The Impact of Psychological Characteristics on
Financial Assets Holding Behavior: An
Empirical Study
By:Xuejiao Zhou1
University of Groningen Faculty of Economics & Business
MSc. Finance
Supervisor: Dr. A.M. (Annika Maxine) Mueller Date: 08-06-2017
Abstract:
This paper is an attempt to explore the impact of psychological characteristics on financial asset holding behavior from an empirical perspective, using a sample of 2030 Dutch household individuals obtained from the DHS data wave 2016 (the 24th wave) that is conducted over the period April 2016 – October 2016. This paper examines the relationship between personalities and household finances focusing on financial decisions regarding financial assets holding amounts. Personality traits are classified into nine categories: conscientiousness, extroversion, agreeableness, neuroticism, openness to experience, risky averse, risk loving, concern present and concern future. The conscientiousness and openness to experience have been found relatively large effects on financial assets holding behavior, while extroversion and agreeableness appear to be least important in affecting financial asset holding. The findings also indicate that different personality traits have different influences across various types of financial assets.
Field Keywords: Behavioral finance, psychological characteristics, Big five personality traits, Financial assets holding behavior, Multiple regression analysis, DHS data sets.
Word count: 11,330
1. Introduction
A relatively new stream of research in the household behavioral finance literature is about the impact of personal characteristics on financial decisions over the last decades. It is obviously that the personal characteristics may influence any decisions made by individuals including financial decisions as well. Therefore, it is important to know what kind of individual’s personal characteristics can influence the financial decisions, and in which way the personal characteristics can influence it.
According to the prior studies, some personality characteristics have been proved as important determinants of finance decisions. The most fundamental and universal classification of personality characteristics is the “Big Five” personality traits developed by Costa and McCrae (1992). They classified the personality traits into 5 different categories: openness to experience, conscientiousness, extroversion, agreeableness and neuroticism. Most of the prior studies have shown that openness to experience and extroversion had a significantly impact on household finance decisions. In addition, some other prior studies measured the personality characteristics by using other criteria rather than Big Five personality traits. Becker et al. (2012) stated that in the standard economic approach, individuals’ behavior would be affected by risk aversion, time preference and social preference, in combination with beliefs, expectations, and strategic considerations. These studies have shown that the other personality measurement can also affect the financial behavior significantly, which motivated us to include other personality measurements available in this study as well for further understanding the relationship between the psychological personality and financial behavior.
research, data sets of the DNB Household Survey (DHS) is used for obtaining the required data in order to investigate the possible relation between personality and financial assets holding behavior.
The outcome of this research shows that conscientiousness and openness to experience are the only two personality traits to exhibit consistent results. Conscientiousness exhibit to have relatively large positive effects on the amount of financial assets held. Openness to experience appears to have large negative effect on the amount of financial assets held. On the other hand, extroversion and agreeableness are the two personality traits that consistently fail to exhibit a significant effect on financial asset holding, indicating that these two personality traits are not important in influencing the Dutch household individual’s financial behavior. Neuroticism consistently exhibit to have strong significantly negative effect on financial assets held if the control variables are not included in the regression models. Risk-averse is found to have a negative effect on risk-free assets holding, but no significant effect on risky assets holding. Risk loving is found to have significant positive effect on financial assets held. The level of risk-free assets held is increasing if Dutch household individuals concerning about the future, and decreasing if individuals concerning about the present. The risky assets held are found to have no significant association with time preference. However, all the risk preferences and time preferences have strong effect on the probability of holding risky assets.
2. Literature Review
This section discusses existing literature and research papers that are relevant for this research. Relevant literatures are helpful to answer the research question before the research has been executed. Based on this literature the research questions will be developed. Furthermore, the definition of the research objectives and hypotheses are included in this section.
2.1. Household financial assets
Household assets can be categorized as financial assets and nonfinancial assets as defied by the International Financial Reporting Standards (IFRS) (Kennickell & Starr-MacCluer, 1994). A financial asset is a tangible liquid asset that derives value because of a contractual claim, such as stocks, bonds, bank deposits. Financial assets are usually more liquid than tangible assets, such as land, property, commodities or other tangible physical assets. A household may consider different categories of assets to be used for different purposes. Xiao and Anderson (1997) stated that Households held varying types and differing levels of assets to meet their financial needs and many of them would hold both financial assets and nonfinancial assets at the same time.
Weagley and Gannon (1991) categorized household assets into four groups: savings, financial security, retirement investments, and housing. They found that the diversification of household asset portfolios toward riskier investment categories increased as wealth increased. As asset levels increased, investments in financial securities and for retirement increased relative to the level of savings. Using Maslow's and other theories, Xiao and Olson (1993) divided household financial assets into three different groups presenting as below:
- Group 1: checking and saving accounts, certificates of deposit, and money market accounts.
- Group 2: individual retirement accounts, Keogh plans, various saving plans, and other financial assets.
- Group 3: mutual funds, bonds, and stocks, excluding ones in Groups 1 and 2.
2.2. Big Five personality traits
The Big Five model is the most widely used taxonomy of personality traits. Sir Francis Galton (1884) first derived the comprehensive taxonomy of human personality traits. Raymond Cattell (1940) constructed the personality traits named “Sixeen Personality Factor Questionnaires”. After that, Lewis Goldberg (1980) started his lexical project with emphasizing five broad factors and coined the term “Big Five” as a label for the factors. Goldberg & Digman et al. (1980) reviewed the available personality instruments in a 1980 symposium. Then the five-factor model has been widespread acceptance among personality researchers during the 1980s. Costa and McCrae (1992) developed the Big Five personality traits and made it been fundamental and universal used by researchers to measure personality. The explicitly description of the Big Five personality traits are as below:
Conscientiousness indicates the individual’s willingness to follow rules and
Extroversion encompasses the preference for human contact, attention and the wish
to inspire other people. People who score high tend to be outgoing and energetic and people who score low are more solitary and reserved.
Agreeableness is the willingness to help other people and to act in accordance with
other person’s interests. Individuals who score high are more friendly and compassionate where a low score means a more analytical and detached person.
Neuroticism indicates adjustment versus emotional stability and addresses the degree
to which the individual is insecure, anxious, depressed and emotional rather than calm, self- confident and cool. People who score high tend to be sensitive and nervous while people who score low are more secure and confident.
Openness to experience includes a person’s curiosity and the tendency for seeking
and appreciating new experiences and novel ideas. People who score high are curious, open to emotion, and willing to try new things. People who score low are more consistent and cautious.
