People prioritize expected growth over expected value

People prioritize expected growth over expected value. Adam Bear, Dorsa Amir, Matthew R. Jordan, Fiery Cushman. Human Behavior and Evolution Society 31st annual meeting. Boston 2019. http://tiny.cc/aa1w6y

Abstract: How should people make decisions? A voluminous literature dating back to Bernoulli suggests that people should maximize expected utility. According to this theory, we should prefer an investment that, at every time step, either grows by 40% or shrinks by 30% to an investment that grows by 10% or shrinks by 5%. But while the former investment offers a higher average return over time, the latter investment is expected to grow over time at a faster rate, as characterized by the geometric mean of its payoffs (Peters & Gell-Mann, 2016). Given that this growth rate will largely determine which people or traits survive over evolutionary time, we hypothesized that people’s investment decisions would prioritize this quantity over the more familiar expected value. In a first experiment, we show that people prefer a ‘safe’ investment to its riskier, but higher expected-value, counterpart when this safe investment has a faster growth rate. However, when the riskier investment has a faster growth rate than the safer investment, this pattern reverses: people are more likely to take the risk. These findings provide initial evidence, consistent with evolution, that people may rationally prioritize the long-run growth of a process over simple expected value.