Leverage Analytics to Predict Behaviour and Boost Collections

Better predict behaviour to raise collection efficiency

As the COVID-19 pandemic affects economies worldwide, personal loans will undoubtedly also feel the pinch as money lenders are forced to focus resources on collections. How can analytics be leveraged to identify customers most likely to repay their obligations? And how can the process be made more efficient and cost-effective?

At TransUnion’s recent 2020 Hong Kong Financial Services Summit, Nidhi Verma, Vice President, International Research and Consulting, spoke on the subject. In her session — Leveraging Analytics to Prioritise Collections — she examined how to: improve delinquencies on unsecured loan products, better predict repayment behaviour of borrowers (based on their overall credit status), and use those insights to enhance the collection process.

As the economy goes, so do cure rates

It’s important to note Hong Kong’s credit market has historically experienced extremely low delinquencies compared to other credit economies. But this trend has worsened in recent years, especially for repayments of 90 days past due (for all loan products). Comparing data from 2017 against data from the second quarter of 2020, delinquencies in this category spiked by 32.8%. Average delinquent balances from Q2 of 2020 also rose to HK$73,000 — an increase of over 11% year-on-year.

By analysing roll rates, we can better understand the impact of recent economic conditions on the repayment status of borrowers. For example, if an account is “30 days-past-due (DPD)” in August, and becomes “60 DPD” in September, this is an example of “roll forward” and shows the situation has deteriorated. As another example, if part of the arrears is paid off and remains 30-DPD, this can be referred to as “stable.” Now let’s say the account has moved from 60-DPD to 30-DPD, this shows the situation has improved and has “rolled backwards.” And when the balance has been finally paid off, the account is said to be “cured.”

Delinquency rates deteriorate the later collection begins

According to TransUnion data (across credit cards, personal loans and revolving lines), the roll forward rate for revolving credit is highest in the event of delinquencies. For accounts with 60–89 DPD (2–3 months), the roll forward rate rises significantly versus 30–59 DPD (1–2 months). Let’s take revolving lines as an example: In June 2019, the 30–59 DPD roll forward rate was 48%, with the 60–89 DPD roll forward rate increasing to 71%. In other words, if collection starts later, the chances of a successful collection will be much lower.

When looking at other characteristics of consumers, age isn’t necessarily a driver for roll rate performance. So, how can we identify customers more likely to meet their obligations early on and use that information to shape more effective and efficient collection strategies? Traditionally, money lenders have used on-us customer behaviour data to assess accounts. This internal data — which includes contact preference, payment history, loan balance and behaviour scores — are key components in a lender’s collections strategy. However, while that information is predictive, it doesn’t provide a comprehensive view of the borrower.

Leveraging off-us consumer data to predict curability

One tool to help predict a consumer’s curability is TransUnion CreditVision^® which analyses consumer data 36 months prior to a specific point in time like 30–59 DPD. By considering data on payments, expenditures, new card applications, etc. (collected from other institutions), and credit scores, it builds a comprehensive view of consumer performance. Compared to a traditional credit history report, this deeper analysis helps predict a consumer’s likelihood to cure with higher accuracy.

Leveraging a set of CreditVision algorithms and analysis, the following insights were created:

1. The higher a consumer’s total card utilisation over the past year, the less likely they are to cure on a bankcard. If utilisation exceeds 96%, the cure rate falls to 25%.
2. Higher total debt obligations indicate lower likelihood to cure. If debt obligations exceed HK$37,000, the cure rate falls to 40%.
3. The more a consumer pays above the minimum payment obligation on non-mortgage accounts, the more likely they are to cure.

By adopting a prioritisation approach that leverages an analytics-driven model, money lenders can tailor forbearance plans, payment deferrals, term extensions, rate reductions and refinance offers to each customer’s individual situation to increase cure rates.

To explore the analysis and insights from this session, you may watch the on-demand recording and download the presentation materials here. Contact your sales representatives or click here for password.

To learn more about our CreditVision solution and how it can help improve collections rates, click here.