2.3. Psychological personality and financial assets holding behavior
Borghans et al. (2008) stated “An integration of the different measures and concepts used by economists and personality psychologists promises much potential for amalgamating evidence about the drivers of human behavior that accumulated disjointedly in the fields of economics and psychology.” Borghans et al. (2008) have shown that scholars have begun to integrate personality into economic decision-making recently. Almlund et al. (2011) enrich theory by incorporating personality traits in a standard economic framework of production, choice, and information. After that, the relationship between personality psychology and economic preference has been extensively studied.
characteristics had been confirmed that it will have a significantly influence on outcomes, there were still a lot of papers have put effort into investigating the unexplored relationship between the economic behavior and psychological measures of personality within the recent years. Some articles focus on the relationship between the personality and the life outcomes. Browning and Lusardi (1996) discuss some analyses of saving behavior over the life cycle. And Borghans et al. (2008) examined the predictive power of personality and the stability of personality traits over the life cycle as well. Almlund and Heckman (2011) measured personality traits are positively correlated over the life cycle. Several articles focus on the relationship between the demography-related factor and the financial behavior. Bertaut and Starr-McCluer (2000) found out significant impact of age, wealth, income risk, and entry/information costs on determining the composition of households' assets and liabilities in the United States. Borghans et al. (2009) demonstrated gender differences in risk aversion and ambiguity aversion. Brown & Taylor (2008) suggested that the poorest and the youngest households are the most vulnerable to adverse changes in their financial circumstances.
Meanwhile, psychological factors related to individuals’ personality seem to play an important role also with regard to people’s income level: Anger and Heineck (2010) provided the first joint evidence on the relationship between individuals' cognitive abilities, their personality and earnings for Germany. Nandi and Nicoletti (2014) estimated the effect on pay of each of the Big Five personality traits for employed men living in the UK. Taylor and Brown (2014) explored the relationship between household finances and personality traits using individual data drawn from the British Household Panel Survey. They analyzed the influence of personality traits on financial decision-making at the individual level focusing on decisions regarding unsecured debt acquisition and financial assets.
(Borghans et al., 2009). The most relative result compared to this study it that the openness to experience has been defined as the most relevant personality trait in explaining the financial outcomes. Nandi and Nicoletti (2014) found that openness to experience is the most relevant personality trait in explaining wages, followed by neuroticism, agreeableness, extroversion and conscientiousness. Openness and extroversion are rewarded while agreeableness and neuroticism are penalized, but the openness pay gap is totally explained by differences in worker characteristics, particularly education and occupation. Taylor and Brown (2014) found that extroversion and openness to experience exert relatively large influences on household finances in terms of the levels of debt and assets held. In contrast, personality traits such as conscientiousness and neuroticism appear to be unimportant in influencing levels of unsecured debt and financial asset holding. Their findings also suggest that personality traits have different effects across the various types of debt and assets held.
Some of the prior studies measured the personality characteristics by using other criteria rather than Big Five personality traits. Becker et al. (2012) stated that in the standard economic approach, individuals’ behavior would be affected by risk aversion, time preference and social preference, in combination with beliefs, expectations, and strategic considerations. Brown et al. (2003) showed that optimistic financial expectations impact positively on both the quantity of debt and the growth in debt, at the individual and household levels. These studies have shown that the other personality measurement can also affect the financial behavior significantly, which motivated us to include other personality measurements available in this study as well for further understanding the relationship between the psychological personality and financial behavior.
research further with the demography-related and human capital-related factors as well. The risk and time preferences factors will also be explored in this research since they are the key determinants for the financial behavior.
2.4. Research Questions
Regarding our research target and the literature review results the following main research question seems relevant:
What is the impact of psychological characteristics (Big Five traits) on financial
assets holding behavior?
In order to answer this main research question the following sub-questions should be considered:
Is there significantly difference between the impacts of the Big Five personality traits on the riskless and risky financial assets?
Will the demographic factors (age, educational level, etc.) influence the impact of the Big Five personality traits on financial assets holding behavior?
Will the human capital-related factors (career background, etc.) influence the impact of the Big Five personality traits on financial assets holding behavior?
3. Data and Methodology
This section contains the data collection and research method. It is explained how the data is processed and analyzed in theory.
3.1. Data Acquirement
The required data will be obtained from the DNB Household Survey from the CentERdata website.2 CentERdata conducts the DNB Household Survey annually since 1993 in order to study the economic and psychological determinants of the saving behavior of households in the Netherlands. The DNB Household Survey consists of 6 questionnaires totally, including General Information on the Household; Household and Work; Accommodation and Mortgages; Health and Income; Assets and Liabilities; and Economic and Psychological Concepts. This research will use the data from 5 questionnaires of them (shown in Table 1) to investigate the impact of psychological characteristics on the financial assets holding behavior in an individual level.
The latest DHS data wave 2016 (the 24th wave) conducted over the period April 2016 – October 2016 will be chosen as a database to analyze the research topic in order to study the latest relationship between psychological characteristics and the financial assets holding behavior. The questionnaires are presented to each member of the CentERdata aged 16 or over of which 2231 households have participated in the 24th wave. The response rate of each selected questionnaire is presented in the table below. Besides, the section General Information on the Household includes all members of the households.
Table 1: Response rates of the data sets
Subject No. of persons No. of households
General information on the household Household and work
Health and income
5307 2415 2355 2231 1914 1872
Assets and liabilities 2243 1789
Economic and psychological concepts 2694 2128
*Source: DHS data wave 2016 codebooks
The data from these five different files are merged by using two variables: nohhold (household index) and nomem (index of the member of the household). As nomem is always smaller than 100, a unique personal index can be generated as follows:
!"#$!% !" = !"ℎ!"#×100 + !"#$#
There are 5,315 observations in total after merging the data, while only 2,030 out of total observations exist among all of the five files simultaneously. Therefore, the other 3,285 observations cannot provide us comprehensive data and should be dropped in this study.
3.2. Methodology
This paragraph explains the dependent variables, independent variables, and regression models.
3.2.1 Dependent variables
The abstract dependent variable is the financial assets holding behavior while the financial assets including checking accounts, saving accounts, annuity insurance, mutual funds, bonds and shares. All of these financial assets holding are measured up to the end of year 2015 in Euro. For example, the question “What was the total balance of your SAVINGS OR DEPOSIT ACCOUNTS on 31 December 2015? Type -99 if you don’t know the answer.” is asked. If respondents type -99 as they don’t know the answer, a second question is asked with saving amount in different categories. After that, the mean of these categories will be used as an input for the dependent variable value. In order to distinguish the different effects of personality traits on various types of financial assets, the risky assets and risk-free assets are generated as dependent variables as well. The risky assets are consisted of mutual funds, bonds, and shares, while the risk-free assets are consisted of checking accounts, saving accounts and annuity insurance.
