How your personality “acts as an invisible architect for your career,” according to new research

Young adults who are willing to wait for larger financial rewards tend to apply to college programs associated with longer study durations and steeper earnings growth, while those who avoid risk lean toward fields where graduates experience more predictable salaries, according to a study published in PNAS. The findings suggest that individual economic preferences predict postsecondary educational choices. This pattern indicates that educational sorting reflects how young people evaluate time and uncertainty, operating alongside academic performance and family background.

The research was conducted by an international team of six economists who contributed equally to the work. Their investigation examines how fundamental economic preferences predict real-world educational investments. In economics, two behavioral traits are central to these decisions: patience and attitudes toward risk. Patience refers to a person’s willingness to wait for larger rewards down the road, while risk aversion describes a preference for stable outcomes over uncertain prospects that carry financial risks.

Economists typically view higher education through the framework of human capital theory, which treats schooling as an investment. Under this model, students spend years out of the workforce gaining skills with the expectation of earning higher wages later in life. Individuals who place greater value on future rewards are theoretically more willing to postpone earnings to complete longer degrees. At the same time, because career paths offer widely varying degrees of earnings stability, people who dislike uncertainty are expected to shy away from fields with unpredictable financial prospects.

Empirical evidence linking these economic traits to higher-education decisions has remained scarce. Testing these theories requires measuring how people handle risk and delayed rewards before they make their educational investments. It also requires gathering those measurements across large groups of students and observing their actual choices. Because laboratory-style experiments with real money are difficult to combine with official educational records, prior research has often relied on smaller samples or brief survey questions rather than incentivized choices.

“While psychologists have long known that personality influences our choices, proving that specific cognitive traits dictate major life paths has remained an empirical challenge,” study co-author Ernst Fehr, a professor of economics and director of the UBS Center for Economics in Society at the University of Zurich, told PsyPost. “Historically, it has been nearly impossible to capture a person’s baseline traits before they make massive life decisions and then objectively track their real-world trajectories over decades.”

Free daily newsletter

Rather than following participants for decades, the team measured their preferences just before they applied to higher education and drew long-term earnings data from an older generation of graduates. They conducted the study in Denmark, which Fehr described as a “pristine real-world laboratory.” Higher education in Denmark is tuition-free, and students receive monthly government stipends to support living costs regardless of family wealth. This arrangement allowed the scientists to examine how personal preferences predict educational investments without major distortions from borrowing limits.

The authors invited a random sample of 18- and 19-year-olds from Danish population registries in 2018, catching them just before they submitted higher-education applications. The final analysis focused on 3,132 young adults who completed upper secondary school, had official ninth-grade exam scores, and applied to programs receiving at least ten applicants from the study. In all, the applicants sorted across 76 educational options, consisting of 72 postsecondary degree programs and four upper secondary school tracks.

The researchers measured patience through an online experimental task involving 16 financial choices. In each decision, participants could take a sum of money earlier or save some of it to receive more later, such as eight weeks versus 16 weeks, with the delays and interest rates varying across decisions. One decision was randomly chosen to be paid out for real, with an average payment of about 250 Danish kroner (roughly 40 U.S. dollars) sent to participants’ phones. The researchers ranked participants on a percentile scale from 1 to 100 based on their average savings rate.

To measure risk aversion, participants completed an investment exercise featuring 15 financial scenarios. In each situation, they could keep an amount of money with certainty or invest part of it into a lottery offering a chance at a larger return alongside the risk of a loss. By observing the share of money each person kept safe rather than risking it, the researchers ranked each participant’s risk aversion on a percentile scale from 1 to 100.

The authors then linked these experimental rankings to official administrative registries, identifying each student’s first-priority application in Denmark’s centralized admission system. The team characterized each program by its official length and, using tax records from roughly 330,000 Danes born between 1975 and 1980, by its graduates’ typical annual earnings growth, how widely their salaries varied, and the share who ended up in the bottom fifth of earners.

The researchers compared the 76 programs based on the average patience and risk-aversion ranks of the students who applied to them. On its own, a 10-percentile difference in applicants’ average patience corresponded to programs that lasted nearly one and a half years longer and offered 2.4 percentage points higher annual earnings growth.

Programs that drew more risk-averse applicants tended to have fewer graduates with very low earnings and less variation in pay. Before any adjustments, a 10-percentile difference in applicants’ average risk aversion corresponded to a 6.2 percentage point lower share of graduates in the bottom fifth of earners, compared with an average share of 9.9 percent across the programs students applied to.

