Gender, Risk, and Retirement
Bernasek, Alexandra, Shwiff, Stephanie, Journal of Economic Issues
Women are more vulnerable than men to poverty in their older age. They earn less than men over their working lives and accumulate less savings for retirement. These savings typically have to be spread over a longer period of time since women live longer on average than men, and they must meet higher expenditures since women also experience more chronic health problems as they age. Saving for retirement is increasingly taking the form of investment in financial assets. The proportion of private pensions that are defined contribution plans is increasing, and policy makers are considering the privatization of Social Security. In this environment women's well-being in retirement will depend more heavily on their attitudes toward risk and the impact of those attitudes on investment decisions.
There is evidence that women are more risk averse than men when their entire portfolio of assets is considered (Jianakoplos and Bernasek 1998, Palsson 1996). If women are on average less willing to take risks than men, they are also expected to accumulate less wealth on average, since lower risk is associated with lower returns on investment. Some researchers have looked specifically at pension assets and individuals' willingness to take risks as evidenced by the allocation of their pensions to stocks versus bonds. Several studies have found that women invest their pension assets more conservatively than men--allocating a smaller percentage to stocks than bonds (Bajtelsmit and Vanderhei 1996; Hinz, McCarthy, and Turner 1996; Bajtelsmit, Bernasek, and Jianakoplos 1999).
One of the major limitations of previous studies of gender and risk aversion is their inability to deal with the problem of who makes investment decisions in married and cohabitating couple households. If a female respondent who is married holds a risky portfolio of wealth, we do not know if this reflects her attitudes toward risk or her partner's, or some combination. Nancy A. Jianakoplos and Alexandra Bernasek (1998) tried to overcome this problem by comparing single women, single men, and married couples in terms of their risk aversion. They found that single women are more risk averse than single men and married couples. One of the problems with this is that single women and men are not necessarily representative of women and men in general. When looking at private pension allocation decisions, it seems more likely that the individual who has the pension makes the allocation decision with regard to the pension, but still it seems reasonable that a partner's attitudes toward risk (especially as it affects how the household portfolio of assets is allocated) will have some effect on that decision. The purpose of this paper is to examine the determinants of an individual's allocation of defined contribution pension assets to stocks when we have detailed information about gender and the household financial decision-making process.
Data and Descriptive Statistics
The data used in this paper came from a survey of faculty employed at five universities in Colorado.' The data were collected in the form of a mail survey conducted in spring 2000. The purpose of conducting such a selective survey was to produce a relatively small data set with very detailed information about household financial decision making which could be used as a pilot study to see if the collection of such detailed information on a larger scale would be warranted. Also, by focusing on academics only, the issue of level of education was controlled for and a potentially more significant factor--the type of education or discipline--could be examined.
Given the specific nature of the questions asked in the survey, there were only 270 observations in our sample--those who had complete information on the percentage of their defined contribution pension invested in stocks, as well as on the other explanatory variables. Means of the variables are presented in table 1. The means of several of the variables differed by gender at the 1 percent level of significance. …