When an unexpected medical emergency strikes, households without deep savings often face a painful choice: exhaust their liquid funds, leave hospital bills unpaid, or turn to high-interest credit cards. Intuition suggests that expanding public health insurance would naturally reduce consumer borrowing. If the government covers doctor visits and hospital stays, families should theoretically carry less unpaid debt on their monthly credit card statements.
Yet real-world financial behavior often defies simple assumptions. According to an NBER working paper, expanding Medicaid under the Affordable Care Act actually led to a sustained increase in credit card borrowing. Rather than discouraging the use of credit, the injection of social insurance appeared to make households more financially resilient, prompting lenders to loosen borrowing limits and lower interest rates.
The Theoretical Puzzle of the Social Safety Net
The study was authored by Gideon Bornstein and Sasha Indarte, economists at the Wharton School of the University of Pennsylvania. They set out to resolve an ongoing question in macroeconomics: does public social insurance crowd out private borrowing, or does it crowd it in?
On one side of the ledger, social programs can lessen a household’s reliance on debt by directly absorbing large expenditure shocks. On the other side, receiving insurance alters both household incentives and lender behavior. With basic medical risks covered, families may feel less need to hold precautionary cash savings. At the same time, households that are less likely to default become far more attractive to banks. If lenders respond to this lower default risk by extending cheaper loans and higher borrowing limits, overall debt can rise.
To understand which dynamic dominates, the researchers analyzed how unsecured credit changed following the rollout of the Affordable Care Act. In participating states, the reform expanded Medicaid eligibility to all adults earning up to 138% of the federal poverty level, removing previous restrictions that had largely excluded adults without dependent children.
Tracking Ten Million Credit Files
Determining the causal impact of Medicaid eligibility is challenging because low income is already linked to reduced credit access. A simple comparison between eligible and ineligible individuals would confuse the effects of low earnings with the effects of health insurance.
To bypass this obstacle, the researchers designed a continuous difference-in-differences strategy. They combined credit bureau records from Experian covering a representative sample of 10 million Americans between 2010 and 2021 with Internal Revenue Service income data across more than 11,000 ZIP codes. This approach exploited two separate sources of variation: differences in state rules prior to the expansion, and variations in how many residents in a given neighborhood fell into the newly eligible income bracket.
By comparing credit outcomes within counties across neighborhoods with higher versus lower exposure to the policy change, the researchers were able to absorb the effects of broader regional economic shifts. They focused specifically on revolving balances, which represent unpaid credit card debt carried from month to month that accrues interest, rather than total balances paid off at each billing cycle.
Higher Balances, Better Credit Scores
The empirical findings revealed a clear rise in borrowing. The authors estimated that a 1 percentage point increase in a ZIP code’s Medicaid-eligible population increased revolving credit card borrowing by 0.38%. For each individual who became newly eligible, this translated to an average borrowing increase of $1,446 spread over several years.
Card adoption also grew. A 1 percentage point rise in eligibility led to a 0.08 percentage point increase in the share of residents holding at least one credit card, alongside a 0.31% increase in average credit limits. These borrowing increases were concentrated in lower-income ZIP codes, where families previously had the least access to mainstream credit.
Despite taking on more debt, borrowers did not show signs of heightened financial strain. Instead, measures of distress fell across the board. The likelihood of having any debt 90 or more days delinquent declined by 0.02 percentage points, while 90-day delinquency on utility bills dropped by 0.13 percentage points. Third-party collection balances fell by 2.5% relative to income, and average credit scores rose by 0.21 points per percentage point of eligibility.
When the authors inspected individual-level credit records, they found that effective annual percentage rates, the true fee-inclusive interest rates paid by cardholders, dropped by 4.9 percentage points for newly eligible borrowers. This sharp drop in borrowing costs provided initial evidence that the surge in debt was driven primarily by banks extending cheaper credit rather than desperate consumers maxing out their accounts.
Inside the Economic Engine: A Structural Model
While the empirical data demonstrated that borrowing rose across exposed neighborhoods, cross-sectional comparisons cannot capture nation-wide equilibrium shifts. To evaluate aggregate effects and welfare, the authors constructed a heterogeneous-agent macroeconomic model.
In this simulated economy, households face unpredictable income swings and medical emergencies. They can save in a risk-free asset or borrow through credit cards, and they hold the option to renege on debt obligations by entering delinquency. Delinquent borrowers suffer a loss in utility and cannot borrow anew, but their debts are subject to stochastic haircuts, which represent negotiated settlements or bankruptcy write-downs.
The model categorized households into three distinct behavioral types:
- Strategic defaulters: Making up 48% of the population, these relatively impatient households borrow despite noticeable interest rate spreads and choose delinquency when debt burdens become unmanageable relative to their earnings.
- Credit-excluded households: Comprising 42% of the population, these individuals face no personal utility penalty from delinquency. Anticipating non-repayment, competitive lenders deny them credit, leaving them entirely outside the credit market.
- Patient liquidity-driven defaulters: Representing 10% of households, this group holds substantial liquid savings and defaults only when catastrophic medical shocks overwhelm their resources.
Why Credit Supply Tells the Whole Story
Using this framework, the authors simulated the long-run impact of expanding Medicaid. When the researchers replicated their neighborhood-level analysis inside the model, it reproduced the empirical borrowing response (a coefficient of 0.42 versus 0.38 in the data), and it projected an overall 6.7% increase in aggregate credit card debt. It also allowed the researchers to isolate three competing mechanisms.
The first was the direct effect. Holding borrower behavior and lender pricing fixed, health coverage reduced out-of-pocket medical bills, lowering borrowing by 0.49%. The second was the credit demand channel. While a weaker precautionary savings motive encouraged borrowing, a countervailing force called debt aversion dominated: because households expected to default less often, debt without repricing effectively felt more expensive to repay, shrinking borrowing by 5.41%.
The credit supply channel completely overturned these contractions. Because households were far less likely to default, competitive lenders reduced their interest rate spreads. Cheaper borrowing terms expanded effective credit lines, lifting aggregate credit card debt by 12.63%. In short, without the positive response from lenders, social insurance would have caused household debt to contract.
The model also shed light on human welfare. Under a budget-neutral expansion funded by a modest flat income tax of 0.02%, the policy generated an average net welfare gain equal to a permanent 0.22% increase in consumption. The authors found that 30% of this net welfare gain stemmed directly from the credit supply channel. Lower interest rate spreads allowed households to smooth their spending across both good and bad years far more effectively.
Policy Considerations and Caveats
These findings point to an overlooked mechanism in consumer finance: public safety nets can actively improve private market access for lower-income households. By protecting borrowers from medical bankruptcy, social insurance lowers the risk profile of low-income applicants, enabling financial institutions to offer credit on terms previously reserved for higher earners.
The authors note several boundaries to their analysis. The welfare calculations reflect financial gains from consumption smoothing and do not account for physical health benefits or mortality reductions, which are substantial in their own right. The model also abstracts from potential behavioral biases, such as overspending driven by poor self-control. The empirical evidence that credit scores improved and collections fell suggests that the additional debt did not push households into deeper financial distress.
Additionally, the credit expansion did not extend uniformly to all loan categories. While auto loans saw a measurable bump in participation and balances, mortgages remained unchanged, likely because low-income households remain farther from the margin of homeownership. For policymakers designing social insurance programs, the research demonstrates that the economic reach of a safety net extends well beyond the hospital door, reshaping how households and banks interact in everyday financial markets.
