Reconciling Self-Assessed with Psychometric Risk Tolerance: A New Framework for Profiling Risk among Investors

Abstract

AbstractFinancial advisors need to assess their clients’ risk profile to properly manage their portfolio risk and comply with regulatory provisions. Assessing an investor’s financial risk tolerance (FRT) is a challenge in the advisory process and none of the existing measures can be easily employed on a large scale. Previous literature has revealed a gap between self-assessed and psychometrically assessed measures of FRT (PA_FRT) but has not yet offered a solution to fill this gap. Thus, we propose a model that consistently estimates the PA_FRT by leveraging retail investors’ self-assessment and other information typically submitted in standard bank questionnaires. Our model represents a promising tool for financial advisors looking to improve their customers’ risk profiling.Keywords: Financial risk toleranceInvestment behaviorInvestor’s preferencesPsychometric measuresSelf-assessmentJEL CLASSIFICATION: C81C83D14D81D83G11 Disclosure statementNo potential conflict of interest was reported by the authors.Notes1 Regulators in Europe and the US have introduced provisions in order to protect investors: Since 2004, the European Markets in Financial Instruments Directive (MiFID) have required investment banks offering financial advice to assess the suitability of retail investors’ portfolios; likewise, since 2010, the US FINRA rule 2111 states that investment firms and their associated persons must have a reasonable basis to believe that a recommended transaction or investment strategy involving securities is suitable for the customer. Assessing a client’s risk profile is therefore a challenge that financial advisors must tackle in their daily activity, both to provide clients with high-quality services and to be compliant with regulatory provisions.2 ESMA (Citation2022), Guidelines on certain aspects of the MiFID II suitability requirements. Paragraph 48: “When assessing the risk tolerance of their clients through a questionnaire, firms should not only investigate the desirable risk-return characteristics of future investments, but they should also take into account the client’s risk perception. To this end, whilst self-assessment for the risk tolerance should be avoided, explicit questions on the clients’ personal choices in case of risk uncertainty could be presented. Furthermore, firms could for example make use of graphs, specific percentages or concrete figures when asking the client how he would react when the value of his portfolio decreases”.3 ESMA (Citation2022) identifies “Defining a client’s risk tolerance solely based on the composition of such client’s existing portfolio” as a poor practice, see page 72.4 The sampling was specifically aimed at ensuring that the overall sample was representative of the entire population of the bank’s customers in terms of socio-demographic characteristics (geographical areas/cities, age), risk profile and financial knowledge. As we refer to a big Italian bank that serves many customers across Italy, our final sample should be representative of Italian retail investors.5 The PA_FRT score has been reparametrized in the scale from 0 to 10 to be comparable with the SA_FRT.6 While our model presents endogeneity, we want to note that the literature on questionnaires commonly applies this kind of methodology to find the best set of items: The same G&L (Citation1999) that we employed in our paper used this method to reduce its original 100 items to 13. What we try to show in Table 4 is that, thanks to other information that the bank has including the SA_FRT, even just having a few of those 13 items we can produce a powerful model.7 In order to obtain a parsimonious model, we performed a stepwise regression with a p-value threshold for inclusion of 0.05.8 The items from the Grable and Lytton (Citation1999) questionnaire are added in the sequential order for maximizing the R-squared of the model.