of financial innovation, technology, and algorithms of increasing sophistication. In emerging markets and developing countries in particular, the increased use of mobile phone data for consequential decisions has raised the profile and potential risks of these tools.

There is a growing recognition that governments require ways to assess consumer risks and opportunities in the deployment of these algorithms and AI systems, as well as methods to amplify the opportunities and mitigate risks — all without quashing innovation. But, how?

Earlier this month, H.M. Queen Máxima of the Netherlands, the United Nations Secretary-General’s Special Advocate for Inclusive Finance for Development (UNSGSA), <a href="https://www.centerforfinancialinclusion.org/unsgsa-and-center-for-financial-inclusion-co-host-virtual-roundtable-for-central-banks-to-tackle-algorithm-biases-in-financial-services">co-hosted a virtual workshop</a> with the Center for Financial Inclusion (CFI) to raise awareness and address algorithm biases for financial inclusion. Queen Máxima and CFI Managing Director Mayada El-Zoghbi co-chaired the meeting, which convened leadership from select central banks and other subject matter experts. The workshop examined how to identify and assess these emerging risks, discussed existing risk mitigation tools, and identified opportunities for further research.

The challenge of bias in financial services is not new and many countries already have financial sector anti-discrimination laws that hold financial institutions to fair lending practices, including just under half of emerging markets. The Equal Credit Opportunity Act in the United States, which was passed in 1974, for instance, makes credit discrimination based on protected classes illegal. But it runs into limits in the face of algorithmic lending as it also <a href="https://www.centerforfinancialinclusion.org/reducing-bias-in-algorithmic-decisions-cannot-rely-on-blind-approaches">restricts lenders</a> from collecting demographic information on borrowers. Given the current methods for testing algorithms, this missing demographic data makes disparate impact testing much harder. While the Equal Credit Opportunity Act and similar antidiscrimination laws could be used to hold companies accountable, they will grapple with understandable privacy concerns.

A second potential avenue for accountability could be through data protection laws, the most well-known being the European Union’s General Data Protection Regulation (GDPR).  While GDPR requires that processors assess any high-risk data system, like a new credit scoring algorithm, it does not amount to an audit. GDPR and the laws it has inspired also enumerate individual rights for consumers including the right to an explanation of an automated decision and the right to have your data rectified by a provider if it is incorrect. But we worry, how likely is it that low-income consumers (or consumers writ-large, really) will be aware enough to be empowered by these rights?  If <a href="https://www.centerforfinancialinclusion.org/trust-of-data-usage-sources-and-decisioning-perspectives-from-rwandan-mobile-money-users">consumers are not aware of how their data is being used</a>, they are less likely to exercise their rights, raise their collective voice, and seek accountability for algorithmic decisions.
<blockquote>If consumers are not aware of how their data is being used, they are less likely to exercise their rights, raise their collective voice, and seek accountability for algorithmic decisions.</blockquote>
In recognition of the limits of existing approaches, some countries have drafted national AI policies or frameworks. These policies work to establish the level of risk for discrete AI systems and how to assign levels of supervision. For instance, the EU introduced a draft law that would require <a href="https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52021PC0206" target="_blank" rel="noopener noreferrer">Algorithmic Impact Assessments</a> for high-risk systems, such as automated credit decisions. Such proposed auditing would require technical expertise within governments and some degree of disclosure of the source code and data. We foresee challenges in implementing such audits in many markets due to gaps in government resources and staffing, as well as challenges in creating standards for auditing that balance supervision with innovation and the proprietary nature of the codes.

One of the ways that governments are engaging is through test and learn partnerships with fintechs. For instance, in the US the Consumer Financial Protection Bureau (CFPB) has granted a <a href="https://www.consumerfinance.gov/about-us/newsroom/cfpb-announces-first-no-action-letter-upstart-network/" target="_blank" rel="noopener noreferrer">no action letter</a> to Upstart Networks, an online lender that uses alternative data. The letter essentially immunizes Upstart from regulatory action and in exchange, Upstart provides information about its loan applications, methodology for decision-making, and whether its loans expand financial inclusion. Other governments have also conducted test and learn engagements with fintechs, as well as create fora for enhanced private-public engagement such as the AI Forum in the Netherlands and Project Veritas in Singapore. Other emerging approaches include working across regulators, such as the Digital Markets Taskforce in the UK.

While there are some strong examples and initial work that could be shared, clearly, there is also still much to learn. What enhancements to existing legal frameworks are necessary to address the new or shifting risks in the use of AI in digital financial services? What digital public goods can help mitigate attendant risks? How can supervisors and regulators build capacity to understand data inputs and use cases, and develop principles and assessment methodologies that are regularly evaluated? These questions would best be answered through research and dialogue among and across regulators, as well as the private sector. Only through a concerted effort will the enormous potential of algorithms and AI extend the reach of financial services in a manner that is inclusive, safe, and fair for customers.

Accountability for Algorithms: A Balancing Act for Governments

The past year was one of building on what became our new reality in 2020.

While the full impact of the pandemic remains unknown, we know that financial services play a crucial role in mitigating the continued effects on livelihoods. We know that financial services are important for changing norms to achieve women’s economic empowerment. We also know that understanding data risks is needed for protecting low-income consumers against fraud and other hazards. And we know that financial services are needed for underserved customers so they can be resilient in the face of a changing climate.

So how have we advanced our learning on these key lines of inquiry in 2021?

The list below – 10 pieces of selected research, analysis, and recommendations from the year – is not exhaustive: CFI published 15 publications and nearly 30 blog posts this year. But it offers a cross-section of how CFI is thinking about the biggest challenges in our sector.

We look forward to furthering our learning in 2022 and continuing to advance inclusive financial services. We wish everyone a healthy new year with shared prosperity for all.
<h3><a href="https://www.centerforfinancialinclusion.org/the-stories-algorithms-tell-bias-and-financial-inclusion-at-the-data-margins">The Stories Algorithms Tell</a></h3>
<em>By Alexandra Rizzi, Alex Kessler, and Jacobo Menajovsky</em>

This flagship publication grounds universal challenges on the use of algorithms, automated decisions, alternative data, and bias in the context of inclusive financial services to present the current state of play among inclusive finance actors. The report is based on desk research and interviews from a sample of fintechs, regulators, and other experts.
<h3><a href="https://www.centerforfinancialinclusion.org/normative-constraints-to-womens-financial-inclusion-what-we-know-and-what-we-need-to-know">Normative Constraints to Women’s Financial Inclusion: What We Know and What We Need to Know</a></h3>
<em>By Julia Arnold, Mayada El-Zoghbi, and Alex Kessler</em>

Financial services design and delivery play a part in ensuring women retain control over resources. But norms are increasingly recognized as a key constraint to women’s financial inclusion. This report reviews evidence on social and gender norms and financial inclusion and identifies areas for further improvement to understand and address the norms that influence women’s participation in the economy and the financial system.
<h3><a href="https://www.centerforfinancialinclusion.org/green-inclusive-finance-a-framework-for-understanding-how-financial-services-can-help-low-income-people-weather-climate-related-risks">Green Inclusive Finance</a></h3>
<em>By Howard Miller</em>

The role of financial services in helping low-income consumers respond to climate change is increasingly gaining acceptance in the inclusive finance realm as a key component of the global response to the great crisis of our time. This video and blog post map out CFI’s framework for green inclusive finance and provide four distinct but interrelated pathways for understanding how inclusive finance can help low-income populations prepare for and respond to climate change while improving green outcomes.
<h3><a href="https://www.centerforfinancialinclusion.org/creating-prosperity-in-an-era-of-uncertainty">Promoting Prosperity in an Era of Uncertainty</a></h3>
<em>By Mayada El-Zoghbi</em>

