Banks’ Credit Algorithms Are Quietly Pushing Americans Deeper Into Debt

New research from King’s Business School at King’s College London and the Federal Reserve Board is raising serious questions about how modern banking algorithms shape consumer debt in the United States. The study shows that most credit card limit increases are no longer something consumers actively ask for. Instead, they are automatically triggered by bank algorithms, often for people who are already carrying debt. While these increases may look helpful on the surface, the evidence suggests they are playing a major role in expanding household debt across the country.

At a time when Americans are preparing for peak spending seasons and relying heavily on credit cards, this research offers a closer look at how behind-the-scenes automated decisions influence borrowing behavior in ways many people barely notice.

Most Credit Limit Increases Are Initiated by Banks, Not Consumers

One of the most striking findings of the study is how rarely consumers actually request higher credit limits themselves. According to the data, around four out of every five credit limit increases in the United States are initiated by banks, not cardholders. These increases are applied automatically, usually without a direct request or negotiation from the consumer.

In practical terms, this means millions of people wake up one day to find their available credit has expanded, even though they never asked for it. These automated increases now account for more than $40 billion in additional available credit every quarter, a massive and ongoing expansion of borrowing capacity across the U.S. economy.

Algorithms Target Borrowers Who Are Already in Debt

The research shows that banks are not randomly increasing credit limits. Instead, their algorithms are far more likely to target borrowers who already carry balances on their credit cards. Rather than rewarding people who consistently pay off their balances in full each month, the systems tend to focus on those who are already revolving debt.

This targeting has major consequences. The study finds that about one-third of all unpaid credit card balances in the United States exist only because credit limits were increased after the card was opened. For borrowers with lower credit scores, the number is even more alarming, rising to around 60%. In other words, without these automatic limit increases, a large portion of today’s credit card debt simply would not exist.

Borrowers Spend More When Limits Increase

When credit limits go up, borrowing tends to follow. The study finds that consumers respond to these automatic increases by raising their revolving balances by roughly 30%. This behavior highlights an important psychological and financial reality: available credit often feels like permission to spend more, even when repayment becomes harder over time.

From the banks’ perspective, this makes sense. Revolving balances generate interest payments, which are a key source of profit. From the consumer’s perspective, however, it can mean higher monthly payments, more interest costs, and increased financial vulnerability, especially during economic downturns or unexpected expenses.

AI and Machine Learning Are Deeply Involved

Another notable finding links automated credit limit increases to banks’ growing use of advanced technology. The researchers observed that banks which publicly advertise their use of artificial intelligence and machine-learning tools in official financial reports are also more likely to rely on automated systems to raise credit limits.

These systems are designed to predict which customers are most likely to borrow more if given extra credit. While such models can improve efficiency and expand access to credit, the study suggests they also quietly amplify debt among already-indebted households, often without borrowers fully understanding how or why these decisions are made.

How Other Countries Handle Credit Limit Increases

To better understand the broader impact of these practices, the researchers used a detailed model of household spending and borrowing to compare U.S. policies with those in other countries.

In the United Kingdom, banks are not allowed to raise credit limits for customers who are already in debt without their explicit consent. Canada goes even further, requiring consumer approval for any credit limit increase, regardless of the borrower’s financial status. These rules are designed to ensure that consumers remain aware and in control of changes that could affect their financial health.

The study also notes that the European Union plans to implement similar regulations next year, signaling a growing international consensus around stronger consumer protections in automated lending decisions.

Modest Regulation Could Improve Consumer Welfare

Using their economic model, the authors tested what might happen if the United States adopted safeguards similar to those used in the UK or Canada. The results suggest that consumer welfare could improve by around 1% overall, a meaningful gain at a national scale.

These policies would likely reduce revolving debt balances and lower the share of household income devoted to interest payments. Importantly, the researchers found that such rules would have only a modest impact on overall credit availability, countering concerns that regulation would severely restrict access to credit for consumers who genuinely need it.

A Massive Dataset Behind the Findings

The strength of the study lies in its data. The researchers relied on detailed regulatory microdata covering more than 70% of the U.S. credit card market, collected through the Federal Reserve’s Capital Assessments and Stress Testing framework. This allowed them to quantify, for the first time, the direct welfare impact of automated credit limit increases and to measure how policy changes could reshape consumer outcomes.

Why This Matters for Everyday Credit Card Users

For most consumers, credit limit increases feel harmless or even beneficial. More available credit can help smooth spending, cover emergencies, or improve credit utilization ratios. However, this research highlights a less visible side of the equation. When algorithms automatically raise limits for people already in debt, the result is often more borrowing rather than more financial flexibility.

The study does not argue that automated credit limit increases are inherently bad. Instead, it suggests that context matters. Used carefully, these tools can support households. Used aggressively, they can quietly deepen financial strain.

Understanding the Bigger Picture of Algorithmic Finance

This research fits into a broader conversation about how data-driven decision-making is reshaping modern finance. Algorithms now influence everything from loan approvals to interest rates and credit limits. While they can reduce bias and improve efficiency, they also raise questions about transparency, consent, and consumer protection.

As financial institutions continue to rely on increasingly sophisticated models, studies like this one highlight the importance of thoughtful regulation that balances innovation with consumer well-being.

Research paper: https://www.federalreserve.gov/econres/feds/files/2025088pap.pdf