Hype or Hope? Implications of Big Data for Financial Inclusion

Alibaba in China is the largest ecommerce platform in the world, processing $250 billion just in 2013 alone. Beyond commerce, Alibaba provides a range of payment, credit and investment services almost entirely digitally without a branch network. Alibaba leverages its e-commerce data to assess risk and offer credit services to businesses and its electronic payments solution to channel mutual fund investments for the mass market.

Does Alibaba signal a future model for financial services globally? Is this the direction in which providers may increasingly reach poor customers in other markets? We hold that global trends in data creation and use are driving towards models that allow delivery with ever less reliance on a physical in person delivery system.

Maze of sand and water Maze of sand and water

Photo Credit: Chi Keung Wong

Data has always been an essential element of financial service delivery. Providers use demographic and geographic information to decide who to target; lenders use income, credit history, age and other factors to make credit judgments; and insurers use demographic data to target specific market segments and set premiums. Often this has been “soft” data based on relationship information managed by individual staff. Or it may have been collected in paper form (sometimes self-reported by customers) which is difficult/costly to verify, and very difficult to aggregate).

What we anticipate as a significant opportunity, is that there is more and more “hard” information becoming available which is:

  • In electronic form (which makes it easy to handle, store, manipulate, analyze)
  • Centralized and aggregated
  • Generated in real time
  • Easy to collect and store

There is a sharp increase in the data being generated and the speed with which it is made available. This is combined with new computing powers and analytic techniques. Digital channels help bring together data from different sources, analyze patterns across numerous and seemingly disparate variables, helping infer something useful about an individual’s financial habits, cash flows, and other aspects of behavior.

This changes financial service delivery in several important ways.

  • First, data that aggregates customer behavior across different dimensions (e.g., airtime consumption and savings behavior) can provide insights for a large aggregated pool of users, while at the same time provide highly individualized insights about individuals. Big Data brings both scope and depth.
  • Second, it opens the possibility of tailoring product characteristics to the needs of individuals. This means that providers can develop tailored offerings, rather than a single mass-market offering. The fact that data can be accessed digitally, centrally and in real time means that the cycle to improve products and service can happen faster.
  • Third, the availability of digital channels allows the deployment of data-enabled services at a large scale. Providers can have information they need to know and make decisions without physical presence. This opens the possibility of remotely issuing loans or insurance over digital channels that can reach large segments of the market fast, including unbanked.
  • Finally, the existence of data itself may begin to change the motivations and behaviors of clients. For instance, being sure to repay a loan knowing that failure to do so might limit their access to future loans.

To leverage the potential and use Big Data in the most beneficial ways, we need to begin to address critical challenges. 

  1. Getting past the limited data available in some environments. Low-income markets have a low starting base of digital data stored. Although some markets are accelerating fast and there are now 50 mobile subscribers per 100 adults, the amount and diversity of data stored are limited. This may require more creativity cultivating data sources that may be available such as low-income focused loan portfolios, satellite imagery and public data.
  2. Finding models that allow better sharing of data. Most new digital data is generated through the delivery of non banking services (such as mobile airtime) and therefore the link to financial service providers is not always strong. There can be regulatory restrictions on data sharing Beyond that, another challenge is properly valuing data. Over valuation could mean owners ask too high a price for others to use. Undervaluing means they often leave the data under-utilized or simply don’t make it available. With new business models so unproven and young it is often hard to make the best judgment.
  3. Ensuring adequate protection for customers. Many customers may not be fully aware of the implications of entering their data digitally, and the need for immediate access to a service, traditional forms of informed consent may have limited impact.

While these challenges exist, the trend towards greater use is strong. New start-ups such as First Access, Cignifi, Tiaxa and DeMyst Data are illustrating the potential of some of these ideas. Financial services traditionally delivered in-person are increasingly going to be complemented by much greater use of virtual models of financial service delivery. Data will be a key enabler. This next generation of services, delivered through large-scale digital channels stands poised to offer a major breakthrough in the reach of financial services to the under-served and those about whom information has traditionally been thin.


06 September 2014 Submitted by Peter Burgess (not verified)

Big data is hype ... and maybe microcredit is hype as well. The for profit investor financed growth of microcredit seems to deliver as much of economic damage as it delivers economic good. The idea that there is 'profit' at the bottom of the pyramid is an oxymoron. There is 'need' at the BoP and there is potential at the BoP, and in due course in a generation maybe there will be opportunities for profit, but not in the short term. The BoP needs access to services like food, water, knowledge, education, health, infrastructure, security etc and microfinance is a side show that diverts attention from all these other matters. The mcrocredit that has been valuable are the situations where the money was an excuse to get involved with all the other things.
Peter Burgess multi dimension impact accounting

07 September 2014 Submitted by Graziosi Ascanio (not verified)

This is a good article but, in my view, provided a partial answer to the question mark. Indeed, the implications aren't few and complicated. The Practitioners should be very aware that elaborating data is a necessary way to manage big numbers and the use of parameters or other mathematical tools is the basic basket to deal with big numbers.

However it is also true that, at the end of the day, the necessary educational background along with the daily exposure to risk make it the difference among Portfolio Managers.

How to match credit risk and big data is the real challenge: whatever the solution shall be, the Portfolio Managers should never neglect that the use of mathematics is a necessary but not sufficient condition, the above mentioned profile of the Portfolio Manager making it the difference.

Here is the hard job: how to translate into mathematical formula qualitative and quantitative variables and minimize the risk.

Referring to the under-served people and/or reduce poverty add a further variable to the credit delivery equation: sustainability. In other words, this does mean how to distinguish between micro finance, micro credit and micro grant.
Dr. Ascanio Graziosi

11 September 2014 Submitted by Ken Zita (not verified)

This post is spot-on.

Big Data – we can bypass the technical and just call it Digital Reality – is already transforming financial services in the OECD and is poised to do so at the bottom of the pyramid. The trend is inevitable, and the potential is huge.

Cloud-based analytics and data aggregation enable an entirely new approach to credit risk, and ultimately, to financial inclusion. It’s no secret banks are risk-averse and retrenching and rely on conventional models for establishing whether companies are creditworthy. Big data enables new and diverse views into borrowers’ financial behavior -- from completed export transactions to utility payment histories, online purchases, social reputation, psychometric profiles and more. This is not the sort of data bankers typically look at to make lending decisions. But it is potentially enough information for non-bank financial institutions and private investors to calibrate credit risk. And this is the point: big data allows consideration of borrower risk according to entirely new criteria, apart from conventional banking measures and expectations. Non-bank actors will leverage big data to create opportunities according to fundamentally different measures and rules than established banks.

Prospects for scalable, cloud-based lending services are especially encouraging for SMEs – companies that are too small to attract the attention of traditional financing sources, and too big for microfinance. Given that banks are reluctant to lend to SMEs – IFC estimates the global “missing middle” funding gap to top $1.6 trillion – alternative lending models based on big data may provide a new way forward.

Some bankers laugh off the big data wave: It’s nonsense! No financials no loan! No collateral no recourse -- no deal! These rumblings are not so different than what we heard from newspapers and music labels and retailers before they got crushed by the web. Big data and digital markets will challenge the core business of banking and underserved companies in the frontier markets stand to gain.

Add new comment