3.2.2. Independent variables
traits. However, risk preference and time preference will also be included in this study as the measurements of individual’s psychological characteristics. Since the prior studies showed that the individuals’ time preferences and attitudes toward risk are significantly important for economic models. Based on Wölbert and Riedl (2013), time preference and risk preference played an important role in research on human decision-making in other social sciences, like psychology. Therefore, the risk preference factor and time preference factor are included in this study as well to test if these factors will determine the financial behavior in a significantly way.
To measure the Big Five personality traits, respondents are asked to rate themselves on a five-point Likert scale3 of total ten different statemetns (1 means “not at all applicable”, 5 means “highly applicable”). There are two statements for each of the big five personality traits that are displayed in detail as below:
Table 2: DHS statements of Big Five personality traits
Big Five Personality Traits DHS Statements
1. Conscientiousness - I like order
- I leave my belongings around. (Reverse) 2. Extroversion - I keep in the background. (Reverse)
- I’m quiet around strangers. (Reverse) 3. Agreeableness - I sympathize with others’ feelings.
- I take time out for others.
4. Neuroticism - I get stressed out easily.
- I have frequent mood swings. 5. Openness to experience - I have a vivid imagination.
- I have excellent ideas.
*Source: DHS data wave 2016 codebooks
As can be seen that three out of ten statements in Table 2 are presented in reverse way, thus, this study reverse the grades of these three statements (from 1-5 to 5-1) in order to add each two statements of different Big Five personality traits up conveniently.
Besides, individuals are also asked to rate themselves on two seven-point scales: risk preference scale and time preference scale. The detailed questionnaires for risk preference and time preference can be found in Appendix B1 and Appendix B2, respectively. This study classifies the risk preference into risk-averse group and risk
loving group, and also classifies time preference into concern present and concern future groups to distinguish the different magnitudes of their effects on financial assets holding.
Because this study is mainly focused on the individual level, we also would like to investigate other non-individual factors that may affect the relationship between the Big Five personality traits and financial assets holding behavior, such as, the regions/cities of the respondents. In this study, a dummy variable of regions/cities will be added in the regression formula to test if the regions will affect the relation or not. Meanwhile, the demography-related variables: age, gender, and marital status will also be tested in this study. For the other human capital related factors that may be affected by the personality characteristics, such as, education background variable, may lead to an inaccurate result because they may mediate the effect of independent variables on dependent variables. Thus, a hierarchical multiple regression analysis should be applied here to avoid this potential problem. The variable definitions are presented in Table 3 as below:
Table 3: Variables Definition
Variable Definition
Dependent variables {Measured in Euros.} Checking accounts Saving accounts Annuity insurance Mutual funds Bonds Shares
Total financial assets
Independent variable
Openness to experience Measured from 1 completely closed, to 5 totally open. Conscientiousness Measured from 1 totally disorganized, to 5 completely
conscientious.
Extroversion Measured from 1 totally extravert, to 5 completely introvert. Agreeableness Measured from 1 totally disagreeable, to 5 completely
agreeable.
Neuroticism Measured from 1 totally calm/relaxed, to 5 completely nervous/ high-strung.
Age
Age-squared
Age in years.
Age-squared in years. Marital status Dummy=1 if married.
Education4 1=low education, 2=middle education, 3=high education. Degree of urbanization Measured from 1 very high degree of urbanization, to 5
very low degree of urbanization. Health
Income Region
Risk preference5 Time preference6
1= poor health, 2=fair health, 3=good health, 4=excellent health.
1=low income, 2=average income, 3=high income. Dummy = 1 if the respondent lives in the three largest cities.
Measured from 1 totally risk averse, to 7 totally risk loving. Measured from 1 totally cares about present, to 7 totally cares about future.
3.2.3 Regression analysis
At the beginning, the correlation matrix will be conducted including all nine different personality traits in case there might be auto-correlation existing among the Big Five personality traits, risk preference and time preference. After that, the Spearman correlation estimation will be applied to test whether the correlation between risk preference/time preference and Big Five personality traits are significantly strong or not. The logarithm of financial assets holding amount is applied before the regressions are conducted so that the coefficients can be expanded to a reasonable range, since all of the independent variables are measured in points of different categories.
regression the control variables are added to the regression to investigate whether the control variables can also affect the financial assets holding behavior and to explore whether the effect of Big Five personality traits on financial assets holding will change or not.
After that, the Big Five personality traits will be regressed on total financial assets one by one as a robustness analysis. The regression will be conducted without control variables firstly, and then add all the other control variables to make a comparison with the first regression.
To distinguish the different effects of personality traits on various types of financial assets, the risky assets and risk-free assets are classified and regressed as dependent variables using OLS regression firstly. Then this study will generate the dummy variables for total assets as 1=risky assets and 0=risk-free assets. Afterwards, both the OLS and Probit analyses will be conducted to model binary outcome variables.
At the end, this study will regress the independent variables (including control variables) on each individual financial asset respectively to test whether the effect of personality traits will differ on various types of assets.
The simple OLS regression model is:
!"#$% !"#$#%"$& !""#$"! = ! + !!! + !!
where the ! is the constant term, !! are the Big Five personality traits, !! is the error term, !! is the coefficients of term !!.
The multiple regression base model is:
!"#$% !"#$#%"$& !""#$"! = ! + !!!! + !!!! + !!
where ! is the constant term, !! are the Big Five personality traits, !! are the control variables, !! is the error term, and !! and !! are the coefficients of term !! and !!, respectively.
4. Empirical Results
This section presents the data analysis and demonstrates the empirical results of the research.
4.1. Histogram of logarithm of total financial assets
Figures 1 shows the distribution of the natural logarithm of the total value of financial assets. It can be seen that, although the distribution is closed to normal distribution, there are outliers existing that may lead the distribution skewed. Therefore, the Kernel density estimation7 and graph box analysis8 are applied to further investigate if the logarithm of financial assets holding amount is normally distributed or not. The Kernel density plot indicates that the distribution does not look normal. As seen from the boxplot, the dots at the end of the boxplot indicate possible outliers. The boxplot also confirms that the distribution is skewed towards the left. Thus, this study will also conduct Probit analysis besides OLS analysis to increase the reliability of the study.
