Future homeowners are doing their homework — literally

An enduring piece of parental advice — do your homework — has been reinforced by new research linking it to long-term financial success.

Researchers revealed a connection between homework completion and future net worth and home ownership. The paper, published June 2026 in the International Review of Financial Consumers, suggests that consistent homework completion is related to valuable traits including perseverance and confidence. With further analysis, the researchers predicted that these benefits will be compromised when students rely on AI.

“It’s important to put this in the context of the emergence of AI and LLM models in the current day,” said Ken Shimizu ’18, M.Eng. ’19. “When students are initially learning concepts, we have to make sure that they are learning on their own first, internalizing the concepts themselves and working through the problems themselves.”

The research team analyzed Cornell’s National Social Survey and the National Longitudinal Survey of Youth 1997, analyzing trends across questions where participants gave their homework completion rate, annual household income and home ownership.

They found that homework completion was positively correlated with future net worth and homeownership. While there was no direct link observed between homework completion and future annual earnings, there was a strong association between homework completion and educational attainment, which in turn contributed to future annual earnings.

“Other studies have shown that homework can improve achievement orientation and self-efficacy,” Shimizu said. “So, through homework, developing certain noncognitive skills related to focus, discipline and decision making can translate into better financial decision making.”

Even after controlling for employment, income, race, marital status, geography and education, those who consistently completed homework as a student were roughly 50% more likely to own a home.

Their findings are especially significant in the age of AI, as some students outsource their homework to large language models (LLMs). AI usage is increasing among people under 30, leaping from 33% to 58% from 2023 to 2025, and LLMs are becoming more sophisticated, making it difficult for educators to tell when a student’s work is AI-generated.

“I tell my students, I want your raw thoughts. I prefer rough language because it’s coming from you,” said Scott Stewart, clinical professor of finance and accounting at the Samuel Curtis Johnson Graduate School of Management in the SC Johnson College. At first, Stewart could recognize common AI tells like the word “moreover” or the phrase “in addition to,” but as these phrasing quirks are smoothed out of models, he can’t confidently tell whether a student’s assignment is AI-generated or not. “Student work is just more professionally written and the LLM text reads more smoothly now,” he said.

When there are no immediate repercussions for AI usage, students may face greater temptation to use it on their homework.

“It’s a shortcut, in a sense,” Stewart said. “It’s like taking some of your friend’s completed homework assignment and submitting it as your own.”

But if students take this option, Stewart thinks they’re cheating themselves of the opportunity to build important skills.

“You need to know what questions to ask and how to interpret the answers, and the only way to do that is to learn by yourself,” he said. “Otherwise, you’re going to rely on other people’s opinions when you really need to rely on your own opinion. You’re going to be forever biased.”

Drawing upon MIT Media Lab research that measured lower rates of learning with AI usage, the researchers predicted that turning in AI-completed homework will lead to a reduction in future income by 2%, net worth by 11%, educational attainment by 0.22 degrees and homeownership rates by 12%. In the future, they are interested in conducting another study with newer datasets controlling for cognitive factors alongside noncognitive factors by incorporating data points such as IQ and SAT scores.

Jay Zagorsky, clinical associate professor at Boston University, co-authored the paper.