In her welcome letter to <a href="https://www.centerforfinancialinclusion.org/about/what-we-do/financial-inclusion-week">Financial Inclusion Week</a>, CFI’s Managing Director discusses the biggest trends in and challenges facing inclusive finance amid a pandemic that already has pushed an estimated 97 million more people into poverty. She provides a bird’s-eye view of the measures taken to avert the crisis, noting the bright spots – rapid mobile money adoption, impressive investments in inclusive fintechs, the flowering of the platform and gig economies – while striking a cautious tone regarding the challenges ahead.
<h3><a href="https://www.centerforfinancialinclusion.org/the-precarious-state-of-msmes-understanding-the-impact-of-covid-19-and-opportunities-to-support-their-recovery">The Precarious State of MSMEs: Understanding the Impact of COVID-19 and Opportunities to Support their Recovery</a></h3>
<em>By Shradha Modi, Katia Huayta Zapata, Lauren Braniff, and Aeriel Emig</em>

This brief shares how COVID-19 has impacted MSMEs over the last year and a half in four emerging markets: Colombia, India, Indonesia, and Nigeria. Backed by <a href="https://www.centerforfinancialinclusion.org/research/covid-19-and-msmes-data-and-analysis-to-understand-impact">six waves of survey data</a> collected from owners and operators of micro, small and medium enterprises (MSMEs), this brief offers four areas to focus on to navigate today’s challenges and build resilience for tomorrow.
<h3><a href="https://www.centerforfinancialinclusion.org/weathering-the-storm-ii-tales-of-survival-from-microfinance-crises-past">Weathering the Storm II: Tales of Survival from Microfinance Crises Past</a></h3>
<em>By Daniel Rozas</em>

A follow up to the <a href="https://www.centerforfinancialinclusion.org/weathering-the-storm-hazards-beacons-and-life-rafts-lessons-in-microfinance-crisis-survival-from-those-who-have-been-there">original report</a> and case studies, published in 2011 in the aftermath of the 2007-09 financial crisis, Weathering the Storm II looks at the decade that followed the original paper and shares what became of the organizations profiled within it and how they’re facing the COVID-19 crisis. Nearly every type of financial institution – from NGO to bank – is included in this report that synthesizes experiences and lessons from 16 institutions in 14 countries on four continents over 15 years (including five new case studies from Azerbaijan, Bosnia and Herzegovina, India, Nicaragua, and Palestine).
<h3><a href="https://www.centerforfinancialinclusion.org/yes-covid-is-the-big-bang-of-digital-payments">Yes, COVID is the Big Bang of Digital Payments</a></h3>
<em>By Loretta Michaels</em>

With over 300 million monthly active users in 2020 (up 17 percent year-over-year), mobile money has seen explosive growth. Not only is it an extremely useful tool for businesses and individuals, but mobile money has also been integral for rapid government-to-person (G2P) transfers in many countries in the first months of the pandemic. But there’s still untapped potential for mobile money in a post-pandemic world, especially for governments seeking universal access to financial services.
<h3><a href="https://www.centerforfinancialinclusion.org/trust-of-data-usage-sources-and-decisioning-perspectives-from-rwandan-mobile-money-users">Trust of Data Usage, Sources, and Decisioning: Perspectives from Rwandan Mobile Money Users</a></h3>
<em>By Alex Rizzi and Tanwi Kumari</em>

Who is fairer: loan officers or digital lenders? The publication shares highlights of a qualitative study of 30 mobile money users in Rwanda that aims to answer this question. The survey focuses on user perceptions of and opinions on consumer data in underwriting and the evolving data ecosystem.
<h3><a href="https://www.centerforfinancialinclusion.org/leveraging-digital-financial-capability-to-drive-womens-financial-inclusion">Leveraging Digital Financial Capability to Drive Women’s Financial Inclusion</a></h3>
<em>By Julia Arnold and Marina Dimova</em>

Financial capability must evolve to include digital financial capability. While digital tools offer promise, there is a risk of excluding women, oral communities, and others from the formal financial system, if focus is not given to helping people gain the knowledge and skills to effectively use digital financial services. This post provides a framework for developing effective digital financial capability programs and action steps to guide the way.
<h3><a href="https://www.centerforfinancialinclusion.org/reducing-bias-in-algorithmic-decisions-cannot-rely-on-blind-approaches">Reducing Bias in Algorithmic Decisions Cannot Rely on “Blind” Approaches</a></h3>
<em>By Alex Kessler and Jacobo Menajovsky</em>

What happens when gender-blind credit scoring is applied to financial products and services? We see products like the infamous Apple Card, which gave women consistently lower lines of credit in what researchers call “fairness through unawareness,” an approach that intentionally excludes data on sensitive attributes like race, ethnicity, or sex to reduce discriminatory outcomes. This post addresses the need for a better understanding of algorithms transforming the financial sector to mitigate against discriminatory outcomes while respecting limitations of data gathering needed to protect privacy.

Top 10 of 2021: CFI Research that Furthered Our Thinking in a Pivotal Year

The <a href="https://www.inclusivefintech50.com" target="_blank" rel="noopener noreferrer">Inclusive Fintech 50</a> application data from the third year signals continued progress on innovations to onboard low-income customers, despite the challenges posed by the pandemic:
<ul>
 	<li><strong>Inclusive fintechs are using creative approaches to serve new customers</strong>. From retail partnerships that help drive customer acquisition, to insurance platforms offering micro-coverage, to native language chatbots, to enhanced biometric identity verification solutions, early-stage inclusive fintechs are pursuing outside-the-box solutions to reach un/underbanked people around the world.</li>
</ul>
Nevertheless, there are two areas of concern:
<ul>
 	<li><strong>Funding concentration persists</strong>. While fintech investments across emerging markets have been increasing, early-stage inclusive fintechs continue to face disparities based on geography, funding status, and gender.</li>
 	<li><strong>Potential customer risks persist around data usage</strong>. Although data is central to the business models of inclusive fintechs, some applicants require customer consent related to data, and some still do not.</li>
</ul>

Insights from the 2021 Inclusive Fintech 50 Applicant Pool

H.M. Queen Máxima of the Netherlands, the United Nations Secretary-General’s Special Advocate for Inclusive Finance for Development (UNSGSA), today co-hosted a virtual workshop with the Center for Financial Inclusion (CFI) at Accion to raise awareness and address algorithm biases for financial inclusion.

Queen Máxima and CFI Managing Director Mayada El-Zoghbi co-chaired the meeting, which convened leadership from select central banks and other subject matter experts. Both the UNSGSA and the Managing Director provided brief opening remarks. President and CEO of Accion Michael Schlein also offered brief remarks.

[caption id="attachment_43591" align="alignright" width="559"]<img class="size-full wp-image-43591" src="https://content.centerforfinancialinclusion.org/wp-content/uploads/sites/2/2021/12/unsgsa-cfi-workshop-algorithm-bias-web-story-foto-21.jpg" alt="" width="559" height="322" /> <em>UNSGSA Queen Máxima (top left) is pictured during a virtual workshop for central banks to address algorithm biases for financial inclusion, which the Special Advocate co-hosted with Center for Financial Inclusion Managing Director Mayada El-Zoghbi (top right). Also pictured are Office of the UNSGSA Director Pia Tayag (bottom left) and Accion President and CEO Michael Schlein. Other attendees are not shown, in accordance with Chatham House Rule.</em>[/caption]

Artificial intelligence (AI)-driven digital financial services hold tremendous potential for financial inclusion. For example, the rapid growth in mobile phone usage and increase in new customer data have resulted in novel methods to deliver financial services by leveraging AI and data — especially in emerging markets and economies.

However, these new innovations can be accompanied with new risks or reinforce previous biases and exclusion issues, with potentially harmful consequences for consumers. This is particularly relevant for vulnerable groups like the poor, women, smallholder farmers, and small businesses.

