Gender Differences In Financial Risk Taking: The Role Of Financial Literacy And Risk Tolerance
Johannes Gutenberg-University Mainz
Johannes Gutenberg-University Mainz
March 16, 2016
We study financial risk taking via standard and sophisticated financial investments. Using survey data on 2,047 individuals, we find that standard investments are strongly associated with both actual and perceived financial literacy for men, but only with actual literacy for women. Sophisticated investments, in contrast, are significantly related to perceived financial literacy with an even stronger association for women than for men. Interestingly, risk tolerance does not affect women’s sophisticated investments.
Gender Differences In Financial Risk Taking: The Role Of Financial Literacy And Risk Tolerance – Introduction
There is a large literature documenting a gender gap in financial risk taking (cf. Charness and Gneezy, 2012). This gap is of enormous economic importance: If women are less willing to invest in risky financial assets, they are expected to accumulate lower wealth over time. Combined with lower labour income and a longer life span on average, this renders women more vulnerable to poverty in old age.
The gender gap in stock market participation is usually explained either by women’s lower financial knowledge (Van Rooij et al, 2011; Lusardi and Mitchell, 2009), lower numeracy (Almenberg and Dreber, 2015), lack of familiarity with financial products (Prast et al., 2014) or lower risk tolerance (Croson and Gneezy, 2009; Dohmen et al., 2011). We extend this literature in three ways. First, we examine a larger investment universe by comparing the willingness to invest in risky but fairly standard financial assets (stocks and real estate funds) with the decision to invest in riskier, more sophisticated assets (discount certificates, hedge funds etc.). Second, we consider different dimensions of financial knowledge (actual and perceived financial literacy) to provide more nuanced insights on how literacy relates to financial risk taking. And third, we look at the combined role that financial literacy and risk tolerance play for men’s and women’s investment decision.
Results are derived from the SAVE panel, a representative survey on German households’ financial behavior. We find that women and men differ in the impact that financial literacy has on their investment decisions: For men both actual and perceived financial knowledge are positively (and individually) associated with standard investments, for women only actual literacy is. With respect to sophisticated investments, in contrast, perceived financial knowledge plays an important role for both men and women. Interestingly, the relation between sophisticated investments and perceived literacy is even stronger for women than for men.
We also observe a diverging role of risk tolerance. Higher risk tolerance increases the willingness of men to invest in both standard and sophisticated financial products. For women, in contrast, risk tolerance plays a role only for standard investments but not for sophisticated investments. Sound actual financial knowledge hence appears to be a prerequisite for women’s financial risk taking: It is positively associated with standard investments and thus helps to counterbalance the curtailing effect of women’s comparatively weak risk tolerance. Riskier, more sophisticated investments, in contrast, are driven by strong perceptions about financial knowledge. For women, this even renders risk tolerance irrelevant, while it is still an important factor for men.
Data and method
Our main data is from the 2009 SAVE panel, a representative survey of German households conducted by the Munich Center for the Economics of Aging (MEA).1 Our dataset consists of 2,047 responses. The survey contains a question on the subjective perception of financial literacy (on a seven-point scale), which we use as the perceived literacy score, and a set of nine objective questions related to basic numeracy and more advanced concepts of financial knowledge (see Appendix A). We take the sum of correct answers to these questions as the score of actual literacy. Respondents are
also asked to assess their willingness to take risks with respect to financial matters (on a ten-point scale). We employ this as our risk tolerance score.2 For a definition of variables, see Appendix B.
As dependent variables in our regressions, we employ the decision to invest in standard asset classes, which are defined as individual stocks, stock mutual funds and real estate funds, and to invest in more sophisticated asset classes such as discount certificates or hedge funds. We furthermore require sophisticated investments to be in addition to standard investments, so that respondents investing in the riskier, sophisticated assets have already been exposed to the experience of holding standard assets as well. Both types of investment decisions are coded as dichotomous 0-1 variables.
Table 1 in Appendix C reports the summary statistics for our dataset. The sample consists of 54% women. The average age is 53 years. Education is roughly equally divided between basic schooling, an intermediate degree and the highest degree that qualifies for tertiary education. 10% of respondents attained a university degree. Total gross wealth includes financial and real estate wealth.3 23% of respondents hold standard investments and 4% invest in sophisticated assets. These are 10% of the individuals that invest in standard assets.
Our main focus is on the relation of financial literacy and risk tolerance with the two types of investment decisions. In order to unveil the subtle effects that actual and perceived literacy may have, we use a composite, two-part measure of financial knowledge following Allgood and Walstad (2016). To this end, we first split the sample along the average actual financial literacy into the groups of high, respectively low, actual literacy. We then do the same for the perceived financial literacy to create the groups of high, respectively low, perceived literacy. From these two splits we create four distinct groups of respondents: perceived low / actual low (I), perceived low / actual high (II), perceived high / actual low (III), perceived high / actual high (IV).
Table 2 shows that men display significantly higher actual and perceived financial literacy than women. The table also reports the sample split into the four different literacy groups. Though we see that most respondents are in the perceived high / actual high group, the difference between men and women is here also particularly stark (44% vs. 29%). The perceived low / actual low group, in contrast, is the smallest group for men but the second largest for women. Interestingly, women are more numerous than men in the perceived high / actual low group, implying that it is more likely for women than for men to perceive their financial literacy as above average even though it is actually below average.
As may have been expected, risk tolerance in our sample is significantly higher for men than for women. To treat financial literacy and risk tolerance similarly in our analysis, we will employ dummy variables that indicate whether an individual holds above- or below-average risk tolerance. Splitting the sample along the mean, we hence find that 40% of all men display a risk tolerance that is above average, while only 29% of all women do.
See full PDF below.