Recap of Financial Inclusion Week Twitter chat: Getting digital inclusion right

Globally, 1.7 billion adults remain unbanked without an account at a financial institution or mobile money provider according to the World Bank’s Global Findex 2017 report. Yet, two thirds of these unbanked adults have a mobile phone, making digital technology a great opportunity to broaden financial inclusion.

On October 31st, J-PAL’s Finance sector and the Center for Effective Global Action’s Digital Credit Observatory hosted a Twitter chat about digital financial inclusion as part of Acción’s fourth annual Financial Inclusion Week 2018.

This year’s #FinclusionWeek theme was “Getting Inclusion Right,” so we invited the Twitter community to join our discussion and ask questions about getting digital inclusion right.

Over fifty people participated by asking questions, sharing evidence, or liking and retweeting posts, including J-PAL affiliates Dean Karlan, Paul Niehaus, and Tavneet Suri, Innovations for Poverty Action (IPA), and the World Bank’s Global Findex.

Some of the main take-aways include:

• Early evidence suggests the promising potential of digital finance to increase financial inclusion and improve households’ welfare.
• Digital finance may also expand access to financial services and products to the most marginalized populations, particularly women and the very poor, but barriers remain to achieving equal access.
• Digital finance also brings risks to consumers and financial providers; regulations and consumer protection are necessary to ensure responsible digital finance.
• Researchers continue to pursue new and exciting research on digital finance (many shared during the Twitter chat), with many interesting, open questions left.

Read on for a recap of the lively conversation about the research on digital finance.

Is digital finance leaving anyone behind?

We kicked off our chat by discussing how digital finance may be a promising way to expand financial inclusion for those who are currently unbanked. A participant asked about how we can achieve financial inclusion for the very poor, especially because traditional financial services tend not to reach low-income customers. Researchers agreed that last-mile access remains a challenge.

But my question is, apart from group lending strategy how financial products can reach the very poor? (In conventional banking system they used to refer them as Unbankable group)

— Inno Rusomyo (@innorusomyo) October 31, 2018

Thanks @innorusomyo for the question. Reaching the very poor can be quite costly, so digital finance is a promising avenue. We cover other ways to reduce transaction costs in our policy insight: https://t.co/igkZrHabRJ @CEGA_UC #FinclusionWeek https://t.co/7AVLMej40H

— J-PAL (@JPAL_Global) October 31, 2018

Indeed, recent large scale surveys from #Kenya and #Tanzania found that digital credit remained out of reach or unused by the most vulnerable groups @CGAP, @FSDKe, and @FSDTanzania https://t.co/XRqUppeZ55

— CEGA (@CEGA_UC) October 31, 2018

There seems to be a consistent patter about who uses digital credit, and who is still excluded. In both Kenya and Tanzania, farmers and casual labour are far less likely to use digital credit compared to salaried employees and entrepreneurs. More details: https://t.co/1IVUAA8hmQ

— Edoardo Totolo (@edso) October 31, 2018

Lowering costs for providers and users

Even when traditional financial services are available, user fees and other barriers may be prohibitively costly for low-income customers. We discussed ways that financial service providers—both digital and otherwise—can help lower these transaction costs.

Low trust in financial institutions is a barrier to #FinancialInclusion, but #digital financial services can ↑ trust and savings. Bachas, Gertler, @SeanKHiggins, Seira explain in @VoxDev https://t.co/xbakTq9gmJ

— CEGA (@CEGA_UC) October 31, 2018

In the case of cash transfers, M-Pesa and other mobile money services reduce transfer costs to a minimum. For other financial products, my sense is increases in smartphone penetration will solve the access problem quite soon.

— Johannes Haushofer (@jhaushofer) October 31, 2018

You may want to check out work by @FSDKe on expansion of access through digital credit in Tanzania and Kenya. I also worked with them on putting together a summary of our results on M-Shwari in Kenya, available here: https://t.co/X9ptXyxU0V

— Tavneet Suri (@SuriTavneet) October 31, 2018

However, updated services design can help! @CGAP, @FSDKe, and @FSDTanzania outlined a few possibilities: more nuanced algorithms, flexible repayment structures, timeframes suited to managing the uncertainty and the seasonality of rural livelihoods, and pricing. #FinclusionWeek

— CEGA (@CEGA_UC) October 31, 2018

Closing the gender gap in digital finance

Digital finance may also be an effective tool to expand access to accounts for women, who currently have lower rates of financial access than men. And with expanded access, there is evidence of improved outcomes for women and their families.