Attendees used the opportunity to exchange knowledge and learn lessons from the experiences of their counterparts from different institutions and countries on these issues. They were encouraged to examine how to more effectively diagnose problems and identify risks related to AI-driven digital financial services.

Participants also examined how governments could better monitor and mitigate the risks associated with using algorithms for credit and insurance products in their own markets. Further, attendees used the time to consider the role of the private sector in finding solutions and ensuring future innovations are safe and fair for consumers.

The workshop was designed to help participants better understand the risks associated with algorithm biases, as well as to create a research agenda to develop possible strategies and solutions to tackle these problems.

The meeting was conducted under Chatham House Rule. For further reading on the topic, please see CFI’s report, <a href="https://www.centerforfinancialinclusion.org/the-stories-algorithms-tell-bias-and-financial-inclusion-at-the-data-margins">"The Stories Algorithms Tell: Bias and Financial Inclusion at the Data Margins."</a>

UNSGSA and Center for Financial Inclusion Co-Host Virtual Roundtable for Central Banks to Tackle Algorithm Biases in Financial Services

Alternative Data

Blockchain and Cryptocurrency

Data Analytics

Alexandra (Alex) Rizzi

Devina Srivastava

Mayada El-Zoghbi

Loretta Michaels

Jayshree Venkatesan

Pia  Roman Tayag

Ezra Mannix

Tanwi Kumari

Daryl  Collins

Derry Moore

Maria Gabriela Coloma Ponce de Leon

Alex Kessler

Jacobo Menajovsky

Sonja Kelly

India

South Asia

Rwanda

Sub-Saharan Africa

Kenya

China

East Asia and Pacific

Entrepreneurs of MSMEs

Financial Technology

Underserved Groups

Credit

Delivery Channels

Digital Financial Services

Financial Service Providers

Fintechs

Insurance

Payments

Remittances

Appropriate Product Design and Delivery

Consumer Protection

Privacy

Refugees and Migrants

Enabling Environment

Financial Inclusion Strategy

Partnerships

Policy, Regulation and Government Initiatives

Transparency

Financial Capability

Women

Challenges and Crises

Human Centered Design

Mobile Banking

Client Treatment

Credit Provision

Digital Credit

Interest Rates

Over-indebtedness

Responsible Digital Finance

Underwriting

Report

Blog Post

Announcement

Brief

India is a global leader in financial inclusion. For over a decade, Accion has worked with and learned from India’s innovators, investors, and policymakers to expand economic opportunity for its people and businesses.

Trust of Data Usage, Sources, and Decisioning: Perspectives From Rwandan Mobile Money Users

<blockquote> “Our data is stored in servers, so government can make strict rules regarding not hampering privacy of citizens. We need to have trustworthy individuals in charge of managing the servers."</blockquote>
&nbsp;
<blockquote>“I felt scared [when looking at the data Google stores] – it wasn’t just that they were collecting my location, they were collecting a pattern of my locations so they knew where I would be at a certain time.”</blockquote>
These quotes illustrate a subtle but important difference in the approach to digital experiences expressed by men and women. The first comes from Prajit, a 57-year-old Indian man, and reflected a dispassionate policy-oriented viewpoint. The second comes from, Alison, an 18-year-old Asian woman in California. Her sense of personal violation is not unusual. From Ohio to Assam, a shared discourse of fear and invasion in digital experiences runs through many women’s digital worldviews.

Thanks to generous sponsorship from PayPal, we have spent the past several months speaking to a small sample of low- to middle-income men and women in the U.S. and India with the intent to probe themes relating to experience, behavior, and attitudes in the digital world, including digital financial services. These results add to a body of work -- such as work on the exposure of negative bias towards women in the algorithms used by digital lenders (see <a href="https://www.womensworldbanking.org/insights-and-impact/report-algorithmic-bias-financial-inclusion-and-gender-primer-opening-up-new-credit-to-women-in-emerging-economies/" target="_blank" rel="noopener noreferrer">here</a> and <a href="https://centerforfinancialinclusion.org/the-stories-algorithms-tell-bias-and-financial-inclusion-at-the-data-margins/">here</a>) -- about how women’s digital data can work against rather than for them.

Our conversations took place in two parts. In the first interview, we mapped out the digital lives of respondents. Through this mapping, we were able to capture the whole of respondents’ “portfolios” of digital tools and services -- including not only digital financial services, but also social media, e-commerce, communications, and any other digital tool or service that might capture and use customer digital data.

In preparing for the second conversation, we informed respondents about how their most-used providers kept -- and used -- specific data. For example, with users of any Google product, we explained -- and, where possible, showed -- how they could see in Google Takeout the locations they had been, their YouTube videos, their search activity, their email, their contacts, and many other pieces of data. We explained that this data helps to create user profiles that suppliers can then target.

We then had an open-ended conversation about what they knew -- or did not know -- and how they felt about this information. What emerged from these conversations was a bifurcation in digital philosophies, a gendered digital worldview perpetrated by and perpetuated in the specific discourses of these men and women, particularly in discussions of digital exposure.
<blockquote class="pullquote--right">What emerged from these conversations was a bifurcation in digital philosophies, a gendered digital worldview.</blockquote>
Men and women in our sample approached the topic of digital exposure -- the extent to which personal, demographic, and behavioral data is accessible or shared with digital service providers -- in very different ways.

In discussions about digital exposure, men overwhelmingly were philosophical. Their concerns lay in abstracted notions of right and wrong, what should be done, and by whom, and tended toward a generalized, high-level discourse.

For example, Collin, a 23-year-old Kenyan migrant living in New York City, spoke in terms of a moral code on the topic of the sale of personal data:
<blockquote class="blockquote">“[For a company to] share your secrets, it's something that it's not in line with accepted code of conduct.”</blockquote>
<h2>Through a Lens of Fear</h2>
Women, however, were more inclined to frame their exposure on a personal level and usually through a lens of fear. For these women, their concerns parallel the challenges and threats they encounter in their physical lives. They tended not to debate about morality or responsibility, or digital exposure as a society-level concern. Rather, they spoke at length -- and in visceral analogies -- about their safety and invasion of their privacy.

For example, Sonam, a 27-year old Indian woman working in banking, expressed personal fear when discussing real-time location tracking:
<blockquote class="blockquote">“You never know who can follow you from your location data, the thought is quite scary.”</blockquote>
Alisha, a 50-year-old woman from Texas, when informed that her phone or device is recording her, drew on lived experience and spoke of violation:
<blockquote class="blockquote">“It reminds me of when my home was broken into… and he rummaged through my bedroom drawers and that was the part that bothered me the most because he went through my underwear drawer… you went somewhere no one goes.”</blockquote>
Worse, women’s comments suggest they are even less sure how to protect themselves from these personal invasions in the digital world compared to invasions in the physical one. Exposure in their digital lives is a mirror to the risks they face in their physical lives – but they have less information about how they navigate them.

In response to threats, many women find it necessary to curtail their behavior and activities in the physical world to avoid feeling or being violated. While the digital world could -- and should -- be different, where women can take full part in all it has to offer without feeling threatened, the reality is starkly contrary. When it comes to physical safety, women provide advice to each other about how to protect themselves in different contexts – do not walk alone at night if you can avoid it; lock your car doors; if you put your drink down, do not pick it up again. These constraints are frustrating and limiting, but at least they are known.

While <a href="https://www.icrw.org/publications/technology-facilitated-gender-based-violence-what-is-it-and-how-do-we-measure-it/" target="_blank" rel="noopener noreferrer">technology-facilitated gender-based violence</a>, such as cyberbullying and online harassment, has been acknowledged and discussed, the difference here is that these women are not reacting to a stranger harassing them online. This is a sense of violation from a service they need to use to engage in their personal and business lives. The strategies for protecting oneself against this kind of invasion are nebulous at best. Indeed, the threats themselves are often unknown and obscured. And this lack of knowledge is perceived as a violation among the women we interviewed.

