The study shows that when people have different levels of risk in their job incomes, using a model with one average person can lead to wrong predictions about how much money you can make without risk and how much extra return you get from investing in stocks. This happens especially if people are more careful when they have more money at stake. This suggests that risks that can't be traded, like job security, can help explain why these predictions are often off.