#Women represent 56 percent of the unbanked. Can mobile money help close the #gendergap in financial inclusion? #FinclusionWeek @CEGA_UC pic.twitter.com/rR0i3ICCCm

— J-PAL (@JPAL_Global) October 31, 2018

We are optimistic. In African countries with high mobile money penetration, we find the gender gaps between mobile money users are lower than those between bank account users. Maybe it’s more difficult for a woman to travel or she trusts her local
agent. pic.twitter.com/s5kfz4EfyS

— World Bank Findex (@GlobalFindex) October 31, 2018

CEGA Post-Doc @SeanKHiggins is exploring how #DigitalCredit algorithms using #AI and #MachineLearning could ↓ bias against women in access to credit https://t.co/K7x2XYCB29 https://t.co/2FkUrM5KO8

— CEGA (@CEGA_UC) October 31, 2018

(1/2) Mobile money may give women more control over household spending. In Niger, women reported that mobile money was easier to hide from their husbands: https://t.co/sE4WKOuc2L). https://t.co/CKvZwoWeAB

— J-PAL (@JPAL_Global) October 31, 2018

Gender-specific barriers in digital finance

Economic constraints like the costs associated with owning a mobile phone and technical literacy required to operate it disproportionally impact women. Additionally, social norms may limit if women have and how they use mobile phones. Researchers are continuing to explore barriers women face in accessing mobile money.

Despite the potential, women still face barriers to accessing mobile money. https://t.co/MQrIxibNFP @EPoDHarvard https://t.co/TuH9MKWE5Q

— J-PAL (@JPAL_Global) October 31, 2018

Absolutely. But Findex data shows men in Bangladesh are 25 percentage points more likely than women to have a mobile phone. Are women being left behind at the starting gate?! https://t.co/dbbYOOxZZL pic.twitter.com/o2GBdIkjHi

— World Bank Findex (@GlobalFindex) October 31, 2018

Great question! I don't know for Bangladesh. But, definitely not in Kenya where we found mobile money caused bigger reductions in poverty for female headed households & occupational changes only for women (in both male and female headed households). See https://t.co/6UnEW7sWv9

— Tavneet Suri (@SuriTavneet) October 31, 2018

We are actually working on some new stuff on Uganda, Tanzania and Bangladesh - stay tuned!! So hopefully we can try to have an answer to this for more countries...

— Tavneet Suri (@SuriTavneet) October 31, 2018

#FinclusionWeek comes a bit early for the most recent study by @Oxford_CSAE’s @EmmaRiley19 on the effect of dispersing micro loans to women through mobile money. This could help women fending of pressure to spend the money on non-business stuff. Results will be available soon!

— Lukas Hensel (@LukasHenselEcon) October 31, 2018

Innovative opportunities with digital transfers

Mobile money accounts can also be used to deliver unconditional cash transfers and other safety net programs. Researchers chimed in with results from evaluations of Give Directly’s programs that provide unconditional cash transfers to eligible households via mobile money and also shared a study on a state welfare program in India.

(1/3) @jeremypshapiro and I evaluated @Give_Directly’s program in Kenya that gives unconditional cash transfers (UCTs) to poor households via the mobile money platform m-pesa #FinclusionWeek https://t.co/7h6gvfeBTm

— Johannes Haushofer (@jhaushofer) October 31, 2018

(2/3) After 9 months, @Give_Directly's program increased households’ assets by 61%, consumption by 23%, and business revenues #FinclusionWeek

— Johannes Haushofer (@jhaushofer) October 31, 2018

(3/3) @Give_Directly households also had improved psychological well-being, lower levels of depression, increased happiness and life satisfaction. Read more about the study here: https://t.co/sO527XrXcI #FinclusionWeek

— Johannes Haushofer (@jhaushofer) October 31, 2018

Three years after transfers, the impacts on assets persist. Other impacts are still there in the within-village comparison, but not across villages, suggesting possible spillovers. Difficult to make conclusive statements because of some differential attrition #FinclusionWeek https://t.co/PvpYQbjvq1

— Johannes Haushofer (@jhaushofer) October 31, 2018

@GiveDirectly treats all eligible households in a village now. Watch this space for evidence on price effects and spillovers on ineligible households, with @tedmiguel, @PaulFNiehaus, and Michael Walker

— Johannes Haushofer (@jhaushofer) October 31, 2018

Also check out our affiliate Craig McIntosh and Andrew Zeitlin conducted an #evaluation of @Give_Directly #cashtransfers in #Rwanda -https://t.co/0uIwnv4yPF https://t.co/vdQ5HjkuoG

— CEGA (@CEGA_UC) October 31, 2018

@PaulNiehaus @Prof_Karthik_M Sandip Sukhtankar found that investing in secure payments infrastructure e.g. biometrically authenticated payments infrastructure (Smartcards) could significantly enhance 'state capacity' to implement welfare programs in India https://t.co/OEvKM3iv0b

— CEGA (@CEGA_UC) October 31, 2018

Evidence on digital financial services and migration

Another participant asked about ways mobile money could facilitate migration and we shared randomized and non-randomized evaluations on this topic in response.