Figure 1: Histogram of logarithm of total financial assets
7 See Appendix A1. 8 See Appendix A2. 0 .05 .1 .15 .2 .25 D e n si ty 0 5 10 15
4.2. Descriptive analysis
The descriptive summaries for financial assets and the independent variables are presented in Table 4 as below:
Table 4: Descriptive summaries
Variable Dummy Obs. Mean Std. Min Max
Dependent variables Total assets No 2030 31919 246670 -75 10000000 Checking accounts No 2030 2062 14621 -75 594224 Saving accounts No 2030 16312 40509 0 579344 Annuity insurance No 2030 6118 222370 0 9999990 Mutual funds No 2030 5613 89025 0 3862576 Bonds No 2030 486 7159 0 232000 Shares No 2030 1328 11934 0 260000 Control variables Male Yes 2030 0.540 0.499 0 1 Married Yes 2030 0.622 0.485 0 1
Big three cities Yes 2030 0.143 0.350 0 1
Age No 2030 55 17 16 93
Education Yes 2030 4.878 1.507 1 9
Education: middle Yes 2030 0.576 0.494 0 1
Education: high Yes 2030 0.141 0.348 0 1
Urbanization Yes 2030 3.018 1.315 1 5
Urbanization: average Yes 2030 0.215 0.411 0 1
Urbanization: high Yes 2030 0.391 0.488 0 1
Health Yes 2030 2.180 0.724 1 5
Health: fair Yes 2030 0.200 0.400 0 1
Health: good Yes 2030 0.625 0.484 0 1
Health: excellent Yes 2030 0.126 0.331 0 1
Income Yes 2030 3.757 1.325 0 6
Income: average Yes 2030 0.436 0.496 0 1
Income: high Yes 2030 0.261 0.439 0 1
Neuroticism No 2030 4.600 1.937 2 10 Openness to
experience No 2030 6.688 1.625 2 10
Approximately 54% of individuals in the sample are male; 62.2% are married; 62.5% are in good health and 43.6% has an average income. The observation number of risk averse variable and risk loving variable are 1874 due to the existing of missing data. The average age of respondents is 55 years old with the youngest age is 16 and oldest age is 93. As observed from the table above, respondents are holding financial assets quite differently. The average level of financial assets till the end of 2015 is 31,919 Euro, while the highest amount reach to 10 million. The lowest amount of total financial assets reach to -75 Euro since one respondent is holding -75 Euro in checking amount. The average level of annuity insurance is 6,118 Euro, while the highest amount achieve to 9,999,990. And the average level of mutual funds is 5,613 Euro, while the highest amount reach to 3,862,576 Euro. These data demonstrates the potential outliers that may lead the distribution showed in Figure 1 skewed towards left.
4.3. Correlation analysis
This paragraph analyzed the correlation among the different personality traits including Big Five personality traits, risk preference and time preference.
4.3.1. Correlation analysis on total nine personality traits
risk-loving, the point of concern future will increased by 0.253 point. There are only 4 out of 36 correlations exceed 0.2 in absolute value and 15 out of 36 correlations exceed 0.1 in absolute value. The correlation between risk averse and risk loving is -0.217, which is the second highest correlation among all the correlations in absolute value. The correlation between concern present and concern future is -0.113. Although these two correlations do suggest the negative association between the reverse factor of risk preference and time preference, the correlations in absolute value is too small. Besides, the other correlations are even much smaller. Therefore, this study finds no evidence for a strong auto-correlation between each pair of the nine different personality traits. We can conclude that it is feasible to include all of these nine different personality traits in the regression model at the same time.
4.3.2. Spearman correlation structure
Borghans et al. (2008) stated that an integration of the different measures and concepts used by economists and personality psychologists provides much potential evidence about the drivers of human behavior in the fields of economics and psychology. To shed more light on the relationship between economic preferences and personality traits the student therefore studies how key economic preferences, such as risk and time preferences, are linked to conventional measures of personality, such as the Big Five model.
Table 5: Spearman correlation structure Risk-averse Conscientiousness 0.0751*** Extroversion -0.0502** Agreeableness 0.1568*** Neuroticism -0.0277 Openness 0.059** Risk-loving -0.0831*** -0.0206 -0.1311*** 0.0935*** 0.1029*** Concern present -0.023 -0.0464** 0.0064 0.0023 0.0236 Concern future 0.0505** -0.0246 0.1342*** 0.131*** 0.1977***
In Table 5, the correlations between the outcomes from the risk and time experiments and the personality traits are displayed. Table 5 shows the 20 raw correlations of the Big Five personality and economic preference measures. Table 5 reveals that 13 out of 20 correlations are statistically significant at the 5-percentage level or 1-percentage level. All correlation coefficients are smaller than 0.2 in absolute value. Thus, there is no correlation with a “medium” effect size or larger. Furthermore, of the 20 raw correlations only 6 exceed 0.1 in absolute value.
As the table above, the higher values of time preference concerning about future indicate higher patience, while the higher values of time preference concerning about present indicate lower patience. Only nine correlations are significant at the 1-percentage level, four are significant at the 5-percentage level and none is significant at the 10-percentage level. In terms of effect size on agreeableness, the coefficients of the association between agreeableness and risk preferences, agreeableness and time preference concerning about future all exceed the 0.1 benchmark level, which is categorized as a small correlation by Cohen (1988). Interestingly, the sign for risk-averse and risk loving are positive and negative respectively, that are in line with the existing study (Becker et al., 2012). In terms of effect size on neuroticism, only the coefficient of the association between neuroticism and time preference concerning about future exceed the 0.1 benchmark level. In terms of effect size on openness to experience, only the coefficients of the association between openness to experience and risk loving, openness to experience and time preference concerning about future exceed the 0.1 benchmark level. The positive sign of risk loving associated with openness to experience is in line with Becker et al. (2012).
for a strong associated between the economic preferences and the Big Five personality traits. Therefore, it can be concluded that these two concepts cannot substitute for each other. The low association between economic preferences and psychological measures of personality can be reasonable, according to Becher et al. (2012), the concepts are formulated in quite different ways. Whereas preferences are rooted in utility theory and the Big Five personality traits are originated in language analysis. Moreover, the Big Five measure rather broad aspects of personality. Therefore, although the result in Table 5 above shows low associations between Big Five personality traits and economic preferences, the possibility that there is a stronger level of association between them still exists. Meanwhile, the prior studies have shown that the time preference and risk preference had significant effect on financial assets behavior. Considering the results in Table 5 and prior studies, the significant effects of economic preferences on financial assets should not be omitted. In short, it is necessary to add the time and risk preferences in this study to enrich the research.
4.4. Determinants of the amount of financial assets held
Table 6 presents the results from the OLS analysis relating to the determinant of the amounts of the financial assets held for the Dutch household, with Panel A excluding other controls and Panel B including other controls. In Table 6, all of the five personality variables are included simultaneously, whilst in Table 7 and Table 8 the Big Five variables are entered one by one.
associated with a 9.7 percentage point decrease in the amount of financial assets held. Hence, our findings suggesting that being worried, insecure, depressed, overly anxious and angry is negatively associated with financial assets held. Conscientiousness also has an effect on the amount of financial assets held, with a significant level at 5 percent. For the sample of Dutch household individuals, a one standard deviation increase in conscientiousness is associated with a 7.3 percentage increase in financial assets accumulation, suggesting that being persevering, organized, responsible, dependable, thorough and industrious is positively associated with the amount of financial assets held. A striking finding is that openness to experience has a negative effect on financial assets, with a significant level at 10 percent, suggesting that being imaginative, intelligent, curious and flexible is negatively associated with the financial assets accumulation. For example, for the sample of Dutch household individuals, a one standard deviation increase in openness to experience is associated with a 6.4 percentage point decrease in the amount of financial assets held.
