Surveys consistently find that women hold a smaller proportion of their assets in investments and a larger proportion in cash than men do. This gap is real and it is costly, because cash loses purchasing power to inflation over long horizons while diversified investments have historically not.
The standard explanation is that women are more risk-averse and less confident about investing. Both claims have some support and both are frequently over-extended.
The performance finding
The most interesting complication comes from studies of actual investment behaviour rather than stated attitudes.
Analyses of brokerage account data have found that women's portfolios have in several studies performed comparably to or better than men's, after accounting for trading costs.
The proposed mechanism is trading frequency. Men in these samples traded more, and higher trading frequency is associated with worse net returns because transaction costs accumulate and because frequent trading correlates with worse timing.
This finding has been widely cited and it is worth stating the caveats: the effect sizes are modest, the samples are drawn from particular brokerages and periods, and the difference in returns is smaller than the popular retelling suggests. But the direction is clear enough to make the simple confidence story hard to sustain.
A person who trades less and holds longer is not being timid. She is doing what the evidence on retail investing recommends.
Disentangling risk aversion from other things
Measured risk aversion differences exist but are frequently confounded with three other variables.
Wealth. Risk capacity depends on how much you can afford to lose. Lower average wealth and lower income produce more conservative allocation for entirely rational reasons, and studies controlling for wealth find the risk-aversion gap narrows considerably.
Income volatility. More variable or interrupted income requires a larger cash buffer, which mechanically reduces the proportion invested.
Time horizon expectations. Someone anticipating a career interruption reasonably holds more accessible assets.
Once these are accounted for, a substantial part of what is described as a psychological difference turns out to be a rational response to a different financial position.
The financial advice problem
There is a supply-side factor that receives much less attention than the demand-side psychology.
Studies of financial advice have found differences in how advisers treat clients, including differences in the products recommended, the assumptions made about risk tolerance without asking, and the extent to which questions are directed to one member of a couple.
Audit studies in this area have found that identical client profiles receive different recommendations depending on client gender.
There is also a well-documented finding that a large majority of widowed women change financial adviser after their husband's death, which is generally interpreted as a signal about how the relationship was conducted while both were present.
An investing gap partly produced by an advice industry that engages less effectively with half its potential clients is a different problem from one produced by client psychology, and it points at a different fix.
The information environment
Financial media aimed at women has historically been oriented towards saving and spending discipline — cutting costs, budgeting, avoiding waste — while financial media aimed at men has been oriented towards accumulation and investment returns.
Content analyses of personal finance coverage have found this pattern consistently. The cumulative effect of a decade of reading about spending less rather than investing more is not neutral.
It also frames the problem incorrectly. Spending reduction has a floor; investment returns do not. For someone with a long horizon, the second is the larger lever by a considerable margin.
What actually closes the gap
Defaults, more than education. Auto-enrolment into workplace pensions with default investment allocations has done more to increase women's investment exposure than any amount of financial literacy programming, because it removes the requirement to make an active decision under uncertainty.
This is consistent with the finding elsewhere that gaps shrink when ambiguity is removed. The pattern recurs: differences in behaviour under ambiguity, largely absent when the default is clear.
Financial education has a mixed evidence base. Reviews have found that general financial literacy interventions have small effects on behaviour that decay over time, and that just-in-time education tied to a specific decision performs considerably better.
The implication is to teach at the moment of the decision — when the first pension enrolment happens, when the first surplus appears — rather than in a classroom three years earlier.
The advice that survives
Automate the decision so it is not made repeatedly. Standing transfers into a diversified low-cost fund require one decision rather than monthly ones.
Trade less. This is the one place where the evidence suggests the average woman's existing behaviour is the correct one and should not be adjusted.
Distinguish emergency cash from long-horizon money, and do not apply the same risk logic to both.
And treat any framing of the gap that locates the entire problem in women's confidence with appropriate scepticism, given that the portfolio data does not support the implied conclusion that the more confident approach performs better.