Debt Literacy: What the Research Says About Interest, Fees and Costly Borrowing

Debt literacy, as Lusardi and Tufano define it, is the ability to make simple decisions about borrowing, starting with how interest compounds. In the 2009 working-paper version of their phone survey of 1,000 Americans (published in 2015), only about one-third grasped compounding or how credit cards work, and people with lower debt literacy tended to use costlier borrowing and more often judged their debt excessive or were unsure about it.

Most of that evidence is survey-based, so it describes association. Studies of state requirements and one Brazilian randomized trial found better credit records or less high-cost borrowing among students exposed to financial education. A Peruvian trial found no effect on whether students had credit or repaid, and a Russian household panel found no link between financial literacy and over-indebtedness.

Teaching interest from the borrower's side

The Council for Economic Education framework lists managing credit as its own pillar, separate from saving, and our 48 modules use project-based learning throughout. The questions that anchor debt literacy are about borrowing: how fast an unpaid card balance doubles, what minimum payments cost, how one payment plan compares with another.

Interest looks different from the borrower's side, and those survey questions test that side. Nothing in the studies below evaluates our program or any other classroom curriculum for its effect on how students borrow as adults.

What did Lusardi and Tufano find?

Start with Lusardi and Tufano, "Debt literacy, financial experiences, and overindebtedness," Journal of Pension Economics and Finance, 2015. The figures below come from the 2009 working-paper version, which reports more detail than the published abstract.

The authors surveyed 1,000 US residents by phone in November 2007. One question asked how long a $1,000 credit card balance at 20% interest, compounded annually and left unpaid, would take to double. Only 35.9% chose the correct answer, less than five years. Ignoring interest on interest gives five; compounding makes it shorter. Another 18.3% said they did not know.

Their main findings, after controlling for demographics, income and wealth:

• Only about one-third of respondents could apply compounding to everyday situations or understood how credit cards work.
• People with lower debt literacy were more likely to fall into groups defined by costly borrowing, such as paying fees or using alternative financial services (payday and similar loans).
• Using industry cost data and an assumption of one event a year, the authors modelled the late, over-limit, cash-advance and minimum-payment charges of cardholders who rated their own financial knowledge 4 or lower on a seven-point scale (28.7% of cardholders). The model puts that group's charges about 50% above the average cardholder's. If the average cardholder incurred $100 of these charges, the group would incur about $150. This is a modelled cost, not observed fees.
• In the authors' model, as much as one-third of those charges are attributable to low knowledge rather than to other observable traits.
• Lower debt literacy went with judging one's debt excessive or being unable to judge it. In the survey, 26.4% said they had or might have difficulty paying off debt, and another 11.1% did not know where they stood.

Overindebtedness was self-reported, and the survey cannot separate pure financial knowledge from numeracy and cognitive ability, which the authors say themselves.

Do other studies find the same pattern?

The pattern shows up in different data and for different kinds of borrowing.

In the last row, among the 10,560 workers, people who answered all three standard literacy questions correctly had odds of using payday, title or pawn loans that were 58% lower (an odds ratio of 0.42) than people who did not, after controls for income, age, risk tolerance, emergency savings and other factors. Odds are not probabilities, so that does not mean a 58% drop in the share of borrowers. The paper's authors describe it as a reduced likelihood, and in the sample about 24% had used one of these loans.

The mortgage rows disagree on the mechanism. The homeowner survey points to risky loan choices. The study that matched survey answers to loan records found that people with weaker numerical skills defaulted more, but not because they picked different contracts.

Does teaching it change borrowing?

Association does not tell a teacher whether instruction helps. Three studies can say more, and all three found better credit records or less high-cost borrowing. Two other school programs, covered in the limits below, did not match them.

• State requirements: a Federal Reserve working paper by Brown, Collins, Schmeiser and Urban compared young adults in Georgia, Idaho and Texas before and after personal finance mandates took effect in 2007 with comparison states, using credit bureau records. For students three years in, credit scores through age 22 were higher by 28.7 points in Georgia, 6.5 in Idaho and 13.0 in Texas, and the share ever 90 or more days delinquent was lower by 3.6, 0.7 and 3.3 percentage points. All of these differences are statistically significant at the 5% level or better. Georgia's gain equals about a third of a standard deviation in credit scores. Idaho's first graduating class scored 7.3 points lower before gains showed up.
• Payday borrowing: a brief by Melody Harvey for the FINRA Foundation, summarizing her paper, estimated that young adults in states whose graduating classes were required to take personal finance (exposure inferred from age and current state) were four percentage points less likely to have borrowed payday loans. Required courses did not necessarily reduce rent-to-own use or use of alternative financial services overall.
• A randomized trial: in Brazil, 892 high schools were randomly assigned a financial education program. Using administrative data on about 16,000 students over the following nine years, treated students were less likely to borrow from expensive sources or repay late.

Where the debt-literacy evidence is thin

• Association is not cause: the survey studies cannot rule out that something else, such as income security or a stable job, drives both knowledge and borrowing.
• One panel found nothing. A Russian study used four survey waves from 2018 to 2024 (about 6,000 households per wave in the survey). After restricting to indebted households above the poverty line, its fixed-effects estimate rests on 728 observations, and pooled and two-stage estimates with larger samples also found no link between financial literacy and over-indebtedness, defined as income left after loan repayments falling below the poverty line.
• The mandate results are sensitive. In the Federal Reserve study, moving the clock to the year a law passed instead of the year it reached classrooms pushed the effects toward zero.
• Not every school program moves credit outcomes. A Peruvian trial (Frisancho, 2023) found literacy gains but no effect on whether students had credit or repaid, with fewer arrears (down 20%) only among the few who had loans. The early Brazilian results (Bruhn and colleagues, 2016) showed better saving and budgeting but also more use of expensive credit in the short run.
• Randomized evidence is thin in the US. The US teaching evidence here comes from mandates, which are policy changes rather than random assignment. The randomized trials cited are from Brazil and Peru.

Classroom moves from the debt literacy research

The research points to compounding applied to debt, fees and the price of different kinds of credit.

1. Start with the doubling question. Ask students how long a $1,000 balance at 20% takes to double before anyone calculates. In the 2007 survey, 9.6% said two years, 18.8% said five to ten, 13.1% said more than ten and 18.3% did not know, so expect a mix and use the wrong answers as your starting point. The survey also asked how long a $3,000 balance would take to clear on $30 monthly minimum payments at 12% interest, and which of two payment plans for an appliance cost less.
2. Put a cost on fees. Have students total what a late fee, an over-limit fee and a cash advance would add to a year of card use. The Lusardi and Tufano result is about fees as much as interest.
3. Name the high-cost lenders. Payday, auto title and pawn loans are the categories in the National Financial Capability Study analysis above, so students should recognize them and be able to compare their cost with a card or a credit union loan. Teaching teens about credit is a good place to start.
4. Separate costly debt from useful debt. What is good debt and student loan interest rates give you cases where borrowing can be reasonable, so the lesson covers price and terms.

Teachers planning a credit unit can request a free Rapunzl demo and ask our team about the curriculum and simulator.

By Nate Thomas, School Partnerships Lead at Rapunzl.