What is the influence of personality traits on Dutch
household’s attitude towards financial decisions?
Gerhard Wemke van der Velde
University of Groningen
Faculty of Economics and Business
MSc Finance
Jan. 2017
Table of Content
1 Introduction ... 4
1 Literature Review ... 6
1.1 The personality traits ... 6
1.1.1 Neuroticism ... 7 1.1.2 Extraversion... 7 1.1.3 Openness to Experience ... 8 1.1.4 Agreeableness ... 8 1.1.5 Conscientiousness ... 8 1.1.6 Overconfidence ... 8
1.2 Personality traits and Financial Decisions ... 9
1.2.1 Attitude towards saving money ... 9
1.2.2 Risk Aversion ... 10
1.3 Personality traits on financial decisions ... 10
1.3.1 Extraversion on financial decisions ... 10
1.3.2 Agreeableness on financial decisions ... 11
1.3.3 Conscientiousness on financial decisions ... 11
1.3.4 Neuroticism on financial decisions ... 12
1.3.5 Openness to Experience on financial decisions ... 12
1.3.6 Overconfidence on financial decisions. ... 13
2 Data ... 14 2.1 Data information... 14 2.2 Preparation of Data:... 14 2.3 Descriptive Statistics ... 15 2.4 Control Variables ... 15 2.5 Operationalization ... 16 2.6 Cronbach’s Alpha’s: ... 18 2.7 Plan of analysis ... 19 3 Results ... 20 3.1 Descriptive statistics: ... 20
3.1.1 Descriptive statistics Attitude towards Saving ... 20
3.1.2 Descriptive statistics Risk Aversion ... 21
3.2 Bivariate Regression... 22
3.2.2 Dependent Variable Risk Aversion: ... 23
3.3. Multicollinearity: ... 24
3.4 Assumptions: ... 24
3.5 Multiple Regression ... 24
3.5.1 Multiple regression for the Attitude towards saving money ... 25
3.5.2 Multiple regression on the Level of Risk Aversion ... 27
4 Conclusion ... 30
4.1 Personality traits and Attitude towards saving money ... 30
4.2 Personality traits and Risk Aversion ... 32
5 Limitations and Further Research ... 35
6 References ... 37
7 Appendix: ... 41
7.1 Appendix Descriptive statistics ... 41
7.2 Appendix operationalization ... 50
7.3 Appendix Multiple Regression on Attitude towards saving ... 52
What is the influence of personality traits on Dutch
household’s attitude towards financial decisions?
Gerhard Wemke van der Velde
Abstract:
This paper investigates the influence of personality traits on the attitude towards saving of Dutch households. The Big Five personality traits plus Overconfidence are used to explain the attitude towards saving and risk aversion of Dutch households. The DNB Household Survey 2015 by CentER Data is used. Doing so, Conscientiousness, Agreeableness, Neuroticism and Overconfidence were found to be significant predictors of the attitude towards saving. Openness to Experience, Agreeableness and Overconfidence are found to be significant predictors of Risk Aversion.
1 Introduction
Financial theories rely on perfect markets and perfectly rational people. During time, people found that markets are not always perfect. This paper will go one step further and follows the idea that people are imperfect, people show irrational behavior. These deviations from rationality can be systematic, which is the basis of the field known as behavioral finance. People are not rational. A great example in finance is the sunlight affecting mood (Hirshleifer & Shumway, 2003). Hirshleifer and Shumway (2003) found that sunshine is highly
significantly correlated with daily stock returns. This cannot be explained with rational price settings and is therefore seen as psychological evidence that sunny weather is associated with a better mood which results in higher stock returns. This is just one example which resulted in the field known as behavioral finance. Some relevant concepts used in this field are framing, heuristics, limited self-control, social influences and emotions. These are all cognitive skills. Cognitive heuristics are, for example, representativeness, availability and anchoring.
