More Credit, More Debt: New Evidence on Automated Credit Decisions

January 16, 2026

More Credit, More Debt: New Evidence on Automated Credit Decisions

Vitaly M. Bord, Agnes Kovacs, and Patrick Moran

Behind the scenes of every credit card lies an increasingly complex algorithmic infrastructure that determines who receives more credit and when, largely outside the inspection or knowledge of credit card users. Credit card issuers deploy sophisticated algorithms that continuously analyze consumer spending and borrowing behaviors, often increasing credit limits without consumers requesting such changes. While regulations in the United States allow lenders nearly unfettered ability to raise credit limits, other countries such as Canada and the UK have begun to place restrictions on bank-initiated credit limit increases. These developments raise two central questions: who do lenders target with limit increases and under what conditions are regulations restricting limit increases beneficial for consumers?

In this note and the accompanying paper, Bord, Kovacs, and Moran (forthcoming), we use regulatory data to answer these questions. First, we show that limit increases are prevalent, and that the majority of increases in the U.S. are bank-initiated. Next, we examine the targeting patterns behind these algorithmic credit decisions and find that banks appear to give more increases to revolving borrowers who carry balances month-to-month. Finally, we explore the implications for consumer welfare of these practices and of restrictions on limit increases that some countries have begun to implement.

Limit Increases and Low-and-Grow Strategies

Our analysis of Federal Reserve Y-14M regulatory data, which covers more than 70 percent of the U.S. credit card market, reveals that limit increases are prevalent in the U.S. About 12 percent of credit cards receive limit increases annually, totaling, in aggregate, about $160 billion dollars of new available credit each year. These limit increases account for approximately half as much additional available credit as new account originations. Importantly, the overwhelming majority-approximately 80 percent-of increases initiated by banks rather than requested by consumers.

The prevalence of limit increases in the U.S. is consistent with lenders implementing "low-and-grow" lending strategies, where they issue credit cards with low initial limits to lower credit-score borrowers and then increasing the limits based on borrower behavior. As shown in Figure 1, credit card limits at origination for accounts of subprime borrowers with credit scores below 600 are only $700, on average. However, by five years after origination, accounts of subprime borrowers have a limit of $2700, on average, a 285 percent increase. By contrast, accounts of superprime borrowers with credit scores above 760 start off with a much higher limit of $12,000 at origination, on average, and grow 25 percent to about $15,000 by five years later.

Bank-initiated limit increases and low-and-grow strategies are prevalent likely because the bank has limited information about a new borrower prior to issuing a credit card. When a new customer applies for a credit card, the lender has information on their credit score, their credit report, and self-reported income. However, if the lender approves the card and the borrower begins using it, the lender obtains extensive information about all aspects of credit card usage, spending, and payments, which the lender can then use to adjust the limit. While we do not observe the algorithms used by lenders to grant limit increases, our data allow us to observe the borrower and account characteristics that are associated with limit increases.

Limit Increases and Algorithmic Targeting of Revolving Borrowers

We find that one important aspect of credit card usage is revolving behavior-that is, whether the borrower carries balances month-to-month and pays interest. Existing literature has shown that revolvers account for the majority of the profitability of credit cards (Adams et al, 2022). This suggests that lenders may be more likely to give limit increases to revolvers since they are the most profitable customers.

Figure 2, Panel A, shows that revolvers are indeed much more likely to receive a bank-initiated limit increase than transactors, i.e. credit card customers who always pay off their credit card. While only 2 percent of transactors receive a bank-initiated increase each quarter, almost 4 percent of revolvers do. By contrast, Panel B shows that the share of credit card users who ask and receive a borrower-initiated limit increase is the same for revolvers and transactors. This suggests that while lenders target revolvers for increases, revolvers do not ask or receive limit increases at a higher rate.

In Bord, Kovacs, and Moran (forthcoming), we find that not all revolvers are targeted for bank-initiated limit increases in the same way. Our analysis shows that the extent of revolving borrowing matters for the probability of receiving a limit increase. In particular, borrowers who revolve between 30 and 70 percent of their credit limit have the highest probability of receiving a limit increase, all else equal, while those with very high or very low revolving utilization are less likely to receive an increase. We also examine other factors that matter for the probability of receiving a limit increase including credit score, income, and other borrower and credit card characteristics. Notably, we find that the increase in the probability of receiving a limit increase from revolving 30-70 percent of the credit limit is approximately equivalent to the increase in the probability associated with a 60-point increase in credit score.