Danielle, for example, analogizes this lack of knowledge to date rape:
<blockquote class="blockquote">“If you’re a woman and someone drugs you and rapes you and you wake up and you don’t know that you’ve been drugged and raped. Then you see the video and you realize that this has happened, but before you see the video you did not even know you had been violated. So, I feel there’s some aspect of this that feels off about it — like I’ve been raped, and I don’t even know that I’ve been raped so I feel not only violated but tricked into not knowing.”</blockquote>
Through these conversations, a gendered narrative emerges that belies a split between men's and women’s digital worldviews. Worldviews that influence and determine their digital behaviors and lives. These interviews <a href="https://dalberg.com/our-ideas/privacy-line/" target="_blank" rel="noopener noreferrer">and others</a> suggest a common desire from many women who say they have or would like to curtail their activities on specific digital products because they are concerned about safety. But opting out of digital tools is at an end of a behavioral spectrum. Simply limiting the use of products creates a digital divide of its own.

Governments and digital service providers alike benefit from an increasingly digital population. For example, government-to-person transfer programs in developing countries are increasingly dependent on digital technologies to deliver services and provide an interface to the public. Digital service providers benefit from a large and growing consumer market. However, both governments and private providers bear increased responsibility in meeting expectations of digital safety for these benefits to be realized.

Both will fail if governments and providers do not better understand the nuances of consumer concerns. This is not currently being done. Eliciting concerns about data security and privacy must consider how experiences are gendered if we want to build – and maintain – trust among women.

Going forward, we will seek to further investigate the nature of women’s digital worldviews and their effects on digital behavior, especially in the realm of digital financial services. Specifically, we look to understand specific barriers women face in using digital services, investigate perceptions of the value of these services, and evaluate the design and function of digital services through a gendered lens.

Gender and Digital Worldviews: Divergent User Perspectives On Data Collection and Use

The majority of Kenyans have mobile money accounts; their activity demonstrates digital financial services’ potential. Accion helps Kenyan innovators harness emerging technologies to create new financial services.

In April 2021, a <a href="http://kenyalaw.org/kl/fileadmin/pdfdownloads/bills/2021/TheCentralBankofKenya_Amendment_Bill_2021.pdf" target="_blank" rel="noopener noreferrer">bill</a> that proposed licensing digital lenders was passed by the Kenyan parliament and is now pending approval from the national assembly. Among other things, this bill would require all digital lenders to submit product pricing to the Central Bank of Kenya for approval before launching, a move that seeks to lower predatory interest rates. Further, non-performing loans would be capped at <a href="https://www.businessdailyafrica.com/bd/economy/mobile-phone-lenders-risk-closure-in-cbk-bill-3376538" target="_blank" rel="noopener noreferrer">twice the amount of the initial loan defaulted</a>, which means lenders won’t be able to perpetually increase the amount to be paid back through late-payment fees.

Much has been <a href="https://www.bloomberg.com/news/features/2020-02-12/tech-startups-are-flooding-kenya-with-apps-offering-high-interest-loans" target="_blank" rel="noopener noreferrer">written</a> on the fintechs that have flooded the Kenyan market since 2016. FSD and CGAP published alarming <a href="https://www.cgap.org/blog/kenyas-digital-credit-revolution-five-years">evidence</a> in 2018 showing that in addition to a saturated digital lending market, an increasing number of digital borrowers in Kenya became over-indebted, which furthered them into <a href="https://www.microsave.net/2019/09/18/11310/" target="_blank" rel="noopener noreferrer">poverty traps</a>. Digital lenders such as OKash and Branch reportedly publicly shamed debtors by <a href="https://restofworld.org/2020/okash-microlending-public-shaming/" target="_blank" rel="noopener noreferrer">texting their friends, family, and colleagues</a>. Further, high-interest rates and non-transparent fees resulted in APR’s above <a href="https://qz.com/africa/1975202/kenya-prepares-to-crack-down-on-high-interest-loan-apps/" target="_blank" rel="noopener noreferrer">400 percent</a>, making it more difficult for some low-income borrowers to repay. In some instances, repayment was possible, yet it meant sacrificing other payments like food, medicine, or school fees. And finally, millions of borrowers were reported to credit bureaus for <a href="https://www.cgap.org/research/publication/consumer-protection-digital-credit" target="_blank" rel="noopener noreferrer">loans smaller than $10</a>.

During the COVID crisis, the Kenyan Central Bank issued a number of measures to ease the debt burden on borrowers, such as <a href="https://www.centralbank.go.ke/uploads/press_releases/850440997_Press%20Release%20-%20Credit%20Reference%20Bureau%20Regulations%20-%20April%202020.pdf" target="_blank" rel="noopener noreferrer">forbidding</a> unregulated mobile-based and credit-only digital lenders from reporting any late borrowers to credit bureaus. Although this helped ease debt stress on consumers during the crisis, this drove many digital lenders out of the market as levels of default skyrocketed, leaving <a href="https://www.theelephant.info/data-stories/2020/09/18/covid-19-regulatory-measures-could-widen-kenyas-financial-access-gap/" target="_blank" rel="noopener noreferrer">many Kenyan borrowers underserved</a>. The entire loan market in Kenya shrank due to COVID-19; today, these platforms lend around <a href="https://www.businessdailyafrica.com/bd/opinion-analysis/editorials/cbk-regulation-of-digital-lenders-is-long-overdue--3378386" target="_blank" rel="noopener noreferrer">2 billion shillings a month</a>, compared to 4 billion shillings in early 2020.

The proposed bill by the Central Bank of Kenya unfortunately will not address the issue of making digital credit both safe and accessible. The bill focuses on product pricing only, and it doesn’t address the underlying challenge of digital lending platforms: a weak credit assessment model. Such a model <a href="https://www.ted.com/talks/shivani_siroya_a_smart_loan_for_people_with_no_credit_history_yet?language=en" target="_blank" rel="noopener noreferrer">extracts large amounts of data from the borrower’s phone</a> and runs it through algorithms that approve or deny the loan. Many digital lenders mitigate the risk of a light-touch credit assessment with above-market APRs and hidden late payment fees, thus being able to generate profit and returns to their investors.

The primary problem is digital lenders do not measure the capacity of someone to repay, but their willingness to repay, and these are drastically different concepts. Willingness refers to the genuine disposition of a person to pay back their loan. Capacity, on the other hand, refers to the person’s ability to generate enough money in the future to pay back the loan without having to forgo essential expenses.
<blockquote class="pullquote--left">Digital lenders do not measure the capacity of someone to repay, but their willingness to repay and these are drastically different concepts.</blockquote>
For example, I might score very well on a digital risk assessment. My geo-trace shows I go to work every day, my call log shows I call my mom regularly, and my browsing history shows I do not engage in sports bets. These measurements indicate I am a responsible citizen who complies with her duties, suggesting that I have a high willingness to repay a loan. However, I might have astronomical medical expenses that the digital lender is unable to detect. Or I might be sending significant remittances to my family. Both of these relate to my repayment capacity and are measures that digital lending apps tend to ignore in favor of automated decision-making based on more readily available data.

By comparison, traditional banks run thorough credit risk assessments, sometimes so thorough that they are slow and expensive. These banks have credit committees and risk analysis departments that challenge relationship managers on the loan proposals they present. The result is that most people granted a loan have the means to pay it back. This method excludes the majority of the world’s population from the financial system, however, it also ensures that most people who are granted a consumer loan actually have the capacity to repay.

