Risk tolerance of men and women

Francesco D'Acunto

20 September 2015

a

A

Heterogeneous preferences and beliefs have major implications for the dynamics of labour markets (Killingsworth 1987), financial markets (Garleanu and Panageas, forthcoming), and the macroeconomy (Chien et al. 2014). Heterogeneity across genders is especially relevant – each group includes half of the population, and there are systematic differences in the preferences of men and women, which in turn have large effects on economic outcomes (Croson and Gneezy 2009, DellaVigna et al. 2013, Bertrand et al. 2015).

Research so far has studied the effects of biology and genetics on risk attitudes across genders (Sapienza et al. 2009), Barnea et al. 2010). In a recent paper (D’Acunto 2015), I test whether social identity may help explain these differences. Social identity is the set of stereotypical behaviours society attributes to men and women (Akerlof and Kranton 2000).

Table 1. Identity stereotypes attributed to men and women by 200 survey responders                        

Male identity increases risk tolerance

How to test for a causal effect of social identity on preferences and investments? I randomly manipulate identity stereotypes at the individual level in a controlled environment, which makes it clear to subjects the types associated with their gender group (Benjamin et al. 2010). Then I ask subjects to perform risk tolerance elicitation tasks and simple investment tasks. Manipulations make identity stereotypes salient or threaten identities (Willer et al. 2013). Both interventions increase the willingness to engage in risky behaviours by men but not women in the psychology literature.

In a first experiment, I elicit subjects’ risk tolerance through lottery choice tasks before and after randomly priming or threatening their identity. This procedure allows measuring the change in risk tolerance due to the manipulations at the individual level. Figure 1(a) shows that the distribution of the risk tolerance of men before the manipulations is similar across treatment groups: all are risk averse on average. Figure 1(b) plots the distribution of the in risk tolerance across treatment conditions. Men in the control group have the same risk tolerance before and after the manipulations. But men in the salience and those in the threatening conditions become more risk tolerant on average after each manipulation.

Figure 1a. Risk Tolerance of Men before Manipulations

Figure 1b. Change in Risk Tolerance of Men due to Manipulations

Identity makes men overconfident

But which mechanisms transmit the effect of identity on men’s risk attitudes? In a second experiment, I find identity makes men overconfident. Primed men believe they will succeed more than what the objective probability of success suggests. To create a subject-level measure of better-than-average beliefs, I ask subjects how often they think themselves and their neighbours would succeed if investing ten times in an opportunity that on average succeeds five out of ten times. Figure 2(a) shows men are overconfident on average while women are not, consistent with the conjecture of Barber and Odean (2001). In Figure 2(b), I plot the measure of better-than-average beliefs across treatments for men only. Primed men (right) believe they are more likely than their peers to succeed in a pure game of chance, that is, the primes make men more overconfident.

In a third experiment, I show this effect on overconfidence acts through the illusion of controlling random processes. I replicate the results inducing overconfidence in the form of the illusion of control. I induce overconfidence by priming a sense of power over others (Fast et al., 2009). Similar to the identity manipulations, men induced with overconfidence become more risk tolerant.

Figure 2a. Overconfidence across genders

Figure 2b. Men’s better-than-average beliefs across treatment conditions

Identity affects men’s investment decisions

I then analyse investment behaviour. In a fourth experiment, I manipulate identity and face subjects with a risky opportunity whose expected value is positive. Subjects receive a virtual endowment of $100, and decide if and how much to invest. Figure 2 shows that men in the identity and overconfidence prime conditions want to invest significantly more often than controls. In an ordinary least squares analysis, I find that men in both conditions invest on average $20 more than controls after accounting for age and education heterogeneity. We can interpret these results as a causal test for the effect of overconfidence on trading behaviour (Barber and Odean 2001). The same results obtain in a fifth experiment, in which I frame the investment decisions as delegated investments on behalf of a principal. We can interpret these results as a causal test of the investment-cash flow sensitivity of overconfident managers (Malmendier and Tate 2005).

