Financial Inclusion 2020 Blog Series banner imageFinancial Inclusion 2020 (FI2020) is a global multi-stakeholder movement to achieve full financial inclusion, using the year 2020 as a focal point for action. This blog series will spotlight financial inclusion efforts around the globe and share insights from key thought leaders in financial inclusion, with a specific focus on quality beyond access.

PERC, a “think and do tank” advancing financial inclusion through information services, has been effective in addressing credit invisibility by advocating the use of alternative data in credit reporting, including in Australia, Brazil, China, Kenya, and the U.S. We invited Michael Turner, PERC’s CEO, to submit an opinion piece, and are publishing the results in a three-part series. Part one and two can be found here and here; the following is part three.

Misperceptions abound about how to impact credit information sharing in emerging markets. Let me weigh in on this debate and set the record straight.

  • Technology is not the problem. There are abundant and affordable platforms to enable robust information sharing in even the most extreme environments.
  • Scoring models are not the problem. FICO, SAS, Dunn and Bradstreet, and a host of multi-national credit bureaus and lenders have plenty of smart mathematicians, computer scientists, statisticians, and others with lots of letters behind their surnames to ensure innovation in this space. The breakthrough that will move markets won’t be found here.
  • End-user capacity and incentives are not the problem. Many pro-poor lenders are already using automated underwriting solutions and can quickly assimilate new data or new scoring models.

So if investing in the technology, risk modeling, and end-user trenches aren’t going to galvanize things, let alone revolutionize them, in which trenches will the revolution begin? The answer lies further upstream, in the consumer and commercial credit ecosystems.

The answer is data access.

This is a deceptively simple response and raises a number of related questions. Which data is both predictive of credit worthiness and covers broad segments of the unbanked and underserved populations? Who owns it? Can traditional credit bureaus access this data? Why haven’t they so far? Are other parties needed to provide lenders access to this data? How can data subjects (people) access and “port” their data from mobile payment systems the same way they can carry their credit report information?

Most of the large, predictive, quality data sets come from lenders and are captured by traditional credit bureaus. This is good news for the already banked—usually the wealthier persons in any given economy. Getting data on the unbanked is a far greater challenge for the following three reasons:

  • First, many of the transactions nodes involving the unbanked don’t capture payment data digitally. These include transactions along the agricultural and retail supply chains—purchases of crops from smallholder farmers or sales from small retailers—which would provide timely and accurate income data that could be used to assess credit capacity. Other large data assets involve pre-paid data from MNOs, bill payment data from energy utilities (gas, heating oil, water, electric), wireless service and handset payment data, and possibly rent for retailers and residences—all of which is helpful for assessing a person’s (or micro-enterprise’s) credit risk.
  • Second, there are data quality and organizational capacity issues among non-financial data furnishers that must be overcome in order to reliably access such data over time. This data gathering/arranging won’t likely be done by traditional credit bureaus. It requires considerable resources that would dramatically reduce earnings and may make the operation non-viable. There are some supply chain automaters (players working to increase the efficiency of supply chains by doing things like integrating control software and automating processes for weighing and sampling crops) that are moving into this space, and this will help, but considerably more effort is needed. This is fertile soil for impact investors and grant-giving institutions with a focus on financial inclusion.
  • Third, most non-financial entities simply won’t share their data with traditional credit bureaus. By and large they don’t trust credit bureaus with their data and have real concerns about losing control of their customer data. Publicized misuse of customer data will result in a customer backlash against the furnishers that could negatively impact customer relations, revenues, and earnings. Further, most non-financial entities don’t consider the traditional credit bureau business model as fair—namely, that they must transfer their data asset to a credit bureau for free or a modest flat fee, and the credit bureau monetizes their asset and records earnings in the 30 to 50 percent range.

There’s no silver bullet to overcome these real obstacles, but let me be clear, unless and until more effort is invested in accessing a growing quantity of predictive data for credit risk assessment, digital financial services for the poor and the missing middle risk remaining stuck in their current state—namely as a payment system and a platform for digital higher cost lenders.

Let me offer a roadmap on overcoming these hurdles in order to empower the financial infrastructure necessary to take digital financial services to its next stage of evolution.

  1. Recognize the need for investing in upstream activities like financial infrastructure development (credit information sharing networks, payment systems, asset and collateral registries, and identity verification tools). Though far less attractive from a “sexy investment” perspective, such infrastructure is vital to the success of the next stage of digital financial services.
  2. Invest in efforts to digitize data from non-financial service providers in the agricultural supply chain (will help many smallholder farmers access affordable sources of productive capital) and the fast-moving consumer goods/retail supply chain (fast growing, cash-starved small business owners will benefit from expanded credit access).
  3. Invest in information sharing systems that are sensitive to the needs and incentives of the data furnishers. For instance, financial services firms have a clear interest in sharing customer payment data with a credit bureau, and need access on the same from other lenders to make responsible lending decisions. Not so for non-financial service providers that aren’t assessing risk for enlisting customers and have other means to punish those who pay late or don’t pay, such as termination of service. These organizations don’t often trust credit bureaus or any third-parties with their data, want remuneration, and fear customer backlash. These must be addressed for this data to be accessed in sufficient quantity for purposes of credit underwriting.
  4. Don’t focus only on efforts afoot in the poorest markets. Focusing exclusively on the base of the pyramid populations in very low-income markets ignores the development impact that can be achieved and sustained by helping those in the missing middle, especially in populous countries with high rates of financial exclusion and sufficient credit capacity (net disposable income among the credit excluded). Countries such as India, Mexico, Indonesia, South Africa, Malaysia, the Philippines, and Nigeria all meet the criteria necessary for success (typically middle-income countries whose large unbanked populations have sufficient credit capacity to capture benefits of scale).

In reality, the traditional credit bureaus remain best positioned to capture vast quantities of alternative data. Proof of this is found in the fact that several of them have made this commitment and have succeeded. Such success stories are few and far between, largely in mature economies. In emerging markets, traditional bureaus and other interested parties are going to have to pay for data access. The lack of large-scale activity in this space suggests that some, if not most traditional credit bureaus may be unwilling to buy much of this data even when it is available in emerging markets, and that smaller efforts are too narrowly focused or lack capacity to have much of an impact beyond small value consumption loans. The undersupply of this data, given the need and demand among lenders, suggests a market failure. Perhaps a large-scale effort can only come with the assistance of the philanthropic community to offset likely losses during the early stages owing to high and front-loaded investment costs associated with data acquisition.

While launching an alternative credit information agency is difficult, the bigger challenge remains with the prospective data furnishers. There is a clear market failure here. Despite the demand for greater quantities of predictive data, it is being grossly undersupplied. Credit bureaus and others have been unwilling to invest in upstream capacity building efforts to ensure access to reliable sources of predictive data and have chosen instead to wait for those furnishers to build their own capacity. This is understandable given their market orientation. This also creates an opportunity for philanthropic and charitable organizations with a financial inclusion mission to galvanize that process and in return receive a big bang for the buck in terms of development impact.

This is nothing short of a call to arms for funders/investors to jump into the upstream trenches with PERC. Help us start the data access revolution so that a broad range of digital financial services may proliferate and lift the boats of the base of the pyramid and missing middle alike around the world! Get your hands dirty by investing in data digitization and in alternative data furnisher capacity building in order to deliver desperately needed new sources of data in the near term.

Image credit: International Maize and Wheat Improvement Center

Have you read?

A Contrarian View of the M-Shwari ‘Revolution’

The Hard Truth about Monetizing Client Data

Financial Inclusion Trends and Innovators – 2015