Spotting risk earlier by tracking consumer credit journeys

How we're investing in data and analytics in consumer finance

Our goal is regulation that is evidence-based, targeted, and achieves good outcomes for consumers. That’s why we’ve been using richer datasets and sharper data science to drive better outcomes in the consumer finance market, widen financial inclusion, and support economic growth. 

This blog post explains one way we've been doing that, in a proof-of-concept undertaken by the team of Isabela Barra, Daniel Bogiatzis-Gibbons, Lawrence Charles, and Wenjin Li (detailed results in the Technical Annex (PDF)[1]). 

We draw on credit file information from a major Credit Reference Agency (CRA), an existing source of data that the FCA has been using since 2018. We apply novel advanced statistical methods to draw new insights from it on which consumers are likely to fall into distress on their credit products and at what time. We can do this given we have a wider market view than an individual firm’s data sources.

This means we can:

• Spot patterns that reveal emerging or disproportionate harm among consumer groups based on past performance data.
• Sharpen our focus on affordability and vulnerability for consumers, separating momentary blips from ongoing strain.
• Get ahead of risks with earlier, more targeted supervision and timely engagement with firms.

This focus gives us a market-wide view related to our rules on strengthening protections for borrowers in financial difficulty. These include requirements for firms to offer supports before customers fall into arrears, for it to be tailored, and to pay attention to customers with vulnerabilities.

In future, we will use our product sales data (PSD) on credit agreements in data science projects. The PSD will further help us to plot trends in consumers' engagement across different credit products and identify triggers of financial distress across different consumer groups. The PSD will also have more comprehensive coverage than CRA data when fully operational.

What we are doing: Looking at whole credit journeys, not just snapshots

New datasets and uses of existing ones are driving exciting improvements in our analysis. Traditional credit indicators include delinquency rates, credit scores, and payment histories. They tend to flag problems after they have already crystallised. 

However, they often miss:

• Direction – whether a person’s financial position is getting stronger or weaker.
• Velocity – how quickly the stress is building.
• Persistence – whether early signs of stress fade or worsen.
• Combined signals – when several modest, individual changes that may be manageable on their own add up to increase risk.

What’s different about our new approach is that it tracks how people move between different states of financial stability, emerging stress, and acute distress. By spotting those common patterns in consumers’ credit journeys early, it helps us prioritise groups of people and firms where financial stress is emerging.

We assign each person to one of 5 segments at a given time:

1. Distress (about 5% of users) – severe credit issues such as going bankrupt or falling more than 3 months behind on credit payments.
2. At Risk (about 5% of users) – early warning signals (for example, recent missed payments, a high level of usage of their available credit limit, or multiple new unsecured accounts (like extra credit cards).
3. Secured Credit Users (about 1 in 3 users) – at least one active mortgage and stable use of credit.
4. Unsecured Credit Users (about 1 in 5 users) – active users of multiple unsecured products with stable behaviour.
5. Low Credit Engagement (about 1 in 3 users) – limited or no use of formal credit.

Using these definitions, we can see transitions between the 5 segments. See figure 3 in our Technical Annex (PDF)[1]. Most consumers remain stable, but there are clear flows from At Risk into Distress, and some recovery back to Unsecured or Secured.

These transitions show that distress rarely just appears without warning signs. It usually comes after a period of instability, such as rising balances or missed payments. Equally, recovery is uneven. Some people stabilise quickly, while others remain in difficulty for longer.

Identifying who is at risk is only part of the job. Timing also matters. We use what statisticians term 'survival analysis' to estimate how long someone is likely to remain financially stable and identify what factors change that timeframe.

Using this analysis to take a forward-looking view across the entire consumer population shows that:

• Individuals in the Low Credit Engagement and Secured Credit groups remain financially stable the longest.
• The At-Risk group have the shortest period of financial stability.
• Having recent missed payments, multiple new unsecured credit accounts, or increasing use of a person’s available credit limit is associated with moving into the Distress group faster.

What’s next: How analytics supports our consumer finance goals

Building on our work here, we will monitor how consumer journeys in credit develop over time. It will help us understand how people are accessing credit products. Then, we can proactively identify potential risks, allowing us to target supervision more effectively. 

That’s how we can help people weather changes in their financial circumstances and navigate their financial lives. For example, we recognise that consumers’ use of credit is evolving all the time. So, we will incorporate Deferred Payment Credit (DPC, often known as Buy Now Pay Later), products in future iterations of this analytical work. 

We’re keen to join forces with academics and tech innovators exploring credit file data to drive consumer outcomes. Please contact [email protected]. We also welcome continued input from financial firms and consumer groups through our Consumer and Practitioner Panels. 

Together, we can work to spot risk earlier, focus support where it helps most, and maintain credit markets that work well for consumers who rely on them.