Publicação
Three essays on financial literacy and financial behavior
| Resumo: | The lack of consistent evidence of a positive effect of financial literacy on financial behavior can be attributed to omitted variable bias, measurement error and reverse causality of financial literacy. This thesis presents insights to solve for the previously mentioned problems. Our study constitutes three papers on financial literacy and financial behavior. In the first study, we present an inter-temporal consumption model that integrates the effect of family ties on the individual’s choice level of investment in financial literacy. We posit that individuals with strong family ties tend to seek financial advice from their family network instead of improving their financial decision-making by investing in financial literacy. This allows individuals with strong family ties to substitute the costs of investing in financial literacy with consumption or saving. The result is a sub-optimal financial literacy level for individuals with strong family ties as opposed to ones with weaker family ties. Furthermore, based on the assumption that a higher financial literacy level allows the individual to make better financial decisions and thus earn a higher return on savings, individuals with strong family ties will have a lower return on savings, which in turn requires them to optimally save more than those with weaker family ties. The model drives our empirical approach to the analysis of the effect of family ties on financial literacy, and the joint effect of financial literacy and family ties on financial behavior. Using microeconomic panel data from SHARE WAVE 5, 6 and WAVE 3 SHARELIFE, we find strong support for the model’s predictions. After controlling for endogeneity, a strong level of family ties shows a significant negative effect on financial literacy. In addition, the initial level of financial literacy positively influences the current level of financial literacy. After applying IV regressions and Generalized Method of Moments (GMM), the effect of the initial level of financial literacy weakens in its effect on the current level of financial literacy. On the other hand, family ties show a significant negative effect on saving, investment in complex financial instruments, wealth and debt accumulation. In compliance with prior research, there is a positive effect of financial literacy on financial behavior. However, the effect weakens once we introduce family ties to our empirical model. After using IV regressions and GMM, the effect of financial literacy increases in its effect on our financial behavior proxies. These findings show that financial literacy is not the main determinant of financial behavior and suggest new implementations of financial education programs that address cultural differences if needed. In our second paper, we present a comparative cross-region study that explains individual financial literacy level through demographic, psychological and cultural factors, where cultural and psychological factors are introduced for the first time in a financial literacy study to control for omitted variable bias. We produce a comprehensive survey that collects all these factors and is applied to 600 individuals that can be subdivided into three regions: Arabic, Germanic and Latin Europe (nine countries). The data is analyzed using OLS regression with country fixed effects, Instrumental Variables Regression, Generalized Method of Moments (GMM), Principal Component Analysis (PCA), and Principal Component Regression (PCR). The regression findings show that financial literacy is positively associated with numeracy, financial socialization, Germanic countries and the psychological trait of confidence. In addition, financial literacy is negatively associated with the cultural traits of family ties and religiosity; the psychological trait of impatience, increased age and being female. Moreover, we apply Artificial Intelligence using machine learning algorithms which are Regression Trees, Random Forest and Quantile Regression Forests. Based on AI techniques using Regression Trees, Random Forest and Quantile Regression Forests, we find that financial literacy is mostly determined by geographic region, numeracy, impatience, confidence, financial socialization by parents, family ties and religiosity. These findings contribute to knowledge by suggesting new implementations of financial education program as an attempt to improve the effectiveness of financial literacy on financial behavior. These findings can be taken into consideration in constructing financial education programs that can be targeted towards focus groups based on gender, country, age, level of religiosity, family ties, impatience and confidence. In our third paper, we present a comparative cross-region study that explains individual five proxy measures of financial behavior level through financial literacy, financial socialization by parents, numeracy, demographic, psychological (impatience, risk-taking and confidence) and cultural factors (family ties and religiosity). Our five proxies of financial behavior are active savings, budgeting, punctual debt payment, stock market participation and retirement saving. We produce a query that aggregates all these factors and is applied to 600 individuals from Arabic countries (300 responses), Germanic countries (149 responses) and Latin European countries (151 responses). The data is analyzed using OLS regression with country fixed effects, IV regression, GMM, PCA and PCR. Moreover, we apply Artificial Intelligence using machine learning algorithms which are Regression Trees, Classification Trees, Random Forest and Quantile Regression Forests. We find that financial behavior is mostly determined by financial socialization by parents, impatience, confidence, risk-taking, financial literacy, numeracy, family ties and religiosity. Our findings show that financial literacy is not always the main determinant of financial behavior. The effect of numeracy, financial socialization, family ties, religiosity, impatience, confidence on financial behavior suggest that relying solely on financial literacy is not a solution to improve individual financial behavior. These findings suggest the need to identify new methods to improve financial behavior. This can be done using financial education programs and public policies that integrate psychological and cultural differences. In addition, the inability of financial education program to consistently improve financial behavior may result in new public policies that help the individual to attain financial well-being. |
|---|---|
| Autores principais: | Ramadan, lzzat Ibrahim |
| Assunto: | Financial Literacy Financial Behavior Religiosity Family Ties Literacia Financeira Comportamento Financeiro Religiosidade Laços Familiares Ciências Sociais::Economia e Gestão |