This paper will focus on non-cognitive skills, or soft skills, which are related to motivation, integrity and interpersonal interaction. The soft skills are associated with an individual’s personality (ACT WorkKeys, 2014) To measure an individual’s personality, the Big 5 personality traits are often used (Brucciol & Zarri, 2015). Individuals scores on these 5 traits reflect the way in which the individual thinks and feels (Rustichini, A., DeYoung, C.G., Anderson, J. Burks, S.V., 2012). The big 5 personality traits are Extraversion, agreeableness, conscientiousness, neuroticism and openness to change(Borghans, L., Duckworth, A.L., Heckman, J., Weel, B.J., 2008). These five factors capture the term personality at the broadest level of abstraction, with each factor being a summary of more specific and distinct
personality characteristics (Borghans et al., 2008).
this paper is: What is the influence of personality traits on Dutch households attitudes towards financial decisions?
The study contributes to the literature in mainly three ways. First, this paper will focus on the Dutch Households since little research had been done using the DNB Household Survey of CentER data in order to explore the influence of personality traits on the attitude towards financial decisions. The 23rd wave of the DNB Household Survey is used, which collected data from April 2015 till October 2015 and is currently the most recent DNB Household Survey available. Recent studies in the field of behavioral finance focused on British
Households (Brown & Taylor, 2014) German Households (Brown et al., 2008) or the Health and Retirement Study from the USA (Brucciol & Zarri, 2015). By focusing on the Dutch Households, this paper will contribute to the existing literature.
Second, this research tries to capture the attitudes towards risk and saving where other studies (Nyhus & Webley, 2001; Brown & Taylor, 2014) focused on the amount of net debt and level of risk aversion. When differences occur between the attitudes towards saving and the actual amount of money saved, interesting conclusions can be made.
Third, overconfidence is added to the standard Big Five personality traits in order to explain both attitudes towards saving and the level of risk aversion. This paper will contribute on the existing literature on overconfidence by using Dutch Household data on the attitudes towards saving and risk aversion. Existing literature focuses on non-Dutch data in order to explain the influence of overconfidence on risk aversion. Overconfidence on the attitude towards saving has not been explored extensively. This paper tries to extend the findings of existing literature concerning overconfidence.
The conclusion of this paper is summarized as follow. Using a multiple regression in order to predict the dependent variables, attitude towards saving money and risk aversion,
conscientiousness, agreeableness, neuroticism and overconfidence were found to be
significant predictors of the attitude towards saving. Openness to experience, agreeableness and overconfidence are found to be significant predictors of Risk Aversion.
This paper is structured in the following way. Firstly, the literature will be explored on
1 Literature Review
This research tries to explain the attitude towards financial decisions through personality traits. Existing literature is explored in order to find useful explanatory variables for the attitude towards financial decisions. First, the variables influencing financial decisions will be explored using existing literature. Second, the financial decisions Dutch households face are explored. Third, literature is used in order to support the variables used that influence the attitude towards financial decisions.
The research conducted by Brown & Taylor (2014) indicated that “certain personality traits are associated with a number of aspects of individuals’ economic and financial decision-making.” These personality traits were extraversion, conscientiousness and openness to experience. The conclusion was based on research using the British Household Panel Survey with 10.000 individuals interviewed annually during three waves from 1995, 2000 and 2005. Using the BHPS, Brown & Taylor (2014) analyzed the relationship between personality traits and financial decision making. Brown & Taylor (2014) focused on unsecured debt and
financial assets. Brown & Taylor (2014) found that extraversion is positively associated with the probability of holding credit card debt, and conscientiousness is negatively associated with holding credit card debt. They also found that individuals scoring high on openness to experience are positively related to holding shares.
Nyhus & Webley (2001) concluded, using the CentER Saving Survey from 1996 and 1997 (N=734 households), that extraversion is a predictor of the household behavior towards savings. The focus of the research by Nyhus & Webley (2001) was to investigate the relation between personality and saving/borrowing behavior. They measured saving behavior by looking at household liquid saving, investment saving, insurance saving, the sum of debt, sum of saving and plans of saving in the next 12 months. These are all fixed numbers on saving or debt. Nyhus & Webley (2001) also found that emotional stability is a predictor of saving and borrowing behavior, and agreeableness has a relation with certain types of saving.
Extraversion, Agreeableness, Neuroticism are all personality traits that are part of the Big Five personality traits (Borghans, L., Duckworth, A.L., Heckman, J., Weel, B.J., 2008). The personality traits will first be explored before exploring the current literature on the
relationship between personality traits and attitude towards financial decision making.
