For their paper, “The Opioid Epidemic and Consumer Credit Supply: Evidence from Credit Cards,” Senior Economic Advisor and Economist Wenli Li and Senior Financial Economist Raluca Roman of the Philadelphia Fed, Professor Sumit Agarwal of the National University of Singapore, and Assistant Professor of Business Nonna Sorokina of the Pennsylvania State University studied how the U.S. opioid crisis has impacted the quantity of credit and other credit terms offered by bank lenders to consumers in severely opioid-affected areas.

More than 1 million people have died of opioid overdoses in the United States since 1999, the authors explain. That’s more than from car accidents during the same period.1 The opioid epidemic has intensified over time, with minorities, young men, and those with low levels of education disproportionately affected. For these groups especially, the opioid crisis has reduced labor force participation, increased unemployment, and reduced income.2 In turn, the crisis has impacted these individuals’ consumer finances and their credit risk. For example, opioid users — particularly those who fund their drug addiction using credit — are more likely to default on loans.

In general, credit card debt is attractive to consumers, the authors explain, as credit cards do not require collateral, unlike other types of loans. In the United States, there are more than 175 million credit cards users, representing 80 percent of all consumers.3 On the flip side, bank lenders often lack the information they need to evaluate an individual’s default risk. Since credit card debt is an important determinant of overall credit risk, increases in defaults on credit cards, including among opioid abusers, can harm lenders’ portfolios.

The authors used a data set of credit card offers that banks mailed to consumers between 2010 and 2019, a period characterized by perhaps the most serious abuse of prescription and illicit opioids.4 These credit offers, the authors note, “are a direct informative measure of consumer credit supply by the banks, helping circumvent challenges of disentangling supply from demand forces that plague other studies.” Also, they measured the extent of the crisis using opioid-related death rates by county from the Centers for Disease Control’s National Center for Health Statistics.

The authors investigated two related questions.5 The first concerns how opioid abuse affects a person’s decision to make loan payments, with the drug usage warping their financial decision-making. The second concerns how a bank, not knowing if any one borrower is abusing opioids, uses public, county-level data on opioid-related overdoses to estimate the risk of lending to a borrower in a specific county.

Also, the authors examined state-level antiopioid legislation to measure their effectiveness in mitigating the opioid crisis in general and their impact on credit supply specifically. They studied six different types of opioid-related laws — three types of laws designed to reduce the supply of opioids (opioid prescription limiting laws, mandatory prescription drug monitoring program laws, and triplicate prescription laws) and three types of demand-side laws designed to discourage opioid usage (naloxone laws, Good Samaritan laws, and medical marijuana permitting laws).

They find that bank lenders, to account for higher credit risk, significantly reduced credit supply (by 10 percent) to opioid-impacted populations. (Indeed, the authors observed that individuals in counties with high opioid usage had more days past due on their credit cards.) When banks did solicit consumers with credit card offers in counties with heavy opioid usage, they charged a higher interest rate on the credit card debt (1–2 percentage points higher), lowered the credit limit (a 12–21 percent decrease), and offered fewer rewards (a 4 percent decrease).

Within high-opioid-usage counties, credit terms were worse for certain groups, including people with incomes under $30,000 per year, those under the age of 25, and minorities (particularly Black Americans). The consequences of incomplete information available to lenders, the authors find, also spilled over to residents in high-opioid-usage counties who do not use opioids. Notably, these residents faced higher interest rates on credit offers from lenders.6

State-level legislation targeting the supply of opioids, the authors find, helped reduce opioid prescriptions and opioid prescription death rates but had only a limited effect on deaths from illicit opioids. These supply-oriented laws also helped alleviate liquidity shortfalls to the affected communities. However, legislation targeting demand was less beneficial and sometimes even aggravated the opioid crisis. In addition, demand-side legislation was less effective in improving the credit supply to high-risk communities and, in some cases, had a detrimental effect on the credit supply.

“The cautious behavior of banks appears to be partially justified by the relatively high credit risk in the highly affected areas,” the authors conclude, but at the same time, the reduced credit has significant negative repercussions, on net, for these communities. A lower credit supply led to a significant decline in consumption spending as measured by credit card purchases. This result has important macroeconomic implications, the authors argue, given consumption’s large contribution to gross domestic product and economic growth. 

The findings by Agarwal, Li, Roman, and Sorokina provide key information on the economic and financial impacts of the opioid epidemic and the effectiveness of existing legislation to address the crisis. Policymakers can use their findings to further support recovery and promote economic growth in opioid-affected counties.

  1. The views expressed here are solely those of the author and do not necessarily reflect the views of the Federal Reserve Bank of Philadelphia or the Federal Reserve System.
  2. Centers for Disease Control and Prevention, “Drug Overdose Deaths,” available at
  3. Dionissi Aliprantis, Kyle D. Fee, and Mark E. Schweitzer, “Opioids and the Labor Market,” Federal Reserve Bank of Cleveland Working Paper 18-07R3 (2022).
  4. For additional data, see Consumer Financial Protection Bureau, “The Consumer Credit Card Market,” October 25, 2023, available at
  5. This data set is from the anonymized Mintel/TransUnion Match File, which is compiled from surveys of a sample of about 4,000 consumers each month. The authors selected the period 2010 to 2019 to avoid the dates of the Great Recession; the COVID-19 pandemic; and the 2009 implementation of the Credit Card Accountability Responsibility and Disclosure Act, which prevents banks from raising interest rates on newly opened credit card accounts for one year.
  6. To allow for a direct comparison of the credit supply in U.S. counties, the authors used various econometric techniques, including instrumental variables, propensity score matching, and contiguous counties techniques. They also controlled for differences in local economic conditions and demographics.
  7. Similarly, the authors note, employers in opioid-crisis communities face an information problem that can lead to lower wages paid to those who do not use opioids.