| Ano: | 2019 |
| País: | Portugal |
| Tipo de documento: | tese de doutoramento |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade do Minho |
| Idioma: | inglês |
| Origem: | RepositóriUM - Universidade do Minho |
| Resumo: | The lack of consistent evidence of a positive effect of financial literacy on financial behavior can be attributed to omitted variable bias, measurement error and reverse causality of financial literacy. This thesis presents insights to solve for the previously mentioned problems. Our study constitutes three papers on financial literacy and financial behavior. In the first study, we present an inter-temporal consumption model that integrates the effect of family ties on the individual’s choice level of investment in financial literacy. We posit that individuals with strong family ties tend to seek financial advice from their family network instead of improving their financial decision-making by investing in financial literacy. This allows individuals with strong family ties to substitute the costs of investing in financial literacy with consumption or saving. The result is a sub-optimal financial literacy level for individuals with strong family ties as opposed to ones with weaker family ties. Furthermore, based on the assumption that a higher financial literacy level allows the individual to make better financial decisions and thus earn a higher return on savings, individuals with strong family ties will have a lower return on savings, which in turn requires them to optimally save more than those with weaker family ties. The model drives our empirical approach to the analysis of the effect of family ties on financial literacy, and the joint effect of financial literacy and family ties on financial behavior. Using microeconomic panel data from SHARE WAVE 5, 6 and WAVE 3 SHARELIFE, we find strong support for the model’s predictions. After controlling for endogeneity, a strong level of family ties shows a significant negative effect on financial literacy. In addition, the initial level of financial literacy positively influences the current level of financial literacy. After applying IV regressions and Generalized Method of Moments (GMM), the effect of the initial level of financial literacy weakens in its effect on the current level of financial literacy. On the other hand, family ties show a significant negative effect on saving, investment in complex financial instruments, wealth and debt accumulation. In compliance with prior research, there is a positive effect of financial literacy on financial behavior. However, the effect weakens once we introduce family ties to our empirical model. After using IV regressions and GMM, the effect of financial literacy increases in its effect on our financial behavior proxies. These findings show that financial literacy is not the main determinant of financial behavior and suggest new implementations of financial education programs that address cultural differences if needed. In our second paper, we present a comparative cross-region study that explains individual financial literacy level through demographic, psychological and cultural factors, where cultural and psychological factors are introduced for the first time in a financial literacy study to control for omitted variable bias. We produce a comprehensive survey that collects all these factors and is applied to 600 individuals that can be subdivided into three regions: Arabic, Germanic and Latin Europe (nine countries). The data is analyzed using OLS regression with country fixed effects, Instrumental Variables Regression, Generalized Method of Moments (GMM), Principal Component Analysis (PCA), and Principal Component Regression (PCR). The regression findings show that financial literacy is positively associated with numeracy, financial socialization, Germanic countries and the psychological trait of confidence. In addition, financial literacy is negatively associated with the cultural traits of family ties and religiosity; the psychological trait of impatience, increased age and being female. Moreover, we apply Artificial Intelligence using machine learning algorithms which are Regression Trees, Random Forest and Quantile Regression Forests. Based on AI techniques using Regression Trees, Random Forest and Quantile Regression Forests, we find that financial literacy is mostly determined by geographic region, numeracy, impatience, confidence, financial socialization by parents, family ties and religiosity. These findings contribute to knowledge by suggesting new implementations of financial education program as an attempt to improve the effectiveness of financial literacy on financial behavior. These findings can be taken into consideration in constructing financial education programs that can be targeted towards focus groups based on gender, country, age, level of religiosity, family ties, impatience and confidence. In our third paper, we present a comparative cross-region study that explains individual five proxy measures of financial behavior level through financial literacy, financial socialization by parents, numeracy, demographic, psychological (impatience, risk-taking and confidence) and cultural factors (family ties and religiosity). Our five proxies of financial behavior are active savings, budgeting, punctual debt payment, stock market participation and retirement saving. We produce a query that aggregates all these factors and is applied to 600 individuals from Arabic countries (300 responses), Germanic countries (149 responses) and Latin European countries (151 responses). The data is analyzed using OLS regression with country fixed effects, IV regression, GMM, PCA and PCR. Moreover, we apply Artificial Intelligence using machine learning algorithms which are Regression Trees, Classification Trees, Random Forest and Quantile Regression Forests. We find that financial behavior is mostly determined by financial socialization by parents, impatience, confidence, risk-taking, financial literacy, numeracy, family ties and religiosity. Our findings show that financial literacy is not always the main determinant of financial behavior. The effect of numeracy, financial socialization, family ties, religiosity, impatience, confidence on financial behavior suggest that relying solely on financial literacy is not a solution to improve individual financial behavior. These findings suggest the need to identify new methods to improve financial behavior. This can be done using financial education programs and public policies that integrate psychological and cultural differences. In addition, the inability of financial education program to consistently improve financial behavior may result in new public policies that help the individual to attain financial well-being. |
|---|