Journal of Economics, Management and Trade

Mutual funds delegate security selection to professional managers, but they do not remove behavioural demands from retail investors. Investors still decide whether to participate, which fund or fund family to select, how much risk to accept, when to contribute or redeem, and whether to switch after gains, losses or salient market events. This critical narrative review integrates three literatures that are often analysed separately: behavioural biases, subjective risk perception and financial literacy. Literature published from 1 January 1990 to 19 July 2026 was considered, while seminal earlier theory was retained when necessary for conceptual development. The synthesis indicates that mutual fund decisions are shaped by selective attention, representativeness and performance chasing, reference-dependent evaluation, disposition tendencies, overconfidence, familiarity and social cues. Their effects are not uniform: some patterns that appear behavioural can also arise from search costs, tax considerations, rational learning or heterogeneous participation costs. Risk perception is best treated as a dynamic translation mechanism between information and choice rather than as a synonym for stable risk tolerance. It is influenced by framing, affect, recent experience, disclosure format and investor knowledge. Financial literacy generally improves fee sensitivity, diversification and interpretation of risk, and it can attenuate some biases, but knowledge is not a universal debiasing device. Perceived literacy and confidence can diverge from objective knowledge, while generic educational interventions show heterogeneous downstream effects. The review therefore proposes an integrated process model in which information environments shape attention; attention and prior experience shape beliefs and perceived risk; literacy and confidence affect the quality of that translation; and realised outcomes feed back into future reference points and choices. The strongest evidence comes from transaction records and incentive-compatible experiments, whereas much recent emerging-market evidence relies on cross-sectional surveys and investment intentions. Future work should combine longitudinal account data, validated psychometrics and randomised disclosure interventions to identify mechanisms and boundary conditions more credibly.