![](https://www.equifax.com.au/knowledge-hub/sites/default/files/intro%20image%20repayment%20history%20%282%29.png)
Once considered a primary data set for predicting borrowers' future financial resilience, recent default and comprehensive credit reporting data (CCR) is not the only data set lenders will need to make accurate, reliable credit decisions.
Lenders looking to manage this shifting dynamic should move quickly to ensure their scoring model has been updated to reflect this new normal. The latest generation Equifax credit score, Equifax One Score, has been developed to do just that. To not only incorporate the learnings of this pandemic, but to provide lenders with a model that offers considerable improvements in predictive power and coverage when compared to earlier generation Equifax credit scores.
A stable and highly predictive credit score is an essential asset to ensure that loan origination and underwriting processes are not adversely affected by the ongoing impact of coronavirus-induced loan deferrals and expected rise in delinquencies.
Under COVID-19 simulation conditions, Equifax One Score has shown its superiority over our previous scores in helping lenders make more precise lending decisions. Here's how the analysis was conducted:
- Up to 30% of the payment history data from all consumers applying for credit in the development sample was removed.
- Accounts were reset to '0' to simulate a payment deferral or '.' missing if the loan was already in arrears to simulate a payment deferral that is now in hardship.
- All defaults were removed over the prior ten months leading up to the point of application/score generation.
- Equifax One Score's performance was compared to our previous generation score using actual outcomes 12 months later.
The results demonstrate that both Equifax scores were robust, but Equifax One Score experienced a Gini drop of only 0.9% compared to 1.6% for our previous generation score.
With this more accurate gauge of a borrower's financial resilience, a lender may be in a position to manage their credit risk more effectively while better serving clients as they journey through very different recovery trajectories in the months ahead.
Managing recent CCR data limitations
Pre-pandemic, the most recent period of a borrower's CCR data was considered the most predictive of future behaviour. Post-pandemic, this same data comes littered with repayment gaps and fraught with uncertainties about individual borrowers' resilience.
Designed to cope with these limitations, Equifax One Score places less emphasis on recent data. Instead, the model incorporates the full history of data – from six months to 24 months – so that it is less sensitive to data that may no longer be useful.
By drawing on recent and longer-term repayment history trends, as well as up to five years of credit enquiry and defaults trends, Equifax One Score provides a robust tool for lenders. Its superior predictive power, when compared to previous Equifax scores, is particularly vital when managing and monitoring credit risk in an environment where the visibility and access to reliable data may be compromised.
Adding more pieces to the data puzzle
No one wants to start on a jigsaw puzzle and then discover pieces are missing. It's the same concept with a credit model – when data is missing from a credit file, the model might negatively interpret this because it doesn't know what else to do.
By adopting the premise that the world is never perfect, Equifax One Score has been designed to cater to these inevitable data gaps. The model understands that it may not receive every data element needed to complete the puzzle. Still, with enough pieces, it can start to form a picture of a borrower's credit behaviour. Rather than being sensitive to any single data point that is missing, our new model bases its risk assessment on a broad range of data and the length of time information has been in the credit file.
Using a model that taps into a vast collection of data can help lenders effectively identify higher-risk borrowers from lower risk borrowers. In turn, this enables lenders to avoid unexpected credit risk while simultaneously extending credit responsibly to more consumers.
Equifax is the market leader in consumer credit information, holding the most extensive collection of data, including CCR data and Buy Now Pay Later data, than any other Australian credit bureau.
In Equifax One Score, the latest CCR data augments our extensive repositories of negative event and enquiry data. In certain circumstances, geo-demographic data is also fed into the score, providing the opportunity to form inferences about credit behaviour according to where a consumer lives.
Improving analytical capability
With the adoption of ground-breaking machine learning analytical techniques, it's now possible to reach wider and deeper into data to look at multiple interactions with many variables.
Equifax One Score uses an explainable AI (xAI) machine learning scoring methodology known as NeuroDecision™ Technology (NDT) to handle the complexity of data residing in different sources. It does so with accuracy, transparency and explainability.
The NDT algorithm rewards positive behaviour (score increases) and penalises negative behaviour (score decreases). When a borrower re-pays debt every month on time, NDT identifies the action as positive and increases the score. Importantly, it generates a logical and actionable explanation of why a consumer would have a low credit risk score which may lead to denied credit. This explanation is crucial for lenders striving to remain compliant and make informed decisions when assessing credit risk in an uncertain environment.
Continual innovation
The rollout of open banking brings with it many new alternative data sets that can add another layer of information to existing credit reporting. Beyond Equifax One Score's release, we will continue to develop score overlays and new generation scores that incorporate CCR data and additional demographic and utility data. We are currently building a predictive model prototype based on transaction data, for example, that searches for previously unrecognised positive and negative correlations in income and expense data.
This continual innovation is made possible by Equifax's A$2 billion global transformation program that has considerably improved our deployment capability. Having migrated to a new, multi-Cloud environment, we are leveraging this architecture to deliver analytics to production within a benchmark of just 30 days. Even very complex models can now be packaged up as a full service and exported once to multiple channels, using rapid testing to attain a faster turnaround than ever achieved by a traditional bureau.
Just as we have incorporated the learnings of the COVID-19 pandemic into our new generation score, we have the data management and analytics capabilities to move quickly to ensure our future generation scores reflect ever-changing consumer credit behaviours. For lenders, this translates into accurate, timely credit underwriting and monitoring decisions both in and out of downturns and changing economic conditions.
With Equifax One Score, lenders can put aside concerns about repayment history gaps and focus instead on the business opportunities which come with this powerfully predictive analytic tool. Speak to your account manager or email us to find out more.
Related Posts
The ongoing housing crisis remains a hot topic and one of the major concerns for many Australians. Figures from Equifax1 confirm a troubling trend: home ownership in Australia becomes a progressively distant aim for younger people.
![Data cleansing, fixing spelling errors in customer names, removing deceased individuals from datasets, identifying duplicate records or reconciling inconsistencies are key for high quality data.](https://www.equifax.com.au/knowledge-hub/sites/default/files/styles/intro_image/public/field/image/Deathcheck%20BLOG%20article%20image_0.jpg?itok=jRvaE6Rp)
Businesses collect and process vast amounts of data, and often overlook the importance of ensuring the data is accurate, reliable, relevant, complete, consistent and timely. For a data-critical sector like insurance, the process of cleaning data to fix inaccuracies, incomplete, duplicate or wrongly formatted records holds valuable benefits.