Risk in the Background: How Men and Women Respond

While generally women are more risk averse than men, when women have even small amount of income, they are more willing to take future risks. Instead, men increase their risk-taking after winning, even if the odds do not favor them subsequently.


Risk influences our decisions and how we act. Deciding between two options, one more risky than the other, can greatly affect our outcomes in jobs and at home. Risk aversion studies typically use isolated situations to explore the effect of gender on risk. However, outside of the lab, decisions to take risks rarely occur in isolated situations. Rather, there is both background risk (impending risk) and realized risk (past risks that may have paid off, such as winning a lottery) to consider. Research tells us that winning produces several outcomes: (1) emotion, (2) subject expectations (i.e. belief an individual will keep winning once they already have, or "gambler's fallacy"), and (3) an elevated level of hormones, namely testosterone. This study explores whether there are gender differences in how exposure to past risk (realized and unrealized) affects subsequent risk-taking.. Understanding how this context affects the relationship between gender and risk-taking may help close micro-gender gaps, which could then decrease larger inequities. By using a randomized control trial that involved a dice game, the experiment accounts for different underlying risk environments and highlights how men and women responded differently to them.

Each participant was placed in one of four treatments: a control treatment (receiving no money), a background risk treatment (a lottery scenario where participants could win either $2 or $30 based on how a dice rolled), a low-fixed sum treatment (everyone received $2), and a high-fixed sum treatment (everyone received $30). Afterwards, all subjects participated in two tasks to elicit their risk attitude. The first task took place before they received the payout from the first part of the experiment, and the second task afterwards. This helped gauge the effects of their past realized and unrealized risk experiences on their future risk-taking.


Women who were initially exposed to the lottery (and didn’t know the outcome) or who were assigned to receive a fixed-sum of money, were less risk-averse than women in the control group. Men did not experience this income effect. On the other hand, men who won the lottery increased their risk-taking, while women did not.

  • In the control group, women were 17% more likely to prefer a safe choice relative to men. This is women’s baseline risk aversion, as it reveals the female participants’ preference for lower risk in a neutral environment.
  • Women who were assigned to the (not yet realized) background risk group, were 13% less likely to prefer safe choices relative to the females in the control group, while men didn’t respond to the unrealized background risk nor to receiving a fixed amount of money.
  • In both fix-sum groups, women took more risks relative to the female controls – an increase of 9%, with no significant difference between the risk-taking behavior of women in the low- or high-fixed sum groups. This effect was fleeting, however, ending once the experimenter moved onto the second risk elicitation task where payouts from the first lottery and fixed sums had already been implemented.
  • Men who experienced a lottery win increased their subsequent risk-taking behavior relative to women. In the second risk elicitation task, men took 33% more risk than women who won the lottery; in the control group this difference was 15%. Given that receiving a fixed amount of money did not have an effect for men, this result suggest that it was the experience of winning itself that altered future behavior.

Women’s reduction in risk aversion could be explained by their response to an income effect, as in, the potential to earn money without risking losing money, since they have a fixed sum to fall back on if they lost. This could potentially mean that women modify their risk-taking behavior based on whether they have financial security. Security does not seem to affect men, however winning does increase their subsequent risk-taking choices.


The study was conducted as a randomized control trial at Harvard’s Decision Science Laboratory and included 160 students split evenly by gender. The 75 subjects in the background risk treatment were provided with a dice and were informed that an experimenter would come to roll this dice later in the experiment and that they would earn $30 when the dice turned up 1, 2, or 3, and $2 otherwise. Subjects in the high fixed sum treatment were notified that they received $30, and subjects in the low fixed sum treatment were notified that they received $2. Subjects in the control treatment received nothing. After reminding all subjects of their assigned group, two risk elicitation tasks that followed allowed subjects. This gauged the effects that their assigned risk environment group had on risky behavior.

Related GAP Studies