Figure 3. Frequency of positive investments by men across treatment conditions

Identity affects older men more than younger men

Gender stereotypes evolve over time and across cultures – their effects on behaviour should fade as long as the difference in perceived male and female stereotypes decreases. I look at the heterogeneity of the effects across cohorts; older subjects should have a starker view of gender identity stereotypes than younger cohorts, and especially those born after 1990. Hence older subjects should react more to the identity manipulations than younger subjects. This is what I find in Figure 4(a) for the change in subjects’ risk tolerance, and in Figure 4(b) for the change in subjects’ willingness to invest.

Figure 4a. Effect of identity on men’s risk tolerance by cohorts  

Figure 4b. Effect of Identity on men’s investments by cohorts

Conclusions

Gender identity, next to biology, helps explain the differences in risk tolerance, confidence, and investment behaviour across genders. Contrary to biology, social identity is a cultural product, and can be manipulated using gender-identity cues. The tests in this column provide a causal validation for the literature relating overconfidence to investment, and they propose identity as a source of overconfidence in men. The results also provide a rationale for marketing campaigns that exploit male and overconfidence cues to increase the take up rates of risky products by consumers. The effects of identity on behaviour fade as gender stereotypes become less stark. Hence, the pervasive differences in behaviour across genders may persist for less than researchers have thought so far.

References

Akerlof, G and R Kranton (2000), “Economics and Identity”, Quarterly Journal of Economics 115(3): 715-753.

Barber, B and T Odean (2001), “Boys Will be Boys: Gender, Overconfidence, and Common Stock Investment”, Quarterly Journal of Economics 116(1): 261-292.

Barnea A, H Cronqvist and S Siegel (2010). “Nature or Nurture: What Determines Investor Behavior?”, Journal of Financial Economics 98: 583-604.

Benjamin D, J Choi and J Strickland (2010), “Social Identity and Preferences”, American Economic Review 100(3): 1913-1928.

Bertrand M, E Kamenica and J Pan (forthcoming), “Gender Identity and Relative Income within Households”, Quarterly Journal of Economics.

Chien Y, H Cole, and H Lustig (2014), “Implications of Heterogeneity in Preferences, Beliefs and Asset Trading Technologies for the Macroeconomy”, St. Louis Fed WP 2014-014A.

Croson, R and U Gneezy (2009), “Gender Differences in Preferences”, Journal of Economic Literature 47(2): 1-17.

D’Acunto, F (2015), “Identity, Overconfidence, and Investment Decisions, Working Paper.

DellaVigna S, J List, U Malmendier and G Rao (2013), “The Importance of Being Marginal: Gender Differences in Generosity”, American Economic Review 103(3): 586-590.

Fast N, N Sivanathan, N Mayer and A Galinsky (2012), “Power and Overconfident Decision Making”, Organizational Behaviour and Human Decision Processes 117(2): 249-260.

Garleanu, N and S Panageas (forthcoming), “Young, Old, Conservative and Bold: The Implications of Heterogeneity and Finite Lives for Asset Pricing”, Journal of Political Economy, forthcoming.

Killingsworth, M (1987), “Heterogeneous Preferences, Compensating Wage Differentials, and Comparable Worth”, Quarterly Journal of Economics 102(4): 727-742.

Malmendier, U and G Tate (2005). “CEO Overconfidence and Corporate Investment”, Journal of Finance 60(6): 2661-2700.

Sapienza P, L Zingales and D Mastripieri (2009), “Gender Differences in Financial Risk Aversion and Career Choices are Affected by Testosterone”, Proceedings of the National Academy of Science.

Willer R, C Rogalin, B Conlon and M Wojnowicz (2013), “Overdoing Gender: A Test of the Masculine Overcompensation Thesis”, American Journal of Sociology 118: 980-1022. 

a

A

Topics:  Labour markets

Tags:  Behavioural economics, beliefs

Assistant Professor of Finance at the R.H.Smith School of Business, University of Maryland

Events