Digitally Included, Yet Poor: What Do Their Data Trails Tell Us?

In recent years, there has been a lot of talk about the digital footprints of the world’s poor. Four years ago, Nandan Nilekani, the architect of India’s Aadhaar infrastructure, predicted, “Indian residents and businesses will be data rich before becoming economically rich.” In fact, the volume of digital records generated by poor people is increasing at an unprecedented rate. This is driven by the expansion of mobile internet and smartphone users, digital accounts, digital IDs, social media and other digital services. According to GSMA Intelligence, in 2019, 57% of the Sub-Saharan Africa’s mobile subscribers used mobile internet (meaning they were digitally included). Yet, during the same time, 51% of the region’s population lived on less than $5.50 per day, according to World Bank poverty estimates.

As highlighted by Nilekani, there are large segments emerging across the globe that are generating significant volumes of digital data – yet they remain economically poor. Do we know who they are, what pools of data they generate, and where this data sits? More importantly, how can data trails be leveraged to unlock economic opportunities and improve the value of financial services for the poor, especially for women?

The expanding volume and richness of data trails resulting from increased digitization, paired with growing data analytics capabilities, presents an opportunity to further advance financial inclusion. Earlier this year, CGAP embarked on a three-year journey to explore some of these areas. The aim is to help low-income people, especially women, benefit from financial services that build on increasingly rich data trails.

As part of our agenda, we have begun a customer landscaping exercise. The goal of this exercise is to shed light on how the development community can use data trails to help digitally included, yet economically disadvantaged segments access financial services, thereby improving their livelihoods and building their resilience. The questions we are asking include but are not limited to:

• Who are the digitally included, yet economically poor? How do women feature in this segment? Arriving at the market size for this segment is relatively straightforward. To do this, we will overlay poverty data with data on mobile internet and smartphone users. These users will then be split by gender to estimate the number of women. Globally, those with digital tools facing economic hardships comprise a large and diverse segment. It will therefore be important to disaggregate data by demographics, income, location (urban-rural), occupation, and whether users receive government-to-person (G2P) benefits. Occupational segments will include smallholder farmers, platform workers, and entrepreneurs with micro, small and medium enterprises (MSMEs).
  
  However, we will need to take the numbers that we arrive at with a grain of salt since they could mask several nuances within this segment. Consider women with digital capabilities, for example. While they may have access to digital devices, their use of these tools could be regulated by family members due to male-dominated societal practices. CGAP has previously explored how gendered social norms hold women back from overcoming the gender-based digital divide – specifically, by limiting women’s access to handsets, internet connectivity, infrastructure and opportunities to improve digital literacy. We would need to take factors like these into consideration before arriving at any conclusions.
   
• What data pools do they generate? The digital data pools generated by the poor include information on identity, demographic and social characteristics, economic disposition, occupation and finances. For example, if we were to look at the digital activity of a woman who is employed as a domestic worker in Jakarta, Indonesia, we might note that she: 
   
  • Uses her smartphone to communicate with her family on WhatsApp
  • Withdraws money from a bank branch for which she receives an SMS acknowledgment from the bank
  • Watches videos on YouTube for entertainment
  • Sets up her son’s online class using Google Meets on her phone
  • Sends money back to her family in the village through GoPay
  • Uses Facebook to sell homemade goods as a supplementary source of income

When we look at the digital activity of economically disadvantaged groups, we need to consider a few overarching questions. Specifically, what are the data trails telling us about users’ economic, social and behavioral characteristics? And how could these trails be used by financial services providers (FSPs) to push the frontiers of financial inclusion?

• Where does this data sit? Some of the obvious sources for accessing data are mobile devices, governments, public registries, private credit bureaus, financial services sector, telecom and utilities sectors, e-commerce providers, platforms and social media platforms. What could be the other sources? More importantly, which sources are most relevant and reliable for FSPs?

While considering these questions, we are looking at how FSPs access and use data to optimize or build new services, with a focus on how FSPs use data to reach traditionally underserved groups, such as women, farmers and other segments. Our hypothesis is that FSPs can use data to improve risk management and enhance efficiency in operations, resulting in lower costs and better products. Additionally, data-driven segmentation and customer research can help FSPs identify and target customers and better serve their needs. We’ll be complementing our work on customers and FSPs by exploring ways to promote inclusive data sharing at the market level. Our work in this area will focus on policy regimes, infrastructure (for example, credit bureaus and APIs), and open banking frameworks that govern a country’s data sharing ecosystem.

We recognize that exploring opportunities to use data to advance financial inclusion should be balanced with an understanding of the emerging risks. Risks go beyond data protection and privacy risks and include other issues such as algorithmic decisioning bias and over-indebtedness.

We see a key role for several stakeholders, including FSPs, players in the platform economy, members of the data sharing ecosystem, and the donor community. Collaborative efforts could uncover how data about lower-income people can be used to help them in capturing income-generating opportunities and build resilience through financial services.

Learn more about CGAP’s work on data and policy here.