@JPAL_Global Is there past/ongoing research on the role of mobile-money in facilitating (internal/external) migration on the intensive and/or extensive margin? Thanks in advance! #FinclusionWeek

— Akib Khan (@akib_kn) October 31, 2018

Thanks @akib_kn for the Q. There is some #evidence that credit constraints may be a barrier to seasonal migration (like in this paper: https://t.co/1EU5e11VKh); digital credit may be way to alleviate this, but this an important question for future research. #FinclusionWeek https://t.co/rRdj0tztmF

— J-PAL (@JPAL_Global) October 31, 2018

This may be a good place to start: https://t.co/Y1RoCpbb3m

— Tavneet Suri (@SuriTavneet) October 31, 2018

Our affiliates Michael Callen and @jblumenstock touch on the issues of #migration and #mobilemoney in a series of papers including https://t.co/qzLFIVWq0b https://t.co/umuzow9zwW

— CEGA (@CEGA_UC) October 31, 2018

Migration and the Value of Social Networkshttps://t.co/4zoqW7JEQE

— CEGA (@CEGA_UC) October 31, 2018

Ethical questions remain

Digital platforms can also facilitate alternative credit scoring methods based on borrower’s mobile phone data. This can help further expand access to credit by improving lenders’ ability to assess creditworthiness—even for potential borrowers who lack a credit history. The DCO shared some key considerations when it comes to ethics of machine learning and credit scoring algorithms, as well as ongoing research on the topic in Kenya.

Hey @JPAL_Global and @CEGA_UC, with an increasing number of alternative methods of assessing credit, what protections should be in place to make sure that the algorithms used are ethical, fair and unbiased? #FinclusionWeek

— Elvis Wong (@elviswcwong) October 31, 2018

“Don’t forget people in the use of big data for development” @jblumenstock talks pitfalls of big data and biased algorithms, and ways to move forwardhttps://t.co/X4ye8QVyBN

— CEGA (@CEGA_UC) October 31, 2018

Watch DCO researcher @SeanKHiggins talk about his work on gender-differentiated credit-scoring algorithms #SmartDev2018 #MachineLearning #ArtificialIntelligencehttps://t.co/AZmmHKbgEU

— CEGA (@CEGA_UC) October 31, 2018

Certainly some potential risks with (alt) credit scores: In #Colombia, computer-generated credit scores led banks to extend larger loans to less risky borrowers and smaller loans to riskier borrowers - who may be more vulnerable. https://t.co/ws5JGlKERu #FinclusionWeek @CEGA_UC https://t.co/yIjezDBKaL

— J-PAL (@JPAL_Global) October 31, 2018

A few possible resources on #ethics and #MachineLearning include this panel discussion from #SmartDev2018 w/ @FKondylis @ermonste @aubra_anthony & Moorea Bregahttps://t.co/HQ6u4H5VoY https://t.co/VHK73JLuAX

— CEGA (@CEGA_UC) October 31, 2018

The DCO's Scientific Director @jblumenstock tackles some of these issues in his research on #Algorithms and #digitalcredit products https://t.co/sEspNW9I6m pic.twitter.com/UJCEUVGt9M

— CEGA (@CEGA_UC) October 31, 2018

With opportunities come challenges

Despite the promise of digital finance, quicker and easier access to credit also poses risks. Appropriate regulation and consumer protection is important—this is an area where more research is required.

Digital credit has reached large segments of the population in East Africa. In Kenya, one in three mobile owners has borrowed a digital loan. No other lending technology has ever developed at this speed. This obviously has risks https://t.co/vw4W9CyPx4 pic.twitter.com/rCKpJc8HJY

— Edoardo Totolo (@edso) October 31, 2018

basic tensions here imho:

(a) generally the best way to make sure you create valuable things is to try to sell them

(b) that's harder when you are creating for people with little $

(c) with financial products you worry more than usual that folks may self-harm, e.g. debt https://t.co/wX9qor1VQe

— Paul Niehaus (@PaulFNiehaus) October 31, 2018

First step is to be more disciplined in giving grant money to consumer lending. I’ve seen a multi million dollar grant to fund a telcos marketing campaign for a new digital loan. How is that anything but a corporate subsidy? https://t.co/cHSmEyO3Mw

— Rafe Mazer (@rkmazer) October 31, 2018

It is not that #digitalcredit per se is damaging (quite the opposite), but the current plain vanilla models need (significant) overhaul - much as microfinance did. See https://t.co/2ZSGapMeSe #FinclusionWeek @CEGA_UC