Table 6: Determinants of total financial assets holding Variables Panel A Panel B Coef. Std. Coef. Std. Intercept 9.560*** (0.512) 5.856*** 0.858 Conscientiousness 0.073** (0.033) 0.061* 0.032 Extroversion 0.036 (0.032) 0.008 0.030 Agreeableness -0.060 (0.041) -0.053 0.041 Neuroticism -0.097*** (0.031) -0.011 0.032 Openness to experience -0.064* (0.037) -0.077** 0.036 Risk-averse 0.017 0.015 Risk-loving 0.036** 0.016 Concern present -0.023** 0.009 Concern future 0.031*** 0.011 Male 0.277** 0.119 Married -0.172 0.124
Age 0.004 0.021 Age-squared 0.000 0.000 Education: middle 0.313** 0.133 Education: high 0.953*** 0.188 Urbanization: average -0.279** 0.143 Urbanization: high -0.168 0.130 Health: fair 0.407 0.282 Health: good 0.727*** 0.264 Health: excellent 0.879*** 0.299 Income: average 0.682*** 0.147 Income: high 0.902*** 0.167 Control No Yes Adj. R-squared 1.55% 16.20% OBSERVATIONS 1434 1360
Note: The standard errors are displayed in parentheses. ***, **, * denotes statistically significant differences at 1%, 5% and 10% level respectively.
Focusing upon the results in Panel B, it is apparent that conscientiousness and openness to experience are the only two personality traits to exhibit consistent findings compared to Panel A. However, the statistical significant power of the effects of these two personality traits on financial assets is weaker than Panel A. Meanwhile, in Panel B, the effect magnitude of conscientiousness on financial assets is decreased by 1.2 percentage compared to Panel A, while the effect magnitude of openness to experience on financial assets is decreased by 1.3 percentage. Furthermore, the neuroticism trait has no longer a significant influence on financial asset holding behavior.
Turning briefly to the other personality variables after focusing on the effect of the Big Five, it is observed that, the level of financial assets is increasing if Dutch household individuals being risk-loving and concerning about the future, and decreasing if individuals concerning about the present.
healthier body will hold more financial assets than the others. Interestingly, the urbanization level of the city has negative effect on financial assets, where those Dutch household individuals living in lower level urbanization city have lower level of financial assets. Since age has no significant effect at all, this study will not compare the two different age groups (18<age<30, age over 30), which is applied by previous study (Brown. S. & Taylor. K., 2011). In terms of income, focusing upon the full sample, household income from employment has a large and positive effect on financial assets. For instance, a one standard deviation increase in household individual income is associated with a 0.682 (0.902) percentage point increase in financial assets for average level of income (high level of income). These findings are in line with the findings in the previous literature, for instance, Gropp et al. (1997), Crook (2001) and Brown and Taylor (2011).
The adjusted R-squared is 1.55 percentage and 16.20 percentage for Panel A and Panel B respectively, which tells that only 1.55 percentage and 16.20 percentage of the variations in financial assets is explained by these two regression models. The regression model with controlling explained the error much better than the model without controlling.
4.5. Robustness analysis
This study also explored the robustness of the results if each of the five personality variables is included separately rather than simultaneously in the OLS analysis. The broad pattern of results described above is also found when each of the five personality variables is entered separately, see Table 7 and Table 8 below. In Table 7, all of the other controls are excluded while in Table 8 are included.
household individual’s financial behavior. Conscientiousness and openness to experience are the two personality traits to exhibit consistent findings across the different regression methods with positive and negative effect with limited statistically significance, respectively. Furthermore, it is apparent that the statistically significant level of the effect of conscientiousness and openness to experience on financial assets holding in the separately regression model are relatively stronger than in the regression model that entered the Big Five simultaneously. Neuroticism appears to exhibit a strong statistically significant negative effect on financial assets holding without controlling for other factors. Because neuroticism is related to pessimism, such finding is also interesting in the context of the positive association found in the previous study relating to optimism personality and financial assets.
Table 7: OLS regression of Big 5 traits on total assets without control variables
Big 5 traits Trait 1 Trait 2 Trait 3 Trait 4 Trait 5 Simultaneously
Conscientiousness 0.098*** 0.073** (0.032) (0.033) Extroversion 0.045 0.036 (0.031) (0.032) Agreeableness -0.057 -0.060 (0.040) (0.041) Neuroticism -0.118*** -0.097*** (0.030) (0.031) Openness to experience -0.079** -0.064* (0.036) (0.037) Control No No No No No No Adj. R-squared 0.57% 0.07% 0.07% 1.00% 0.27% 1.55% OBSERVATION 1434 1434 1434 1434 1434 1434
Note: The standard errors are displayed in parentheses. ***, **, * denotes statistically significant differences at 1%, 5% and 10% level respectively.
found that both the extroversion and agreeableness had a statistically significant negative effect on financial assets holding. They found that extroversion had a relatively large inverse effect on the amount of financial assets held. However, in Table 8, neuroticism has no longer a statistically significant effect no matter the personality trait is entered into the regression separately or simultaneously when control variables are included in the regression models.
Table 8: OLS regression of Big 5 traits on total assets with control variables
Big 5 traits Trait 1 Trait 2 Trait 3 Trait 4 Trait 5 Simultaneously
Conscientiousness 0.068** 0.062* (0.031) (0.032) Extroversion -0.003 0.009 (0.029) (0.030) Agreeableness -0.063 -0.053 (0.040) (0.041) Neuroticism -0.020 -0.011 (0.031) (0.032) Openness to experience -0.089*** -0.078** (0.034) (0.036)
Control Yes Yes Yes Yes Yes Yes
Adj. R-squared 15.97% 15.67% 15.82% 15.70% 16.09% 16.20%
OBSERVATION 1360 1360 1360 1360 1360 1360
Note: The control variables in this table (not reported here for brevity) are as in Table 5. The standard errors are displayed in parentheses. ***, **, * denotes statistically significant differences at 1%, 5% and 10% level respectively.
Compared the results in Table 8 to Table 7, it is apparent that entered the Big Five personality traits separately into the regression model will lead to stronger statistically significant results. Furthermore, the regression model excluding control variables will also lead to stronger statistically significant results compared to the regression model including control variables. However, the adjusted R-squared in Table 7 are extremely low and are considerably smaller than in Table 8, indicating the regression model excluding other control variables explained the variation of the regression poorly on the amount of financial assets held.