So where do we draw the line? How do we expand access to credit to low-income people and informal workers and also protect them from default? How can we measure their capacity to repay in a cheap, digital, and rapid manner? Should we prohibit alternative, digital, phone-based metrics that only assess someone’s willingness to repay?
<h2>A Possible Solution</h2>
The answer could be a measure in between. One idea is that the Central Bank of Kenya could only grant a license to digital lenders that incorporate the borrower’s capacity to repay in their credit assessment algorithms. This means, finding a way to measure, prove and integrate the person’s real income and expenses. Another option could be asking lenders to integrate proxies for repayment capacity. For example, mobile money records could be matched with the borrower's geolocation to determine where and why they are being paid and predict how stable his or her income is. Coupled with strong data protection measures, this and similar metrics could help curb the high default rate and loan-stacking among digital borrowers in Kenya without causing financial exclusion.

Further, as suggested in a <a href="https://centerforfinancialinclusion.org/making-digital-credit-truly-responsible/" target="_blank" rel="noopener noreferrer">publication produced by CFI</a>, the Central Bank of Kenya should withdraw a digital lender’s license upon surpassing a threshold percentage of non-performing loans as a share of the total loan portfolio. This would force the lenders to improve their algorithms and creatively design more capacity-measuring metrics. And finally, the CBK should explicitly require licensed digital lenders to cross-check an individual’s loan performance with each other before issuing further credit to prevent loan-stacking. Although the Digital Lenders Association of Kenya (DLAK) advocates for digital lenders to be able to receive and submit borrower information to credit bureaus, this practice has not addressed loan stacking in the past and only served to blacklist millions of borrowers for small amounts. Therefore, digital lenders should be allowed to view and report outstanding loans, yet with very limited power to blacklist a late borrower.
<blockquote class="blockquote">The CBK should explicitly require licensed digital lenders to cross-check an individual’s loan performance with each other before issuing further credit to prevent loan-stacking.</blockquote>
Kenya cannot go back to the old bank-driven model which financially excluded most of the population, yet Kenyans cannot keep paying the price for the financial experimentation of tech startups. Digital credit should be responsible, affordable, and inclusive. With this new bill, the Central Bank of Kenya is getting closer to achieving it. However, the root cause of the problem is the evaluation algorithm and the business model that stems from it. Without addressing this, little will change for Kenyan borrowers.

Digital Lending in Kenya: Willingness vs. Capacity to Repay

When Goldman Sachs came under fire for offering women lower lines of credit than men on the now <a href="https://www.nytimes.com/2019/11/10/business/Apple-credit-card-investigation.html" target="_blank" rel="noopener noreferrer">infamous Apple Card</a>, they defended the product against charges of sexism by explaining that gender had not been an input to the credit scoring model. If an applicant’s gender was never fed into the algorithm, how did credit decisions end up discriminating against women who had similar credit histories as men?

Gender-blind credit scoring is an example of what many AI bias researchers call “fairness through unawareness,” an approach that intentionally excludes data on sensitive attributes like race, ethnicity, religion, or sex with the goal of reducing discriminatory outcomes. Behind this approach is the idea that eliminating markers of different group identities will prevent basing decisions on those characteristics. This has led to failures in the past, notably <a href="https://cpb-us-e1.wpmucdn.com/sites.suffolk.edu/dist/3/1172/files/2014/01/Rice-Swesnik_Lead.pdf" target="_blank" rel="noopener noreferrer">redlining</a>. Recent research has shown it can also make <a href="https://arxiv.org/abs/2009.06251" target="_blank" rel="noopener noreferrer">discrimination harder to detect and address</a> in the era of algorithmic decisions.
<blockquote class="pullquote--right">“Fairness through unawareness” intentionally excludes data on sensitive attributes like race, ethnicity, religion, or sex with the goal of reducing discriminatory outcomes.</blockquote>
As algorithms <a href="https://www.centerforfinancialinclusion.org/the-stories-algorithms-tell-bias-and-financial-inclusion-at-the-data-margins" target="_blank" rel="noopener noreferrer">transform the financial sector</a>, we need a better understanding of how to mitigate against discriminatory outcomes to ensure that those left out of the digital ecosystem today do not face further financial exclusion. Otherwise, we may not know how many Apple Card equivalents exist in inclusive finance.

This post explores the challenges and tradeoffs of using sensitive attribute data for detecting discrimination in the context of evolving data protection norms and regulations.
<h2>The Perils of Fairness through Unawareness</h2>
The idea may seem counterintuitive: if you remove gender from the data inputs of a credit model, how does the algorithm consistently assign a lower limit to women? The reason is that credit scoring algorithms are crunching larger amounts and new sources of data, which means that the model can inadvertently use proxies for sensitive attribute data in decisions, even if that was not the intention of the model developers. For instance, the <a href="https://journalofbigdata.springeropen.com/articles/10.1186/s40537-019-0180-9" target="_blank" rel="noopener noreferrer">type of phone</a> someone uses or the <a href="https://dl.acm.org/doi/10.1145/2721896.2721908" target="_blank" rel="noopener noreferrer">apps a person installs on their phone</a> can act as proxies for gender, age, or other characteristics.
<blockquote class="pullquote--left">Models can inadvertently use proxies for sensitive attribute data in decisions, even if that was not the intention.</blockquote>
A very simple method for testing algorithmic bias within a credit scoring model requires only a few variables like the actual default rate, the actual score of the individual when approved, and the variable you want to test with your algorithm, such as gender or race.
<h2>Data Tensions: Privacy vs. Anti-Discrimination</h2>
So, why isn’t this done? The answer lies in data limitations, sometimes based on very legitimate concerns about histories of injustice. It illustrates one of the challenges of detecting and monitoring for bias in algorithmic systems without infringing on privacy rights.

Many countries around the world prohibit collecting sensitive data or basing credit decisions on protected categories either under their banking regulations or under new data privacy legislation. In India, lenders are prohibited from discriminating on the <a href="https://www.rbi.org.in/Scripts/PublicationReportDetails.aspx?ID=764" target="_blank" rel="noopener noreferrer">grounds of sex, caste, or religion</a>. In the U.S., the Equal Credit Opportunity Act (ECOA) made it illegal for credit providers to discriminate based on sex, race, national origin, and other protected characteristics in response to widespread denials of credit to women and people of color. ECOA prohibits lenders from collecting this data in the first place, even for the purpose of detecting discrimination, because of fears it will be used improperly to further perpetuate bias in consumer lending. Research in the US shows that discriminatory trends have been <a href="https://arxiv.org/pdf/1912.06171.pdf" target="_blank" rel="noopener noreferrer">hard to detect in consumer lending</a> and that a lack of sensitive attribute data has also created enforcement challenges.

While not all lending regulations prohibit the collection of sensitive data, data privacy legislation introduces new limits on this data. The General Data Protection Regulation (GDPR), which has become the model for legislation in many developing markets, bars certain sensitive data from being collected or processed -- including data on race and ethnic origin, political opinions, philosophical beliefs, sexual orientation, and biometrics -- without individual consent.
<blockquote class="pullquote--right">Concerns about safely and properly using sensitive attribute data are valid given histories of discrimination and surveillance of minority groups in both wealthy and developing countries.</blockquote>
On top of GDPR, the European Commission has recently put forward <a href="https://digital-strategy.ec.europa.eu/en/library/proposal-regulation-laying-down-harmonised-rules-artificial-intelligence" target="_blank" rel="noopener noreferrer">draft legislation to regulate artificial intelligence</a>, with specific provisions for AI systems classified as high risk, which includes credit scoring. The draft legislation suggests that AI providers should process “special categories of personal data” to monitor and correct for bias in the public interest, but does not clarify collection practices or consent processes for such data. Concerns about safely and properly using sensitive attribute data are valid given histories of discrimination and surveillance of minority groups in both wealthy and developing countries. Changes in regulatory frameworks to allow for use of this data to mitigate against further discrimination require public trust bolstered by technical solutions that protect against misuse. Depending on a country’s history or in places experiencing active or simmering conflict, technical solutions may not be enough. For example, in a country like Rwanda where <a href="https://www.nytimes.com/2004/04/09/world/a-decade-after-massacres-rwanda-outlaws-ethnicity.html" target="_blank" rel="noopener noreferrer">ethnic distinctions are banned</a>, the tensions around sensitive identities make it difficult to fathom collecting and using such information, especially by the government.
<h2>Data Ecosystems in Emerging Markets: It’s Complicated</h2>
This tension between privacy and anti-discrimination may be exacerbated in developing markets by the <a href="https://centerforfinancialinclusion.org/rethinking-financial-inclusion-at-the-data-margins/" target="_blank" rel="noopener noreferrer">limited digital footprints</a> of customers at the data margins, as well as paltry knowledge on how data trails and protected groups intersect.

