Picture yourself in a store, deciding between a nicer pen and a cheaper one. Now imagine that once you choose, you have to stand up in front of a room of strangers and announce which pen you picked. Would that change your decision? And if you had the option to borrow money to buy the fancier one, would the audience push you toward debt?
These are the kinds of questions that motivated a new study in the Journal of Behavioral and Experimental Finance, which set out to untangle why social pressure seems to drive people into household debt. Previous research has linked keeping up with neighbors to higher borrowing and even bankruptcy, but the exact reasons have remained murky. The authors ran a controlled experiment to see whether people borrow more to impress others, to mimic what others are doing, or for some other reason entirely.
Two theories about peer pressure
Antonia Grohmann of Aarhus University and Melanie Koch of the Oesterreichische Nationalbank wanted to separate two very different explanations for peer effects on spending. The first is what economists call social image: the idea that people consume visibly to shape how others see them. The second is peer information: the idea that people copy others because they assume others know something they don’t, or simply want to fit in.
In the real world, these two channels are tangled together. If your neighbor buys a new car and you buy one soon after, is it because you want to look successful, or because seeing the purchase told you something about what’s normal or desirable? Observational data can’t easily separate these motives, so the researchers built a lab experiment where they could switch each channel on and off.
A pen, an IQ quiz, and a loan
The researchers recruited 305 students at Technische Universität Berlin. Participants first took a seven-minute IQ-style test with twelve questions. Their earnings depended on how they ranked against others in the session: the top quarter earned 3 euros, the next quarter 2 euros, the third quarter 1 euro, and the bottom quarter just 50 cents.
Next came a shopping round. Participants could use their quiz earnings to buy one of five pens, priced from 50 cents to 4 euros, with star ratings signaling quality. Any quiz money not spent on a pen was forfeited. Participants could also borrow up to 3.50 euros to afford a nicer pen than their quiz earnings allowed. The idea was that buying an expensive pen would hint at being a top performer, while buying a cheap one would suggest the opposite.
After shopping, everyone did a tedious computer task called the slider task, moving on-screen sliders to specific positions for four minutes. Earnings from this task could be used to pay off any loan taken earlier. If someone didn’t earn enough to cover their debt, the difference came out of their participation fee.
The twist was in how the pen choice was made and shared. In the control group, each person’s choice stayed private. In the public group, participants had to stand up and announce their pen choice aloud after everyone had decided. In the information group, participants made choices one at a time, and each person could see on their screen how many of each pen had been bought so far, though not who bought what.
The public condition isolated social image: others learn about you, but you learn nothing new about them. The information condition did the opposite, providing data about others without exposing your own decision.
What the researchers expected, and what happened
The authors predicted that in the public condition, lower-scoring participants would take out loans to buy nicer pens, hiding their low quiz performance behind a pricier purchase. They also expected participants in the information condition to drift toward whatever choices they saw others making.
Neither prediction held up. Overall, about 20% of participants took a loan, and the amount borrowed didn’t differ meaningfully across the three conditions. In fact, when controlling for small differences between groups, the effect of the public condition on borrowing was essentially zero.
An exploratory analysis, which the authors note was not pre-registered, turned up something different. Participants in the public condition were significantly more likely to buy a pen of lower quality than their earnings could afford, leaving money on the table. This pattern was driven largely by the highest-performing participants, who chose cheaper pens than their earnings warranted.
Under the information condition, the researchers found no statistically significant effects in either direction. Descriptive patterns hinted at some conformity, but the evidence didn’t hold up in formal tests.
Why would high scorers downgrade?
The finding that top performers chose cheaper pens when everyone was watching is the opposite of the classic “conspicuous consumption” story. The authors offer several possible explanations, all of which still involve social image concerns, just not the kind they originally anticipated.
One possibility is debt stigma. If being seen as a borrower carries social cost, high performers might pick a cheaper pen to make clear they didn’t need to borrow. Another is what the authors call a “smarty-pants effect,” where people avoid appearing smarter than peers to dodge seeming arrogant. A third, which the authors find most plausible, is blame aversion: because the quiz was graded on a curve, one person’s high score meant someone else’s low score. Top performers may have wanted to avoid publicly flaunting a success that came at others’ expense.
The researchers also looked at whether personality traits predicted who responded to the treatments. They measured cognitive reflection, locus of control, self-esteem, self-monitoring, and the Big Five traits. They found few meaningful interactions, with one exception: in the information condition, more open participants were more likely to deviate from their pre-experiment pen preference.
A puzzle in the effort task
The researchers expected that participants who took loans would work harder at the slider task to earn back what they owed. Instead, they found the opposite. Loan amount was significantly negatively related to effort in the slider task, even after controlling for general ability. This effect was strongest in the public condition. The authors suggest that taking a loan and publicly announcing one’s pen choice may have had a demotivating effect on subsequent effort, though they don’t have a clean explanation for why.
Caveats and open questions
The authors are candid about the limits of what they can conclude. Lab borrowing is artificial: participants couldn’t leave indebted, and any “loan” was repaid within the same session rather than carried for weeks or months. The final sample was about 17% smaller than planned because pilot sessions were excluded, which reduced the study’s ability to detect small effects. The effect sizes for loan take-up in both treatment conditions were too small to be economically meaningful, but the confidence intervals leave some uncertainty.
The pen itself may matter as well. Pens are cheap, familiar, and low-stakes. Products where quality differences are more visible and consequential, like cars or clothing, could produce different patterns. The authors also note that their setting involved an unusual feature: one person’s success on the IQ quiz mechanically meant another person’s failure, which may have amplified blame aversion in ways that wouldn’t apply to ordinary consumer choices.
The researchers interpret their results as evidence that social image concerns do shape borrowing decisions, just not always in the direction the existing literature suggests. Rather than spending more to signal status, their participants sometimes spent less, seemingly to manage how they appeared to peers in a specific setting where achievement was zero-sum and publicly visible.