4.6. Analysis of risky assets and risk-free assets holding
4.6.1. OLS analysis on risky assets and risk-free assets
In Table 9, the determinants of amount of risky assets held and risk-free assets held are presented. The risky assets are consisted of mutual funds, bonds, and shares, while the risk-free assets are consisted of checking accounts, saving accounts and annuity insurance.
Focusing upon the risky assets holding, in terms of control variables, it can be observed that the age has a positive statistically significant effect on holding risky assets, with the significant level is at 5 percentages. Age-squared has a negatively statistically significant effect on risky assets holding with the increasing of ages, indicating that the effect of age on financial assets holding behavior will be fewer and fewer with the increase in age. For example, for the Dutch household individuals, a one year increase in age will lead to 11.8 percentage points increase in risky assets holding. Meanwhile, the results indicate that the educational level achieved by Dutch household individuals has a positively statistically significant effect on risky assets holding. But with the increasing of educational level, the effect of educational level on risky assets holding is decreased.
Table 9: Determinants of risky holding amount and risk-free holding amount
Dependent variables Risky holding Risk-free holding
Risk-averse -0.064 0.036** (0.041) (0.014) Risk-loving 0.050 0.019 (0.042) (0.016) Concern present -0.019 -0.019** (0.023) (0.009) Concern future -0.031 0.029*** (0.029) (0.011) Male -0.360 0.232** (0.316) (0.117) Married -0.137 -0.141 (0.280) (0.122)
Big three cities 0.497 -0.046
(0.392) (0.160) Age 0.118** 0.005 (0.058) (0.021) Age-squared -0.001 0.000 (0.000) (0.000) Education: middle 0.976*** 0.253* (0.343) (0.131) Education: high 0.965** 0.838*** (0.428) (0.186) Urbanization: average 0.173 -0.307** (0.335) (0.141) Urbanization: high -0.465 -0.119 (0.299) (0.128) Health: fair -0.705 0.540* (0.816) (0.277) Health: good -0.411 0.749*** (0.789) (0.260) Health: excellent -1.018 0.958*** (0.853) (0.295) Income: average 0.718* 0.701*** (0.383) (0.145) Income: high 0.936** 0.883*** (0.411) (0.166)
Control Yes Yes
Adj. R-squared 17.39% 13.70%
OBSERVATIONS 246 1353
Focusing upon the risk-free assets holding, in terms of control variables, the amount of risk-free assets held is positively associated with male, high educational level, healthier body and higher income, and negatively associated with average urbanization level. These findings are in line with the existing study (Brown and Taylor, 2011). Besides, our previous results in Table 6 on the control variables are preserved.
In addition, Table 9 indicates significant effects of two personality traits on risky assets held: conscientiousness and extroversion. The positive effect of conscientiousness on risky assets holding is in line with the results showing in Table 7 and Table 8. Conscientiousness is the personality traits to exhibit consistent findings across the different regression methods with positive effect with limited statistically significance. Moreover, the conscientiousness has same effects on financial assets holding, irrespective of type. However, it is apparent that the magnitude of the effect of conscientiousness on risky assets holding is relatively stronger than in the risk-free assets holding. Comparing the results to Table 6, openness to experience has no longer significant effect on risky assets holding. But extroversion appears to have a negative effect on holding risky assets with 5 percent significance. Besides, agreeableness appears to have a negative effect on holding risk-free assets with 10 percent significance.
more about present are more likely to undertake the risk currently, and people who concerning more about future are less likely to bear the risk.
4.6.2. Probit analysis of risky and risk-free assets
In previous sections, this study explored the effect of different personality traits on the financial assets holding amount. In this paragraph, the effect of different personality traits on the probability of holding risky and risk-free assets is investigated to enrich the research.
The dummy variables are generated as risky assets=1 and risk-free assets=0. Both the OLS regression model and Probit regression model are applied to test the influence of personality on probability of holding risky or risk-free assets. The results are displayed in Table 10 as below:
Table 10: Probit analysis of risky and risk-free assets with control
OLS Probit
Variables Coef. Std. Coef. Std.
Intercept -0.131 0.111 -3.837*** 0.794 Conscientiousness 0.001 0.004 -0.003 0.026 Extroversion -0.002 0.004 -0.014 0.025 Agreeableness 0.002 0.006 0.015 0.033 Neuroticism -0.004 0.004 -0.011 0.025 Openness to experience -0.003 0.005 -0.027 0.028 Risk-averse -0.017*** 0.002 -0.098*** 0.012 Risk-loving 0.012*** 0.002 0.062*** 0.012 Concern present -0.003*** 0.001 -0.0135* 0.007 Concern future 0.006*** 0.001 0.037*** 0.009 Male 0.068*** 0.016 0.411*** 0.098 Married -0.043** 0.017 -0.260*** 0.101
Big three cities -0.038* 0.023 -0.204 0.136
Age 0.005* 0.003 0.058*** 0.020
Age-squared -9.41E-06 0.000 -0.0002* 0.000
Education: middle 0.049*** 0.018 0.285*** 0.111
Education: high 0.127*** 0.026 0.593*** 0.149
Urbanization: high -0.007 0.018 -0.042 0.104 Health: fair 0.033 0.037 0.206 0.253 Health: good 0.049 0.035 0.322 0.241 Health: excellent 0.066 0.040 0.428 0.266 Income: average 0.036* 0.020 0.233* 0.121 Income: high 0.061*** 0.023 0.387*** 0.137
Control Yes Yes
Adj. R-squared 15.93% Chi-squared (23) 336.17 (P=0.000) OBSERVATIONS 1,859 1,859
Note: ***, **, * denotes statistically significant differences at 1%, 5% and 10% level respectively.
From the table above, the likelihood ratio chi-square of 336.17 with a p-value of 0.0000, indicating that the Probit model as a whole is statistically significant. However, comparing the results in the two models above with the previous models, all of the Big Five personality traits fail to have a significant effect on probability of holding risky and risk-free assets, while both the risk and time preferences show strong effect on probability of holding different assets. Since only 246 respondents were holding risky assets given the sample size is 2030 observations in total, this data issue may be the potential reason for the insignificance of the impact of Big Five personality on financial assets holding. Furthermore, having a ton of control variables may affect the relationship between the Big Five personality traits and financial assets holding behavior. Therefore, the Probit and OLS model are applied again as robustness model to check if the Big Five personality traits have a correlation with financial assets holding without controlling for other factors. The results are displayed in Table 11 as below:
Table 11: Probit analysis of risky and risk-free assets without control
OLS Probit
Variables Coef. Std. Coef. Std.