Sex-disaggregated data is still <a href="https://data2x.org/wp-content/uploads/2020/03/MappingGenderDataGaps_Economic.pdf" target="_blank" rel="noopener noreferrer">largely unavailable</a> in the financial sector and for many economic indicators in developing markets. Beyond availability, there is less understanding of how women’s lives and the barriers that shape their economic behaviors are reflected in data used for credit decisions. The reality is that gender, ethnic, or religious identities interact with other indicators (like income, asset accumulation, or technology use) and will impact individuals’ data trails in ways that may misrepresent individuals’ potential and result in disparate outcomes (i.e. access to a financial product). Fairness through unawareness approaches are blind to this intersectionality.
<h2>Where Do We Go From Here?</h2>
As an industry, we can do more to understand the data inputs for digital financial services and how socioeconomic realities present in these datasets, especially for traditionally excluded groups. There will still be tradeoffs between privacy and anti-discrimination, but managing those tradeoffs requires a conversation between providers, regulators, donors, and the public to identify what is possible and explore the right balance.

These conversations can start with the many open questions that exist from a technical perspective. For instance, can <a href="https://ieeexplore.ieee.org/document/8970908" target="_blank" rel="noopener noreferrer">third parties</a> play a role in securely storing sensitive data, or are there options to use <a href="http://proceedings.mlr.press/v80/kilbertus18a/kilbertus18a.pdf" target="_blank" rel="noopener noreferrer">encryption</a> or <a href="https://spectrum.ieee.org/artificial-intelligence/machine-learning/engineering-bias-out-of-ai" target="_blank" rel="noopener noreferrer">synthetic data</a> techniques? Can regulators develop tools that allow them to estimate the demographic characteristics of a dataset, like the <a href="https://www.consumerfinance.gov/data-research/research-reports/using-publicly-available-information-to-proxy-for-unidentified-race-and-ethnicity/" target="_blank" rel="noopener noreferrer">Consumer Financial Protection Bureau has done</a> in the U.S.? Can investors encourage investees to design monitoring and evaluation indicators to help answer questions about how their algorithms perform on inclusion?

While the road ahead may be complex, starting these conversations and bringing the public into them, especially those whose lives will be affected by algorithmic decisions, will be crucial for building trust. We need to engage with emerging research on possible solutions to see what might be appropriate in developing markets. At the very least, financial inclusion stakeholders should know enough to respond to any gender-blind or fairness through unawareness approaches in digital finance with healthy skepticism.
<blockquote>Starting these conversations and bringing the public into them, especially those whose lives will be affected by algorithmic decisions, will be crucial for building trust.</blockquote>

Reducing Bias in Algorithmic Decisions Cannot Rely on “Blind” Approaches

In 2020, CFI and Women’s World Banking began workstreams on algorithmic bias. This conversation between Alex Rizzi (CFI) and Sonja Kelly (Women’s World Banking) builds on recent reports by <a href="https://centerforfinancialinclusion.org/the-stories-algorithms-tell-bias-and-financial-inclusion-at-the-data-margins/" target="_blank" rel="noopener noreferrer">CFI</a> and <a href="https://www.womensworldbanking.org/insights-and-impact/report-algorithmic-bias-financial-inclusion-and-gender-primer-opening-up-new-credit-to-women-in-emerging-economies/" target="_blank" rel="noopener noreferrer">Women’s World Banking</a>. Below are their answers to some important questions about the future of AI and machine learning as they apply to financial inclusion.
<h3>Where are we now in the conversation on algorithmic bias? Is it too late to address this?</h3>
<strong>Alex</strong>: It’s just the right time! While it may feel like global conversations around responsible tech have been going on for years, they haven’t been grounded squarely in our field. For instance, there hasn’t been widespread testing of de-biasing tools in inclusive finance (though Sonja, we’re excited to hear about the results of your upcoming work on that front!) or mechanisms akin to credit guarantees to incentivize digital lenders to expand the pool of applicants their algorithms deem creditworthy. At the same time, there are a bunch of data protection frameworks being passed in emerging markets that are modeled on the European GDPR and give consumers data rights related to automated decisions, for example. These frameworks are very new, and it’s still unclear whether and how they might bring more algorithmic accountability. So it’s absolutely not too late to address this issue.

<strong>Sonja</strong>: I completely agree that now is the time, Alex. Just a few weeks ago we saw a <a href="https://www.federalreserve.gov/newsevents/pressreleases/bcreg20210329a.htm" target="_blank" rel="noopener noreferrer">request for information</a> here in the U.S. for how financial service providers use artificial intelligence and machine learning. It’s clear there is an interest on the policymaking and regulatory side to better understand and address the challenges posed by these technologies, which makes it an ideal time for financial service providers to be proactive about guardrails to keep bias from algorithms. I also think that technology enables us to do much more about the issue of bias -- we can actually turn algorithms around to audit and mitigate bias with very low effort. We now have both the motivation and the tools to be able to address this issue in a big way.
<h3>What are some trends leading us down the wrong path when it comes to algorithmic bias?</h3>
<strong>Sonja</strong>: At the risk of being too broad, I think the biggest trend is lack of awareness. Like I said before, fixing algorithmic bias doesn’t have to be hard, but it does require everyone -- at all levels and within all responsibilities -- to understand and track progress on mitigating bias. The biggest red flag I saw in our interviews contributing to our report was when an executive said that bias isn’t an issue in their organization. My co-author Mehrdad Mirpourian and I found that bias is always an issue. It emerges from biased or unbalanced data, the code of the algorithm itself, or the final decision on who gets credit and who does not. No company can meet all definitions of fairness for all groups simultaneously. Admitting the possibility of bias costs nothing, and fixing it is not that difficult. Somehow it slips off the agenda, and that is very discouraging to me.
<blockquote class="pullquote--left">Bias is always an issue</blockquote>
<strong>Alex</strong>: One of the concepts we’ve been thinking a lot about is the idea of how digital data trails may reflect or further encode existing societal inequities. For instance, we know that women are <a href="https://www.gsma.com/r/gender-gap/" target="_blank" rel="noopener noreferrer">less likely to own phones than men</a> and less likely to use mobile internet or certain apps; these differences create disparate data trails, and might not tell a provider the full story about a woman’s economic potential. And what about the <a href="https://centerforfinancialinclusion.org/rethinking-financial-inclusion-at-the-data-margins/" target="_blank" rel="noopener noreferrer">myriad of other marginalized groups</a>, whose disparate data trails are not clearly articulated?
<h3>Who else needs to be here in this conversation as we move forward?</h3>
<strong>Alex</strong>: For my colleague Alex Kessler and me, a huge takeaway from the exploratory work was that there are plenty of entry points to these conversations for non-data scientists, and it’s crucial for a range of voices to be at the table. We originally had this notion that we needed to be fluent in the code-creation and machine learning models to contribute, but the conversations should be interdisciplinary and should reflect a strong understanding of the contexts in which these algorithms are deployed.
<blockquote class="pullquote--right">There are plenty of entry points to these conversations for non-data scientists.</blockquote>
<strong>Sonja</strong>: I love that. It’s exactly right. I would also like to see more media attention on this issue. We know from other industries that we can increase innovation by peer learning. If sharing both the promise and pitfalls of AI and machine learning becomes normal, we can learn from it. Media attention would help us get there.
<h3>What are the immediate next steps here? What are you focused on changing tomorrow?</h3>
<strong>Sonja</strong>: When I share our report with external audiences, I first hear shock and concern about the very idea of using machines to make predictions about people’s repayment behavior. But our technology-enabled future doesn’t have to look like a dystopian sci-fi novel. Technology can increase financial inclusion when deployed well. Our next step should be to start piloting and proof-testing approaches to mitigating algorithmic bias. Women’s World Banking is doing this over the next couple of years in <a href="https://www.womensworldbanking.org/insights-and-impact/awardees-announced-as-part-of-data-org-inclusive-growth-and-recovery-challenge/" target="_blank" rel="noopener noreferrer">partnership with the University of Zurich and data.org</a> with a number of our network members, and we’ll share our insights as we go along. Assembling some basic resources and proving what works will get us closer to fairness.
<blockquote class="pullquote--left">Our technology-enabled future doesn’t have to look like a dystopian sci-fi novel.</blockquote>
<strong>Alex</strong>: These are early days. We don’t expect there to be universal alignment on de-biasing tools anytime soon, or best practices available on how to enforce data protection frameworks in emerging markets. Right now, it’s important to simply get this issue on the radar of those who are in a position to influence and engage with providers, regulators, and investors. Only with that awareness can we start to advance good practice, peer exchange, and capacity building.

