The new anatomy of a score – how the building blocks of credit decisioning are changing
New Zealand Credit Risk and Fraud Summit insights series – by Brian Bond, Managing Director, Credit Bureaux, Data and Analytics and Tenderlink, Experian, and Barrett Hasseldine, Head of Data Science, Experian
At the New Zealand Credit Risk and Fraud Summit, we explored how the building blocks of scoring are evolving. Bureau data continues to provide a valuable view of credit history and repayment behaviour, while transaction behaviour, affordability inputs, fraud and identity indicators, permissioned data and more targeted analytical models can add further context around risk, resilience and intent.
That is what we mean by the new anatomy of scoring. Our discussion was about how different scores, data sources and analytical inputs can work together to support a more complete assessment. For New Zealand lenders, the practical question is how to use these additional inputs in a way that improves precision, remains explainable and supports existing governance standards.
Uncertainty and risk are different problems
Bureau data is most informative when a consumer has a meaningful history of credit use. Where that history is limited, lenders may have less evidence about current financial capacity, commitments and repayment behaviour.
That challenge is relevant for several consumer groups. People who are new to New Zealand, new to credit, younger, less active in traditional credit products, or using debit and BNPL more often than credit cards may have a thinner traditional credit footprint. A limited file can make assessment more cautious, even when the consumer’s current financial behaviour may support a more nuanced view.
However, the distinction between uncertainty and risk is important. Some thin-file consumers will present higher risk, while others may have stable income, manageable commitments and evidence of disciplined financial behaviour. The practical challenge for lenders is to distinguish limited visibility from elevated repayment risk.
For lenders, richer data and more precise scoring approaches can help reduce blind spots. The objective is to maintain risk standards while giving marginal or low-visibility consumers a more proportionate assessment.
Transaction-derived insight can add current behavioural context
Transaction data can add a more current view of financial behaviour alongside bureau data. While bureau data shows how consumers have used and repaid credit over time, transaction behaviour can provide context around income regularity, account conduct, recurring commitments, buffers, volatility and signs of financial pressure.
Open banking and Consumer Data Right are expected to change the practical availability of permissioned transaction data in New Zealand. Before transaction-derived measures can support credit risk, fraud or affordability assessment, the underlying data still needs to be enriched, categorised and interpreted with enough precision for the process using it.
This becomes particularly relevant where a consumer’s traditional credit visibility has been affected by life events. In the session we spoke to a few examples of this, including the experience of consumers returning to New Zealand from abroad, consumers affected by regional flooding, consumers dealing with financial abuse, and consumers trying to access essential services after family hardship. In these circumstances, a traditional credit file may show only part of the consumer’s current financial position, and even less about their efforts and intent to improve their financial situation.
Transaction-led assessment can provide another lens on income, commitments, account conduct and potential capacity to manage essential services. A disrupted credit profile may reflect a temporary visibility issue, a recent change in circumstances or a need for more current evidence. For banks, telcos, utilities and other credit providers, transaction-derived scores may help support more proportionate assessment where traditional data provides an incomplete view.
We have recently launched a new transaction enrichment solution for New Zealand - Experian Money Tracker.
Score innovation is becoming more targeted
As the available evidence base expands, scoring can become more specific to the risk question being asked. A broad bureau score can continue to play an important role, while additional scores or score blends may help lenders assess particular behaviours, segments or treatment decisions with more precision.
Marginal application assessment may benefit from a different combination of bureau data, transaction behaviour and affordability inputs than an early-risk strategy. Thin-file assessment may require a different evidence base from fraud review. A consumer with limited traditional credit history may need a different analytical lens from a consumer with a long credit file but emerging signs of financial pressure.
In the session we also discussed “Never Pay” behaviour as an example of a more targeted scoring problem. Never Pay behaviour can differ from standard delinquency or non-payment patterns, which means it may require more specific analysis. The indicators that help identify potential Never Pay risk may differ from those used to understand ordinary arrears behaviour.
For New Zealand lenders, this points to a more precise approach to scoring – different risk behaviours, consumer segments and treatment strategies may require different combinations of bureau data, transaction-derived measures, affordability inputs, policy rules and operational controls.
The practical test is whether a scoring approach supports the decision or treatment that follows. A score used for marginal application review, fraud escalation, affordability st
ress, early loss prevention or account management should be designed, validated and monitored against that context
What New Zealand lenders should focus on next
The new anatomy of scoring is about strengthening the evidence base behind credit assessment. For New Zealand lenders, three priorities stand out.
- First, understand what each scoring input is helping to answer. Bureau data, transaction behaviour, affordability, fraud, identity and policy rules contribute different forms of evidence, and their value depends on how they improve the assessment being made.
- Second, separate uncertainty from risk. Thin-file consumers, new-to-country applicants, marginal applications and consumers affected by life events may require a more current view than traditional credit data alone can provide.
- Finally, start exploring how you could connect innovative scores to the decisions or treatments it is designed to support. Marginal applications, affordability stress, fraud review, early loss prevention and Never Pay behaviour may each require a different combination of inputs, controls and monitoring.
While scores remain central to credit assessment, the way they are built, combined and interpreted is evolving. The opportunity for New Zealand lenders is to use richer data and more targeted analytics to support more informed, explainable and proportionate assessments.
To find out how Experian is helping New Zealand organisations strengthen credit decisioning with connected data, analytics and scoring capability, please get in contact through the form below.