— Graham A. N. Wright (@GrahamANWright) October 31, 2018

Our current research will have some insights on #regulation, but in the meantime, the DCO has relevant resources here: https://t.co/FT0Dg8R2zt https://t.co/bU6Gtm715Q

— CEGA (@CEGA_UC) October 31, 2018

Look at these snazzy infographics explaining links between financial inclusion and the sustainable development goals! Excellent work @UNSGSA and @BetterThan_Cash https://t.co/BTVZU76tuf

— World Bank Findex (@GlobalFindex) June 6, 2018

Moving beyond access to impact

A big picture question cuts across all financial inclusion research: how does access to financial products and services affect wellbeing? The Global Findex shared how digital financial inclusion can help us achieve the Sustainable Development Goals.

Look at these snazzy infographics explaining links between financial inclusion and the sustainable development goals! Excellent work @UNSGSA and @BetterThan_Cash https://t.co/BTVZU76tuf

— World Bank Findex (@GlobalFindex) June 6, 2018

What’s next?

With so many unanswered questions around digital finance, J-PAL’s Executive Director Iqbal Dhaliwal asked what interesting open questions researchers are exploring.

My question for the #finclusionweek chat happening now with @JPAL_Global @CEGA_UC: What are the most exciting open questions J-PAL affiliates like @SuriTavneet and @deankarlan are asking about digital finance? (Join the discussion or ask your own question with #finclusionweek!)

— Iqbal Dhaliwal (@iqbaldhali) October 31, 2018

At the DCO we're looking at
1/3 What are the short-and long-run impacts (both positive and negative) of #digitalcredit on consumers in emerging markets? https://t.co/WezsJEusMU

— CEGA (@CEGA_UC) October 31, 2018

2/3 Is there heterogeneity in impacts along borrower characteristics (i.e. financial literacy, time preferences, income, and/or #gender)?

— CEGA (@CEGA_UC) October 31, 2018

3/3 How can non-traditional credit-scoring algorithms, regulations, and other #consumerprotection measures be designed to minimize default, over-indebtedness, leakage, fraud and other risks to consumers?

— CEGA (@CEGA_UC) October 31, 2018

How large are the spillovers from #digital #fintech adoption? My #JMP looks at this question for epayments in #Mexico. Gov policy ↑ dynamic fintech adoption on both sides of market (consumers, merchants, then more consumers). Abstact at https://t.co/FMEQLOcOtD #FinclusionWeek https://t.co/1kqLrY0UGs

— Sean Higgins (@SeanKHiggins) October 31, 2018

How large are the spillovers from #digital #fintech adoption? My #JMP looks at this question for epayments in #Mexico. Gov policy ↑ dynamic fintech adoption on both sides of market (consumers, merchants, then more consumers). Abstact at https://t.co/FMEQLOcOtD #FinclusionWeek https://t.co/1kqLrY0UGs

— Sean Higgins (@SeanKHiggins) October 31, 2018

In response, Dean Karlan, our Finance Sector chair, raised questions he and other researchers are considering and offered his insightful summary of the opportunities and challenges that digital finance presents.

(1/4) We (@deankarlan, @SuriTavneet, along with others) raise some important questions and concerns in our paper https://t.co/NVEuP2boeB #FinclusionWeek

— Dean Karlan (@deankarlan) October 31, 2018

(2/4) How to scale safe digital credit? Digital finance providers’ incentives may not align with maximizing client’s welfare #FinclusionWeek

— Dean Karlan (@deankarlan) October 31, 2018

(3/4) How do we regulate the growing industry of digital finance? How do we ensure consumer protection for private client data? #FinclusionWeek

— Dean Karlan (@deankarlan) October 31, 2018

(4/4) How do we link digital finance to end goals? Finance is a means to an end: a way to respond to shocks and live happier, healthier lives #FinclusionWeek

— Dean Karlan (@deankarlan) October 31, 2018

Glad to join @JPAL_Global and @CEGA_UC #FinclusionWeek Twitter chat. My summary on #digital credit: an economist fantasy and a behavioralist nightmare. Transaction costs plummet, but also makes it easy to mess up. How to protect consumers?

— Dean Karlan (@deankarlan) October 31, 2018

Thanks to all who participated and shared resources during the discussion. Check out the J-PAL website to read a recap of the Financial Inclusion week Twitter chat we co-hosted last year and to access more evidence on digital finance and more topics in financial inclusion. For a review of the current literature on digital credit, check out the Digital Credit Observatory’s landscape analysis “Digital Credit in Emerging Markets: A Snapshot of the Current Landscape and Open Research Questions.”