<em>Follow CFI (<a href="https://centerforfinancialinclusion.us5.list-manage.com/subscribe/post?u=394352ae963270aef6c51680d&amp;id=9df382106a">newsletter</a>, <a href="https://www.linkedin.com/showcase/center-for-financial-inclusion-at-accion" target="_blank" rel="noopener noreferrer">LinkedIn</a>, and <a href="https://twitter.com/CFI_Accion" target="_blank" rel="noopener noreferrer">Twitter</a>) and Women’s World Banking over the coming months to stay up-to-date on algorithm bias and financial inclusion.</em>

Help Wanted: A Global Push Toward Algorithmic Fairness

China’s economy has experienced tremendous growth in recent years, with a booming micro, small, and medium enterprise (MSME) market. There are over 50 million MSMEs in China that collectively account for over 60% of China’s GDP and over 50% of the government’s tax income. These businesses create 80% of total employment in China’s urban areas.

Despite China’s overall economic growth, the need to foster full financial inclusion is still imperative. In 2010, 38 million MSMEs lacked access to formal financial services, forcing them to borrow informally from friends, family, and moneylenders.

Headlines late last year breathlessly chronicled the <a href="https://www.theguardian.com/business/2021/jan/23/the-strange-case-of-alibabas-jack-ma-and-his-three-month-vanishing-act" target="_blank" rel="noopener noreferrer">whereabouts</a> of former Alibaba CEO Jack Ma and tensions between Ma and the Chinese government. The missing billionaire, the last-minute <a href="https://www.nytimes.com/2020/11/06/technology/china-ant-group-ipo.html" target="_blank" rel="noopener noreferrer">cancellation of a history-making IPO</a>, and the <a href="https://entrepreneurshandbook.co/the-speech-that-cost-jack-ma-37-billion-dollars-5ad42393c9ba" target="_blank" rel="noopener noreferrer">speech</a> that apparently started this tailspin of events captured the attention of the world. But beyond the headlines, this story is part of a larger signal that China is sending to big tech: put on your helmets, because we are getting off the sidelines.
<blockquote class="pullquote--left">China to big tech: put on your helmets, because we are getting off the sidelines.</blockquote>
In the last four months, China has published a slew of regulations (some still in draft form) aimed at big tech and their outsized role in the economy. This includes the draft <a href="https://www.natlawreview.com/article/china-releases-draft-personal-information-protection-law" target="_blank" rel="noopener noreferrer">Personal Information Protection Law</a> (PIPL), new rules including capital requirements and caps for online lenders, as well as draft anti-trust rules for digital platforms. And, dramatically, China called off Ant Group’s globally anticipated IPO in November, in December launched an anti-trust probe into Ant’s ecommerce parent Alibaba, and in February 2021 told Ant to restructure itself into a financial holding company.

The scale of big tech’s reach into China’s financial sector is worth reiterating. A billion <a href="https://www.cgap.org/research/publication/china-digital-payments-revolution" target="_blank" rel="noopener noreferrer">Chinese consumers</a> use Alipay, owned by Ant, to pay for everything from groceries (<a href="https://progressivegrocer.com/major-asian-supermarket-chain-adopts-alipay-mobile-wallet" target="_blank" rel="noopener noreferrer">even in the U.S.</a>) and utilities, to school tuition and traffic tickets. Alipay, and its competitor WeChat Pay (owned by TenCent), have deftly created business models that leverage increasing use of smartphones, allow for easy funding of digital wallets through linked bank accounts, and coax users into their digital ecosystem through low transaction fees and an ever-increasing array of services. Cross-selling through this ecosystem can be lucrative. For instance, Alibaba’s product Yu’E Bao rapidly became the largest money market fund in the world for a time.

More recently, Ant has received attention for facilitating hundreds of millions of online loans to consumers, many of whom could not qualify for bank-issued credit cards. Despite the mammoth amount of outstanding debt, Ant has taken on limited risk and acted as more of a facilitator than a lender. For instance, at the end of June 2020, Ant had approximately $267 billion in outstanding consumer debt, but just 2 percent of that amount was funded by Ant.

Whose portfolios are on the hook? Ant has partnered with more than 100 commercial lenders and regional banks, collecting fees on interest income from them, and utilized colossal troves of transactional Alipay data to facilitate underwriting. This frenzy of unsecured lending has raised concerns around the concentration of risk that sits with lenders and not facilitating platforms.
<blockquote class="blockquote">Ant has utilized colossal troves of transactional Alipay data to facilitate underwriting.</blockquote>
<h2>New Rules of the Road</h2>
New regulation from China’s Banking and Insurance Regulatory Commission clamps down on both sides of the online lending partnerships. For the lenders, regional banks are no longer permitted to lend outside their geographies, and there will be new caps on how much of a bank’s capital and portfolio can be lent online. And for the platforms, starting in 2022, they will have to fund <a href="https://www.reuters.com/article/us-internet-loans-china-banks/china-regulator-finalizes-guidelines-on-banks-internet-loan-businesses-idUSKBN2AK0F2" target="_blank" rel="noopener noreferrer">at least 30 percent</a> of every loan they make in concert with banks. These measures may take some needed steam out of the ballooning online lending market.

Digital platforms are also the subject of new draft <a href="https://www.reuters.com/article/us-china-internet-anti-monopoly/china-issues-new-anti-monopoly-rules-targeting-its-tech-giants-idUSKBN2A70DT" target="_blank" rel="noopener noreferrer">anti-monopoly rules</a> by the State Administration for Market Regulation, which targets practices such as the use of data and algorithms to offer different prices to different consumers, as well as self-dealing through elevating their in-house products compared to competitors who are selling on their platform. As they take comments on proposed revisions, antitrust authorities have increased the number of investigations and in recent months fined Alibaba, Tencent Holdings as well as smaller platforms like Pinduoduo and Meituan.

The PIPL, when approved, will be the country’s first comprehensive data protection regime applicable to all data processors, including big tech. The draft PIPL contains <a href="https://www.dataguidance.com/opinion/china-draft-pipl-and-gdpr-comparative-perspective" target="_blank" rel="noopener noreferrer">some similarities to the European GDPR</a>, including provisions for the legal basis for processing data and enumerating data rights for consumers such as the right to rectify data mistakes. PIPL diverges from GDPR in terms of a larger focus on data localization, however.

Taken in aggregate, the package of emerging regulations demonstrates resolve among Chinese authorities to get tougher in the face of the enormous market power of big tech, which during the pandemic only embedded itself further into daily life.

