If smart people are more likely to notice ways to save their lives that cost some money, in statistics this may appear as a negative correlation between smartness and wealth. That's because dead people are typically not included in the data.

As a toy model to illustrate what I mean, imagine a hypothetical population consisting of 100 people; 50 rational and 50 irrational; each starting with $100,000 of personal wealth. Let's suppose that exactly half of each group gets seriously sick. A sick irrational person spends $X on homeopathy and dies. A sick rational person spends $40,000 on surgery and survives. At the end, we have 25 living irrational people, owning $100,000 each, and 50 living rational people, owning $80,000 on average (half of them $100,000, the other half $60,000).