Intercept 0.320*** 0.078 -0.265 0.424
Extroversion -0.001 0.004 -0.003 0.023 Agreeableness -0.003 0.005 -0.013 0.030 Neuroticism -0.016*** 0.004 -0.080*** 0.023 Openness to experience -0.004 0.005 -0.027 0.026 Risk-averse -0.0175*** 0.002 -0.090*** 0.011 Risk-loving 0.014*** 0.002 0.068*** 0.011 Concern present -0.0039*** 0.001 -0.0183*** 0.007 Concern future 0.007*** 0.001 0.037*** 0.008 Control No No Adj. R-squared 10.20% Chi-squared (9) 203.02 (p=0.000) OBSERVATIONS 1,874 1,874
Note: ***, **, * denotes statistically significant differences at 1%, 5% and 10% level respectively.
4.7. Analysis of the type of asset holding
In Table 12, the OLS analysis of the effect of Big Five personality traits on different types of financial assets is presented. It is apparent that the influence of the Big Five personality traits varies across the different types of financial assets. As can be seen from the table, the amount of bonds held by Dutch household individuals, arguably the risky financial assets are not influenced by any of the five personality traits. The potential reason for these results may due to the few observations. Considering the observation number of bonds held is too less to be reliable, the extra-ordinary result of bonds held may be related to the small sample leading a biased result.
Table 12: OLS regression of Big 5 traits on individual financial asset
Checking accounts Saving accounts Annuity insurance Mutual funds Bonds Shares
Conscientiousness 0.020 0.097*** -0.010 0.097 -1.440 0.149 (0.025) (0.033) (0.119) (0.089) (0.271) (0.121) Extroversion 0.034 0.010 0.078 -0.142* 0.411 -0.064 (0.024) (0.032) (0.100) (0.083) (0.241) (0.119) Agreeableness 0.045 -0.046 -0.271* 0.071 -2.867 0.340** (0.033) (0.044) (0.147) (0.105) (0.549) (0.159) Neuroticism 0.020 0.032 -0.031 0.071 -1.188 0.012 (0.026) (0.034) (0.101) (0.089) (0.215) (0.125) Openness to experience -0.067** -0.051 0.153 0.090 -0.038 -0.181 (0.028) (0.039) (0.117) (0.089) (0.229) (0.135)
Control Yes Yes Yes Yes Yes Yes
Adj. R-squared 8.07% 13.85% 25.07% 12.17% 85.71% 21.95%
OBSERVATIONS 1129 1052 177 173 23 99
The amount of checking accounts held is negatively associated with the openness to experience with a significant level at 5 percentages, suggesting that the Dutch household individuals being imaginative, intelligent, curious and flexible will hold less checking accounts. This result may be related to that the checking account is definitely a risk-free asset.
The amount of saving accounts held is positively influenced by the conscientiousness, with a relatively large and highly statistically significance. This result reflecting the importance ofthe characteristics of being persevering, organized, responsible, dependable, thorough and industrious here.
Annuity insurance is considered as a risk-free financial asset in the Netherlands. The amount of annuity insurance held is negatively influenced by the agreeableness, with a significance level at 10 percentages. This result reflecting that the characteristics of being friendly and compassionate will lead to less amount of annuity insurance held by Dutch household individuals.
In terms of mutual funds, only extroversion has a negative effect on mutual funds amount held. The finding may be related to that the diversified funds portfolios have spread the risk. Brown and Taylor (2011) stated that a fund manager invests in a range of companies and then formed a portfolio, thereby spreading the potential risk associated with the different sorts of bonds and shares, with individuals only purchasing the mutual funds and receiving the dividends or interest as determined by the performance of the portfolios. Since extroversion has a significant negative correlation with risk-averse found in Table 5, the more extroversion leads to less risk-averse, therefore it is reasonable to hold less mutual funds.
shares. Hence, our findings associated with the relationship between agreeableness and the holding of shares is not in line with the previous study.
In brief summary, it is clearly observed from the table above that Big Five personality traits are crucial determinants of the different types of financial assets held. And the influences of Big Five personality traits vary across the range of financial assets. Considering the observation number of bonds held is too less to be reliable, the extra-ordinary result of bonds held may be related to the small sample leading a biased result. Therefore, the aggregate risky assets and risk-free assets analysis is necessary in order to figure out the distinct of effect of Big Five personality traits on different types of financial assets reliably.
5. Conclusion
In this paper, the effects of psychological characteristics on financial assets holding behavior are analyzed. This study focused on the influence of Big Five personality traits and economic preferences on the amount of financial assets held. The Big Five personality traits developed by Costa and McCrae (1992) classified the personality traits into five different categories: conscientiousness, extroversion, agreeableness, neuroticism and openness to experience. The risk preference and time preference factors are contained as personality factors in this study as a complementary analysis. The correlations between the outcomes from the risk and time experiments and the personality traits are analyzed to enrich the research and show the evidence of necessary to add economic preferences into the study. The analysis of representative data therefore confirms that the level of association between preference personality measures is rather small. However, the possibility that there is a stronger level of association between them still exists and the significant effects of economic preferences on financial assets should not be omitted.
Turning briefly to the other personality variables after focusing on the effect of the Big Five, it is observed that, the level of financial assets is increasing if Dutch household individuals being risk-loving and concerning about the future, and decreasing if individuals concerning about the present. In terms of other control variables, this study indicates that Male have on average 35.7 percentage more financial assets holding than female. Married respondents have on average 14.6 percentage less financial assets holding than unmarried respondents. Furthermore, respondents with higher educational level and healthier body will hold more financial assets than the others. Interestingly, the urbanization level of the city has negative effect on financial assets, where those Dutch household individuals living in lower level urbanization city have lower level of financial assets. Meanwhile, age has no statistically significant effect on total amount of financial assets holding, but has a statistically significant effect on the amount of risky assets holding. This study also shows that household income from employment has a large and positive effect on financial assets. These findings are in line with the findings in the previous literature, for instance, Gropp et al. (1997), Crook (2001) and Brown and Taylor (2011).
This study indicated that entered the Big Five personality traits separately into the regression model will lead to stronger statistically significant results. Besides, the regression model excluding control variables will also lead to stronger statistically significant results compared to the regression model including control variables. However, the adjusted R-squared of regression model excluding controls are extremely low and are considerably smaller than including controls, due to the poor explaining of regression variations.
negative effect on holding risky assets with limited statistically significance. Besides, agreeableness appears to have a negative effect on holding risk-free assets with 10 percent significance. However, in the Probit model, all of the Big Five personality traits fail to have a significant effect on probability of holding risky and risk-free assets with controlling for other factors, while both the risk and time preferences show strong effect on probability of holding different assets. After excluding the control factors in the Probit model, neuroticism appears to exhibit the negative effect on probability of holding risky assets.
