Shares have returned far more than safe bonds over the long run, by more than any plausible degree of risk aversion explains.
Mehra and Prescott stated it in 1985: reproducing the historical gap inside a standard model requires a level of risk aversion that would imply implausible behaviour everywhere else. The candidate explanations are rare disaster risk, loss aversion combined with short evaluation horizons, borrowing constraints, and survivorship in the record itself, since the American market that produced the numbers is the one that did not collapse.
In practice
- A pension fund holds bonds because its committee reviews results yearly, and the yearly odds look worse than the thirty-year odds it is actually running.
- A saver who checks a portfolio daily sells during a fall that would have been invisible on an annual statement.
- Standing
- Still open. Each proposed resolution explains part of the gap, and the premium has narrowed in recent decades and in other countries' data. Compare survivorship bias and loss aversion.
- Subject
- Economics & Incentives
- Essay
- Rules for choosing, and the paradoxes that humbled the theory