However, beyond the heavyweight bouts between big tech and political authorities and implications for the companies, it is important to monitor how consumers are impacted by the new rules. Hopefully, consumers will benefit from more responsible online lending practices, fewer scams and predatory pricing, and more communication around the use of their data. As one Chinese internet regulator stated, consumers should not be “<a href="https://www.wsj.com/articles/chinas-regulators-prepare-to-roll-out-new-antimonopoly-rules-11605013205" target="_blank" rel="noopener noreferrer">prisoners to algorithms</a>.” We’ll be closely watching the market to understand how fintechs and consumers are responding and to understand what lessons other markets might draw from the Chinese approach to checking the power of big tech.

In the meantime,  read our recent publication, <a href="https://centerforfinancialinclusion.org/the-stories-algorithms-tell-bias-and-financial-inclusion-at-the-data-margins/" target="_blank" rel="noopener noreferrer"><em>The Stories Algorithms Tell: Bias and Financial Inclusion at the Data Margins</em></a>, about how FSPs are using algorithmic data to make credit decisions, how those affect low-income consumers globally, and what regulators can and should do.

Regulatory Headwinds on the Horizon for Chinese Fintechs: Where Do Consumers Fit In?

Across all industries, predictive, data-driven algorithms -- mathematical recipes ranging from the simple to the complex -- are being used to tell stories about individuals and can drive high-stakes decisions: who receives a loan, what sentencing a judge will recommend, what therapeutics a doctor will provide.

In the financial services industry, the exploding data ecosystem has created billions of new stories. At the Center for Financial Inclusion (CFI), we are most interested in the ones they try to tell (or don’t tell) about low-income consumers.

This publication has two main objectives:
<ul>
 	<li>to ground some universal challenges on the use of algorithms, automated decisions, alternative data, and bias in the context of inclusive financial services</li>
 	<li>to present the current state of play among inclusive finance actors from desk research and interviews with a sample of fintechs, regulators, and other experts.</li>
</ul>
Our insights are aimed primarily at financial inclusion stakeholders who can influence the trajectory of the inclusive finance industry, with specific recommendations for regulators, investors, and donors. But anyone with an interest in algorithmic decision making or financial inclusion will find this publication thought provoking and useful.

The full publication is divided into the following sections:
<ul>
 	<li>The <strong>introduction</strong> explores algorithms and bias in inclusive finance, presents a framework for understanding algorithmic bias, and delves into what we want to know about algorithms, bias, and financial inclusion.</li>
 	<li>Section 1 is about <strong>data trails in the digital economy</strong>.</li>
 	<li>Section 2 digs into the <strong>risks of bias and emerging tools</strong> through the three categories: inputs, code, and context.</li>
 	<li>Section 3 lays out <strong>recommendations for inclusive finance stakeholders</strong> to advance evidence, solutions, and incentives for responsible algorithms.</li>
</ul>

The Stories Algorithms Tell: Bias and Financial Inclusion at the Data Margins

In emerging economies, the growth of mobile phone use has contributed to an explosion of data that can be fed into algorithms to make predictions about individuals, including who is creditworthy, and offer financial products and services. As of 2020, <a href="https://www.gsma.com/r/wp-content/uploads/2020/09/GSMA-State-of-Mobile-Internet-Connectivity-Report-2020.pdf" target="_blank" rel="noopener noreferrer">half of the world’s population used mobile internet</a> and digital finance offerings have grown in tandem, often framed as engines of financial inclusion. The massive growth in available data and companies to serve customers that had previously been invisible to the financial system might suggest that the problem of thin-file customers has been solved and long-standing financial inclusion gaps are closing.

But the truth is, we don’t know what the impact is yet.

So amidst the rush to generate and leverage new data trails, CFI is embarking on a learning agenda around financial inclusion at the data margins. In pursuit of that, we are rethinking assumptions, and learning from broader tech-oriented conversations about the intersection of alternative data, algorithms and digital profiling, and how these more technical issues interact with socio-economic and gender-oriented issues surrounding digital literacy, financial capabilities, and social norms.
<h2>What Does it Mean to Lend to a Phone Line?</h2>
In the context of financial inclusion, making people visible to the financial market nowadays increasingly means having a digital financial identity based on volumes of (mobile) data that’s both related and unrelated to the actual financial behaviors of an individual. The expansion of digital lending, for example, utilizes non-traditional mobile data such as SMS, device usage and model information, social network activity, contact lists, call logs, app use, location data, and more to construct a picture of an individual’s cash flow or income.

Some of these sources are intuitively linked to financial activity, such as SMS receipts for mobile wallet transactions. Other data -- referred to by some as <a href="https://www.upturn.org/static/files/Knowing_the_Score_Oct_2014_v1_1.pdf" target="_blank" rel="noopener noreferrer">“fringe” alternative data</a> -- like social media activity, is further removed and its value as a fair and accurate predictor of creditworthiness is less established.

Given that data-driven models leverage phone-specific variables, the limits on women’s phone use or migrants’ access to a SIM card will change the data trail they leave and, in turn, the decision that an algorithm may make about their access to a loan.
<blockquote class="pullquote--right">The limits on women’s phone use or migrants’ access to a SIM card may change the decision an algorithm makes about their access to a loan.</blockquote>
In developing economies, women are less likely to own a phone and <a href="https://www.gsma.com/mobilefordevelopment/wp-content/uploads/2020/05/GSMA-The-Mobile-Gender-Gap-Report-2020.pdf" target="_blank" rel="noopener noreferrer">access mobile internet</a>. Phone-sharing commonly acts as a barrier to phone ownership for women as well as other vulnerable groups like migrants or refugees. It’s prevalent due to cost constraints or gaps in identity documents required for registering a SIM card. What does it mean when an algorithm is facilitating lending to a phone line, rather than an individual? What are the implications of this and other extant barriers in determining how inclusive these new tools are?
<h2>Risks of Data Marginalization: Where to Begin?</h2>
In the U.S. and Europe, risks and evidence of data marginalization have prompted calls for greater oversight, but little research has documented what these risks look like in the data environments of developing markets. There is much that we need to learn:
<ul>
 	<li>How large is the risk of cementing exclusion in a data-driven financial system for consumers at the data margins?</li>
 	<li>What strategies and tools are available to companies, investors, and regulators to address these risks, and what resources and skills do they require to be effective?</li>
 	<li>What are the incentives for private sector companies to act on this risk?</li>
 	<li>Globally, we see strides made on data privacy and emerging digital rights, but how will people with limited digital capability be able to meet these new rights and responsibilities?</li>
</ul>
CFI and others, like our friends at <a href="https://www.womensworldbanking.org/insights-and-impact/report-algorithmic-bias-financial-inclusion-and-gender-primer-opening-up-new-credit-to-women-in-emerging-economies/" target="_blank" rel="noopener noreferrer">Women’s World Banking</a>, are beginning to address these questions and establish a learning agenda for the sector.

It is critical we set out on a path for answers now in this period of global data ecosystem transformation. As a fintech from Southeast Asia explained, “You can’t make those [anti-bias] changes when you’re doing a million loans a day… a little bit of a loss now and a gain later is better than taking a loss at that scale.”

Let this remark serve as a wake-up call for all development stakeholders that are building data solutions today not to lose sight of the people behind the mobile phones and to examine how technology and data can help us build an inclusive tomorrow.

This topic is one of CFI’s strategic pillars, so consider this the first mile of a learning journey that will take all of us to a better understanding of data risks and opportunities as they relate to financial inclusion. Stay tuned to this space, sign up for <a href="https://centerforfinancialinclusion.us5.list-manage.com/subscribe/post?u=394352ae963270aef6c51680d&amp;id=9df382106a" target="_blank" rel="noopener noreferrer">CFI email updates</a>, and follow us on social media for coming explorations of data risks and opportunities and how they affect low-income consumers.