Consumer Behavior After Limit Increases

Consistent with previous studies, we find that consumers tend to increase their borrowing after receiving a credit limit increase, even when they are not near their previous limit (Gross and Souleles, 2002; Aydin 2022). Figure 3 shows revolving utilization for accounts that receive a bank-initiated credit limit increase, relative to accounts that do not, during the months before and after the limit increase. Time t=1 is the first month after the increase. The figure shows that revolving utilization initially declines following a limit increase but then recovers to its previous level within about 6 months. Since the limit is now higher, the same rate of revolving utilization means that the credit card user has increased their debt following the increase in their credit limit, with the increase in revolving balances comprising about 30 percent of the credit limit increase. Consistent with existing literature, we find that this effect persists even among unconstrained accounts with low utilization rates prior to the limit increase (Gross and Souleles, 2002).

Policy Implications

Why do credit limit increases lead to more borrowing? A large literature has pointed to the importance of self-control issues in credit card purchase and payment behavior (Meier and Sprenger, 2010; Gathergood, 2012; Kuchler and Pagel, 2021). Our findings align with this research and suggest that if some consumers struggle with self-control, algorithmic limit increases may encourage additional borrowing and thus be profitable for banks but potentially harmful to vulnerable consumers who end up with more debt than they would have chosen otherwise.

Policy-makers in some countries have begun to respond to these concerns. The UK now prohibits limit increases for borrowers in persistent revolving debt, while Canada requires explicit consumer consent for all limit increases. The EU will implement similar consumer consent requirements across member states in 2026.

These policies recognize the tension between the benefits and drawbacks of algorithmic credit decisions. On the one hand, algorithmic credit limit increases may allow for greater access to credit and may help borrowers to better withstand financial shocks, as they can use their credit limits to borrow during hard times. On the other hand, as Figure 3 suggests, limit increases may also lead to excessive debt, especially if lenders target accounts of credit card users who already revolve balances.

In Bord, Kovacs, and Moran (forthcoming), we show that under certain conditions, policies that restrict bank-initiated credit limit increases may be beneficial for credit card users, on balance. Figure 4 illustrates the welfare effects of implementing a policy that prohibits credit limit increases for consumers who have been in persistent revolving debt for more than 12 months, similar to recent regulation implemented in the UK. The figure shows that the overall effect of the policy is positive, as the restriction has large positive effects on "tempted" consumers who struggle with self-control issues and who take on much more debt when granted a limit increase (far right bar). This positive effect outweighs the slight negative effect for "standard" consumers without self-control issues (middle bar) who do not receive the benefits of greater access to credit. On balance, the policy's consequences are positive under the assumptions of the model discussed in Bord, Kovacs, and Moran (forthcoming). In the paper, we show that the effects are similar for a Canada-type policy that stipulates that all limit increases need to receive consumer permission.

References

Adams, Robert, Vitaly M. Bord, and Bradley Katcher, "Credit Card Profitability," 2022. Feds Notes.

Aydin, Deniz, "Consumption Response to Credit Expansions: Evidence from Experimental Assignment of 45,307 Credit Lines," American Economic Review, January 2022, 112 (1), 1-40.

Bord, Vitaly, M., Agnes Kovacs, and Patrick Moran. Forthcoming. "Automated Credit Limit Increases and Consumer Welfare. Journal of Monetary Economics.

Gathergood, John, "Self-control, financial literacy and consumer over-indebtedness, Journal of Economic Psychology, 2012, 33 (3), 590-602.

Gross, David B. and Nicholas S. Souleles, "Do Liquidity Constraints and Interest Rates Matter for Consumer Behavior? Evidence from Credit Card Data*," The Quarterly Journal of Economics, 02 2002, 117 (1), 149-185.

Kuchler, Theresa and Michaela Pagel, "Sticking to your plan: The role of present bias for credit card paydown," Journal of Financial Economics, 2021, 139 (2), 359-388.

Meier, Stephan and Charles Sprenger, "Present-biased preferences and credit card borrowing," American Economic Journal: Applied Economics, 2010, 2 (1), 193-210.