How Bank CROs Can Use Data to Predict Bad Loans

25 February 2025

How Bank CROs Can Use Data to Predict Bad Loans

Bank chief risk officers (CROs) have one job which they can’t afford to get wrong – identifying bad loans before they become a problem. In the US, as of 2023, less than one percent of bank loans were non-performing or bad loans. This is a good sign, highlighting the fact that most people were paying back their debts in time. 

However, US banks are currently dealing with a shortage of analysts who can spot bad loans. Thus, CROs need to step in and oversee this side of things themselves at times. For CROs, it’s no secret that lending comes with risk, but in today’s digital world, relying on gut feelings just won’t cut it. Data gives CROs the power to foresee potential trouble and take preventive action before losses start piling up. Here’s how. 

Using Loan Origination Intelligence for Smarter Lending

When it comes to approving loans, the loan origination process is where it all begins. This stage is crucial because every piece of data collected during loan origination can shape a bank’s future financial health. 

Commercial real estate lenders are already relying on loan origination intelligence to make informed decisions. Such technology doesn’t just help with crunching real estate figures. In the right hands, it paints a vivid picture of market trends, allowing real estate professionals to make data-driven decisions.

According to Blooma, such technology allows CRE lenders to identify high-potential borrowers and projects. At the same time, it helps them identify risky borrowers as well. Bank CROs can get the same benefits from using data and loan origination intelligence. 

Loan origination intelligence pulls insights from vast datasets, including borrower spending habits, employment stability, and even market conditions. In the mortgage industry, for example, lenders no longer rely solely on traditional credit reports. They’re looking at alternative data points, such as utility payments and rental history, to assess borrower reliability. The more comprehensive the data, the clearer the risk assessment.

How AI and Machine Learning Are Changing the Game

Artificial intelligence and machine learning have transformed the way banks predict loan defaults. Traditional risk models used static criteria – income levels, debt-to-income ratios, and credit scores – to assess borrowers. While these factors are still important, they don’t always tell the full story. AI-driven models analyze patterns that human analysts might miss.

For instance, machine learning algorithms can track subtle shifts in spending behavior that may signal financial distress. If a borrower starts maxing out credit cards or missing small payments, AI can flag these behaviors as potential warning signs. This gives banks an early heads-up to step in before the situation escalates.

Even more impressive is AI’s ability to adapt in real-time. Unlike traditional risk models that need manual updates, machine learning systems continuously refine themselves based on new data. That means the more loans a bank processes, the smarter the AI becomes at detecting bad ones.

External Data and Market Trends

In the US, a chief risk officer makes around $162,111 per year on average. The workload they take on, of course, demands such a high salary. A part of their workload in the modern day is to keep an eye on external data and market trends.

Today, predicting bad loans isn’t just about looking at individual borrowers; it’s also about understanding the broader economic landscape. External data sources, such as market trends, unemployment rates, and industry-specific risks, can provide valuable context. 

Real estate markets are another critical factor. If housing prices are declining in a specific area, mortgage lenders need to reconsider the long-term value of properties being used as collateral. Even geopolitical events and inflation rates can impact a borrower’s ability to repay loans. 

Taking Action Before a Loan Goes Bad

Early intervention strategies, such as restructuring payment plans or offering financial counseling, can prevent defaults from happening in the first place. Some banks are even using predictive analytics to create personalized repayment solutions.

If a borrower is showing early signs of financial trouble, the bank can offer tailored solutions like temporary lower interest rates or deferred payments. This approach helps the borrower and also reduces losses for the bank.

The key is to act before a loan becomes a non-performing asset. With data-driven insights, banks don’t have to wait for a borrower to miss multiple payments before stepping in. Instead, they can identify warning signs early and address issues before they turn into full-blown defaults.

The Future of Risk Management in Banking

Here’s the thing: US bad loans stood at $170.521 billion as of June 2024. Hence, bank CROs have a lot on their plate already. But, as data analytics continues to evolve, bank CROs will have even more sophisticated tools at their disposal. 

The days of relying solely on credit scores and financial statements are long gone. Now, it’s about leveraging technology to build smarter, more resilient lending practices.

The future of risk management is proactive, not reactive. Predicting bad loans is no longer a guessing game; it’s a science, and it’s one that banks can’t afford to ignore.

Leadership Resources

For more leadership resources look at our great-value guides. These include some excellent tools to help your personal development plan. The best-value approach is to buy our Leadership bundle, available from the store.

We’ve bundled together these five e-guides at half the normal price! Read the guides in this order, and use the tools in each, and you’ll be well on your way to achieving your personal development plan. (6 guides, 167 pages, 27 tools and 22 insights, for half price!)

• Leadership Essentials
• Defining Leadership
• Leading Insights
• Leading with Style and Focus
• Transformational Change
• Making Change Personal

>> Return to the Leadership Knowledge Hub