Connect with us


‘Alternative data’ makes the lending difference

By JACO VAN JAARSVELDT, chief decision analytics officer at Experian Africa



The devastating economic impact of Covid-19 has created a much tougher lending environment for credit providers that now need to go far beyond traditional criteria when assessing risk and modelling future losses on credit granted. Customer credit profiles are drastically changing and “business-as-usual” strategies will no longer be sufficient in predicting future credit performance and extrapolating loss events based on observed payment performance.

The traditional credit score has been used by the financial services industry for decades, but the way in which consumers are managing their money and credit, particularly in the times of Covid, has changed. When we layer on additional sources of data – like alternative financial services information, consumer-permissioned data, rental and utility payments, full-file public records and spending, travel and association habits, a much more comprehensive picture of the consumer at a hyper-personalised individual level emerges.

These insights help lenders to enhance their risk decisions, refine their provision overlay models based on statistically proven behavioural variables, and most importantly expand their target credit universe by using non-traditional data to assess risk and predict future performance on individuals that have no credit track record. In addition to supporting a better risk decision, access to alternate data sources like device IDs that link to an ID document and photo which in turn links to a selfie helps prevent fraud, thus putting the power to self-protect in the hands of consumers.

Jaco van Jaarsveldt, chief decision analytics officer at Experian Africa

Alternative credit data provides supplemental data to enrich decisions across the entire credit spectrum. Ultimately, this new data drives greater access to credit for consumers and profitable growth for lenders through more informed lending decisions.

The impact of COVID-19 on historic credit data

When you build a predictive risk model you look at historic data and predict the future based on the behavioural aspects of a group of ‘like’ consumers. However, the simple traditional generic use of now historic 2020 data, the year of COVID, is not prudent nor will it be statistically predictive due to the black swan nature of the event.

Consumers with historically flawless credit records who fell into hardship due to the impact of the extended lockdowns, resulting in loss of income and worse case unemployment, will miss payments, many for the first time ever. Does that make them ‘bad’ credit customers? Applying traditional methods based on stable historical data will result in them being defined as defaulted and in time ‘bad’ credit customers. The real question however is, are they? Or are they just a ‘false bad’ that should be treated very differently to a traditional bad credit consumer during these unique, trying times?

Financial institutions need to look at consumers differently and should be able to identify financial distress based on the additional, non-credit data that is individual specific and not based on statistical samples that historically were alike.

What are the pros of using alternative credit data?

In general, many people face barriers to accessing credit or pay more for credit for several reasons. Some have trouble documenting their income. Others have either no credit history or a credit history that is too scarce, or “thin” to generate a credit score. This issue more often affects low-income consumers, and this is where alternative credit data can play a positive role.

Suddenly, we can see if a consumer on the traditional file is using alternative finance products at the same time, revealing how they are treating payments or purchase decisions in both instances. Additionally, alternative credit data can provide insights on an entirely different consumer population not on the traditional credit file by assessing other behavioural aspects that are similarly predictive of future behaviour.

Measuring consumer stability

One of the most predictive factors to determine intent to repay is a consumer’s stability – both in relation to their financial affairs and personal information. In the subprime market, consumer stability is linked directly to future loan performance. For example, it has been statistically proven that the more changes a consumer has in cell phone number, bank account or home address, the higher the likelihood of a default or delinquency.

An “unscoreable” individual is not necessarily a high credit risk – rather they are an unknown credit risk. Many of these individuals pay rent on time and in full each month and buy airtime and data on a consistent, frequent basis and could be great candidates for traditional credit. They just don’t have a traditional credit history to be used as the benchmark.

Looking forward

Improving financial access for thin-file and credit invisible consumers will play an integral role in our road to economic recovery.

While it is difficult to predict when the economy will return to pre-COVID-19 levels, it has never been more important to protect consumers’ financial health and maintain the integrity and accessibility of the credit economy.

Combining traditional credit data with alternative data to access consumer creditworthiness and setting effective collection strategies for the new ‘false bad’ population plays a critical role in achieving this. Doing so will help determine a consumer’s stability, ability, and willingness to repay during the current financial landscape and more importantly beyond.

Ultimately, alternative data can improve credit access and decisioning for millions of consumers in South Africa who may otherwise be overlooked. The data exists. The opportunity now is how to leverage it to support today’s diverse consumer base.