These relationships weakened but did not disappear once the authors accounted for school grades, parental education and income, gender, and immigrant background. With those factors included, a 10-percentile difference in patience corresponded to programs a little over two months longer, and a 10-percentile difference in risk aversion corresponded to a share of low earners about 3 percentage points lower.

Risk aversion also exhibited an unexpected link with degree duration and earnings trajectories. In both simple comparisons and models with controls, programs drawing more risk-averse applicants had longer study durations and higher earnings growth.

“Intuitively, one might assume that risk-averse individuals would avoid long, demanding university programs out of fear of failure,” Fehr pointed out. “Instead, we observed that risk-averse individuals sort disproportionately into some longer, and high-earnings-growth, programs. The reason for this is that they tend to select programs with lower downside earnings risk and lower earnings dispersion; because longer and high-growth programs also tend to have safer labor-market outcomes, risk-averse individuals consequently sort disproportionately into some longer, high-growth programs. Economics illustrates this pattern: it combines a multi-year program and high earnings growth with a relatively low probability of very low earnings.”

Comparing the predictive power of preferences to school grades highlighted their relative importance. Patience was about one-fourth as predictive of program length as school grades were, and its link to earnings growth was about one-sixth that of grades. For avoiding programs with many low earners, risk aversion was about as predictive as grades, and it was roughly twice as predictive as grades when it came to avoiding programs with widely spread earnings. These patterns suggest that sorting across degree tracks reflects personal preferences regarding time and risk, not just past academic achievement.

“The psychological footprint on career selection is remarkably powerful, rivaling traditional cognitive metrics like intelligence,” Fehr told PsyPost. He noted that in their data, which used school grades as a proxy for cognitive ability, an individual’s aversion to risk was “just as powerful a predictor – and sometimes stronger – than their academic performance” when predicting which applicants avoided fields with unstable earnings.

The findings are in line with research covered by PsyPost in 2026, which indicated that a greater orientation toward future rewards was associated with completing higher levels of education. That earlier study evaluated genetic predispositions for delay discounting rather than directly elicited behavioral preferences with cash stakes, but both studies point to a link between patience and educational investments.

“The core takeaway is that your deep-seated psychological profile – specifically your capacity for delayed gratification and your tolerance for risk – acts as an invisible architect for your career,” Fehr observed. “Highly patient individuals naturally gravitate toward marathon educational paths with steep, back-loaded rewards, while risk-averse people choose fields that offer psychological safety through predictable, stable income streams. This reveals that structural societal inequalities aren’t just born from academic talent or family wealth, but from how our minds evaluate risk and time.”

There are a few things to keep in mind when interpreting these findings. The study examined students’ first-choice applications rather than the programs they actually enrolled in or completed, and about 74 percent of applicants were admitted to their top priority. Additionally, because the main analysis evaluated programs by the average traits of their applicants, it captures sorting across fields rather than predicting each individual student’s choice. An alternative model evaluating individual-level choices yielded similar directional patterns, though the estimated effects were smaller.

Fehr emphasized that these patterns should not be viewed as direct proof of cause and effect. “A critical nuance is that our study uncovers predictive patterns rather than a strictly causal mechanism, as it is impossible to ethically manipulate a person’s core personality traits in a lab,” he explained. “Furthermore, we want to clarify that our findings do not hinge on the exact monetary amounts used in our behavioral tasks. Rather, the true value lies in the experiment’s robust psychological validity – its capability to accurately rank where individuals sit on the spectrum of risk and patience relative to their peers.”

Participation patterns and historical data also present caveats. Only 39 percent of the invited young adults logged on to take part, and participants had somewhat higher grade point averages than the general population. In addition, the earnings benchmarks were derived from graduates born between 1975 and 1980, meaning current labor market conditions could differ. The Danish setting is also distinct, as universal living stipends and free tuition reduce financial barriers that might interact differently with risk aversion and patience in systems where students face tuition fees.

Looking ahead, the researchers hope these insights help reshape economic theories and policy discussions regarding educational inequality. “Our long-term objective is to dismantle the outdated assumption that all humans evaluate risk and time identically,” Fehr concluded. “By taking into account the vast diversity in human preferences, we hope to influence how both scientists and policymakers approach systemic inequality. Ultimately, interventions designed to help people climb the socioeconomic ladder must account for individual psychological diversity rather than relying on one-size-fits-all financial incentives.”

The study, “Economic preferences predict higher-education choices,” was authored by Thomas F. Epper, Ernst Fehr, Kristoffer B. Hvidberg, Claus T. Kreiner, Søren Leth-Petersen, and Basit Zafar.