With respect to the type of financial assets held, the analysis suggests that influences of Big Five personality traits vary across the range of financial assets. For instance, conscientiousness is positively associated with the amount of saving account held. Openness to experience is negatively associated with the amount of checking accounts held. Extroversion is negatively associated with the amount of mutual funds held. Agreeableness is negatively associated with the amount of annuity insurance held. Furthermore, the amount of bonds held by Dutch household individuals, arguably the risky financial assets are not influenced by any of the five personality traits due to the fact that few respondents were holding bonds. The amount of holding shares, arguably the riskiest form of financial assets in terms of rate of return, is positively affected by agreeableness with a limited statistically significance.
Reference List
Almlund, M., Duckworth, A. L., Heckman, J., Kautz, T., 2011. Personality Psychology and Economics. IZA DP. 5500.
Becker, A., Deckers, T., Dohmen, T., Falk, A., Kosse, F., 2012. The relationship between economic preferences and psychological personality measures. Annual Review of Economics 4, 453-478.
Borghans, L., Duckworth, A. L., Heckman, J. J., Weel, B., 2008. The
economics and psychology of personality traits. Journal of Human Resources 43(4), 972-1059.
Borghans, L., Golsteyn, B. H., Heckman, J., Meijers, H., 2009. Gender differences in risk aversion and ambiguity aversion, Institute for the study of labor discussion papers, 3985.
Brown, S., Taylor, K., Garino, G., Price, S.W., 2003. Debt and financial expectations: an individual and household level analysis. Discussion Papers in Economics 03/5, Department of Economics, University of Leicester, revised Feb 2004.
Brown, S., Taylor, K., 2008. Household debt and financial assets: evidence from Germany, Great Britain and the USA. Journal of the Royal Statistical Society Series A 171(3), 615-643.
Brown, S., Taylor, K., 2011. Household finances and the ’Big Five’ personality traits. IZA Discussion Paper Series 6191, Institute for the Study of Labor (IZA), Bonn. Brown, S., Taylor, K., 2014. Household finances and the ‘Big Five’ personality traits. Journal of Economic Psychology 45, 197-212.
Bucciol, A., Zarri, L., 2015. Does investors’ personality influence their portfolios? Netspar Discussion Paper 01/2015-006.
Bertaut, C., Starr-McCluer, M., 2000. Household portfolios in the United States, Finance and Economics Discussion Series 2000-26, Board of Governors of the Federal Reserve System (U.S.).
Costa, P.T. Jr., McCrae, R.R., 1992. Revised NEO personality inventory (NEO-PI-R) and NEO Five-Factor inventory (NEO-FFI) manual. Odessa, FL: Psychological Assessment Resources
Crook, J., 2001. The Demand for Household Debt in the USA: Evidence from the 1995 Survey of Consumer Finance. Applied Financial Economics 11, 83-91. Gropp, R., Scholz, J.K., White, M. J., 1997. Personal Bankruptcy and Credit Supply and Demand. Quarterly Journal of Economics 112, 217-51.
Heineck, G., Anger, S., 2010. The returns to cognitive abilities and personality traits in Germany. Labour Economics, Elsevier 17(3), 535-546
Kennickell, A. B., Starr-McCluer, M., 1994. Changes in family finances from 1989 to 1992: Evidence from the Survey of Consumer Finances. Federal Reserve Bulletin 80, 861-882.
Lusardi, A., Browning, M., 1996. Household saving: micor theories and micro facts. Journal of Economic Literautre 34(4), 1797-1855
Lusardi, A., 1998. On the importance of the precautionary saving motive. American Economic Review, American Economic Association 88(2), 449-453.
Nandi, A., Nicoletti, C., 2014. Explaining personality pay gaps in the UK. Taylor & Francis Journals 46(26), 3131-3150.
Xiao, J. J., Olson, G. I., 1993. Mental accounting and saving behavior. Home Economics Research Journal 22, 92-109.
Appendix A
Figure A1: Kernel Density distribution of financial assets
Figure A2: Graph box of financial assets
0 .05 .1 .15 .2 D e n si ty 0 5 10 15 20
logrithm of total financial assets holding amount Kernel density estimate
Normal density
kernel = epanechnikov, bandwidth = 0.4319
Kernel density estimate
Appendix B
B1. Risk Preference Measurement Questionnaire9
The following statements concern saving and taking risks. Please indicate for each statement to what extent you agree or disagree.
To what extent do you agree with the following statements?
Please indicate on a scale from 1 to 7 to what extent you agree with the statement.1 means ‘totally disagree’ 7 means ‘totally agree’.
totally totally disagree agree
1 2 3 4 5 6 7
- I think it is more important to have safe investments and guaranteed returns, than to take a risk to have a chance to get the highest possible returns.
- I do not invest in shares, because I find this too risky.
- If I think an investment will be profitable, I am prepared to borrow money to make this investment.
- I want to be certain that my investments are safe.
- If I want to improve my financial position, I should take financial risks.
- I am prepared to take the risk to lose money, when there is also a chance to gain money.
B2. Time Preference Measurement Questionnaire10
Now we present you some statements about the future. Please indicate for each statement to what extent you agree or disagree.
To what extent do you agree with the following statements?
Please indicate on a scale from 1 to 7 to what extent you agree with the following statements.
1
means‘
extremely uncharacteristic’ 7
means‘
extremelycharacteristic
’
1 2 3 4 5 6 7
- I think about how things can change in the future, and try to influence those things in my everyday life.
- I often work on things that will only pay off in a couple of years.
- I am only concerned about the present, because I trust that things will work themselves out in the future.
- With everything I do, I am only concerned about the immediate consequences (say a period of a couple of days or weeks).
- Whether something is convenient for me or not, to a large extent determines the decisions that I take or the actions that I undertake.
- I am willing to sacrifice my well being in the present to achieve certain goals in the future.
- I think it is important to take warnings about negative consequences of my acts seriously, even if these negative consequences would only occur in the distant future. - I think it is more important to work on things that have important consequences in the future, than to work on things that have immediate but less important
consequences.
- In general, I ignore warnings about future problems because I think these problems will be solved before they get critical.
- I think there is no need to sacrifice things now for problems that lie in the future, because it will always be possible to solve these future problems later.
- I only respond to urgent problems, trusting that problems that come up later can be solved in a later stage.
B3. Big Five Personalities Measurement Questionnaire11
For the following statements on human behavior, please choose the statement that applies most to you. Describe yourself as you are, not as how you want to be. Describe yourself in comparison to other people you know of the same sex and of about the same age.
1 means ‘not at all applicable’ 5 means ‘highly applicable’ - I have excellent ideas.
- I like order.
- I keep in the background.
- I sympathize with others’ feelings. - I have frequent mood swings. - I have a vivid imagination. - I’ll leave my things lying around. - I am quiet around strangers. - I take time out for others. - I get stressed out easily.
