This project aims to investigate the gender gap in financial literacy. Survey data consistently shows that women are less financially literate than men (Lusardi & Mitchell, 2014). Women are less likely to correctly answer conceptual questions about the interest rate, inflation, and risk diversification, and are more likely to answer “don’t know”. This gender gap exists even when controlling for socio-demographic characteristics such as age and education. However, it is unclear whether this gender gap in financial literacy is due to gender differences in financial knowledge (women lack the knowledge to answer the questions) or gender differences in confidence (women are more likely to choose “don’t know” even if they know the answer).

To clearly identify the relative contribution of financial knowledge and confidence to this gender gap, we plan to conduct a large-scale online experiment that uses two treatments to create exogenous variation in each of these factors. In the "knowledge" treatment, participants view a 4-minute video that explains financial concepts. In the "confidence" treatment, participants receive feedback on their score in a pretreatment financial literacy test. Participants in the control group do not receive any information or feedback. The variation in knowledge and confidence created by this experimental design will help identify the relative contribution of each factor to the observed gender gap in financial literacy.

Our focus on measuring and decomposing the gender gap in financial literacy complements existing correlational studies on potential explanations for this gender gap, particularly gender differences in question-answering styles (Chen & Garand, 2018; Hospido et al., 2021), and studies on the correlation between this gender gap and real or self-reported financial behaviours (Bucher-Koenen et al., 2017; Lind et al., 2020; Lusardi & Mitchell, 2014). Some correlational studies aim to adjust for the confounding effects of confidence on performance in financial literacy tests by using econometric techniques (Bucher-Koenen et al., 2021; Ooi, 2020). In contrast to these correlational studies, our experimental setting allows us to investigate the causal impacts of interventions aimed at changing confidence and/or knowledge.

This project makes one methodological and two empirical contributions with potential applications in other contexts: we (1) introduce a survey tool that measures financial literacy in a confidence-independent way, (2) use an experimental setting to examine how confidence and knowledge affect the gender gap in financial literacy and real financial decisions, (3) investigate how responsiveness of these outcomes to confidence and knowledge treatments varies with gender and other socio-demographic characteristics.