Financial overconfidence, promotion of financial advice, and aging

Abstract

This paper proposes and validates a method to measure financial overconfidence in older people through incentivized economic experiments instead of answers to hypothetical questions. Specifically, financial overconfidence is quantified in terms of incorrect answers to a standardized financial literacy questionnaire on which participants reject the help of a financial advisor (overconfidence based on advice aversion) or declare that they are confident that they know the right answer (overconfidence based on confidence self-assessment). Using experimental data from a sample of 295 older Internet users in Spain, analysis shows that the two proposed behavioral measures of financial overconfidence are consistent. Moreover, the analysis confirms the well-documented finding of financial overconfidence in older people. The estimation of two analysis of covariance (ANCOVA) models suggests that, among older people, overconfidence decreases with age and education level (i.e., university studies). However, no statistically significant effect of gender on overconfidence level is found.