Researchers have long relied on traditional statistics to identify factors like education, occupation, and gender that influence earnings potential. In a pioneering study, experts applied machine learning to pinpoint their relative importance, uncovering that the ability to delay immediate gratification ranks among the top predictors of wealth.
For the first time, machine learning has ranked the key drivers of future earnings. As expected, education and occupation emerged as the strongest predictors. Remarkably, self-control—measured by delay discounting—outperformed factors like age, race, ethnicity, and height in forecasting higher income.
Numerous variables correlate with income, including age, occupation, education, gender, ethnicity, and even height. Behavioral traits matter too, such as those from the iconic "marshmallow test," which assesses delay discounting: the tendency to value immediate rewards over larger future ones. Children showing greater self-control went on to earn more as adults. Yet, traditional analyses struggled to prioritize these factors—until now.
Lead researchers analyzed data from over 2,500 participants, splitting it into training and test sets. Models trained on the former were validated on the latter for robust accuracy.
Occupation and education topped the list, followed by location (by ZIP code) and gender (with men earning more on average). Delay discounting proved more predictive than age, race, ethnicity, or height.