1.1 The personality traits
(Almlund, M., Duchworth, A.L., Heckman, J.J., Kautz, T.D., 2011). The Big Five personality traits capture the term personality at the broadest level of abstraction, with each factor being a summary of more specific and distinct personality characteristics (Borghans et al., 2008). In the world of psychology peoples personality are often described using the ‘Big 5’ also called the Five Factor Model or NEO-PI (McCrae, 1989). McCrae & Costa (1989), first adopted the NEO model, which includes neuroticism, extraversion and openness to experience. The model was seen as incomplete, the traits agreeableness and conscientiousness were therefore added. The Big 5 personality traits are neuroticism, extraversion, openness to experience,
agreeableness and conscientiousness (Borghans et al., 2008). Combining the scores on all 5 factors gives insights on the way individuals think and feels (Rustichini et al., 2012). It is argued that these personality traits are stable over time, they do not change during the lifetime of people (Ardelt, 2000; Borghans et al., 2008). An interesting trait is developed and more explored by Barber & Odean (2000) and is called overconfidence. Overconfident individuals are too confident by misjudging their own abilities and skills (Montier, 2007). It is confirmed that overconfidence has an impact on financial decision making (Peón, D., Antelo, M., Calvo, A., 2016). Since little research has been done on the influence of overconfidence on the attitude towards financial decisions, overconfidence is also included as a variable explaining the attitude towards financial decisions. The 6 factors are described in more detail below. 1.1.1 Neuroticism
Neuroticism is about the tendency of individual to view things with negative emotions. Individuals scoring high on neuroticism overreact to bad feelings and therefore experience psychological effects as anger and depression (Zhang, J., Wang, H., Wang, L., Liu, S., 2014). Neuroticism can be seen as a summary of more specific and distinct personality traits
(Borghans et al., 2008). Examples of such traits that are included within neuroticism are, nervousness, moodiness and tempera mentality (Goldberg, 1993).
1.1.2 Extraversion
are examples of specific traits of individuals scoring low on extraversion and therefore being more introverted (Goldberg, 1993).
1.1.3 Openness to Experience
Openness is about the level of conservatism or open mindedness (Almlund et al., 2011). Individuals scoring high on openness are independent and think critically when solving problems. These individuals are able to get into the core of a problem (Zhang et al., 2014). High scoring openness individuals are knowledgeable and high intellectuals. More specific traits that are captured within openness to experience are, imagination, curiosity and creativity (Goldberg, 1993). Contrasting traits are, for example, shallowness and imperceptiveness. 1.1.4 Agreeableness
“Agreeableness refers to individuals courteousness, trustworthiness, tolerance, compassion, generosity and cooperative demeanor” (Zhang et al., 2014). Agreeableness measures the trust, harmony, compassion and cooperation of an individual. Individuals scoring high on
agreeableness are often seen as naïve and can be easily be manipulated by others, this due to the above mentioned factors. Agreeableness can be seen as broader trait capturing more specific traits as, kindness, trust, warmth, selfishness and distrust (Goldberg, 1993). 1.1.5 Conscientiousness
Conscientiousness is about an individual acting in an organized, disciplined and effective manner (Goldberg, 1993). Individuals scoring high on conscientiousness are realistic and work in an organized way (Zhang et al., 2014). It is seen as one of the best predictors of work performance and high scoring individuals do not take decisions without adequate information and experience. Therefore, conscientiousness can be seen as a summary trait of more specific traits such as, thoroughness, reliability, carelessness and unreliability (Goldberg, 1993).
1.1.6 Overconfidence
“Overconfidence occurs when someone has more confidence than they should have based on the situation and they misjudge their ability or opinion” (Montier, 2007). It can be divided into three groups, miscalibration, Above average effect and excessive optimism (Glaser and Weber, 2007). Miscalibration is about individuals answering a question, and afterword
Indicating that 50% of the managers found themselves falsely above average, which is caused by the above average Effect.
Overconfidence is driven by several factors. These factors are, memory issues, biased self-attribution, illusion of knowledge, illusion of control and confirmation bias (Kramer, 2016). Memory Issues can best be explained by the Hindsight Bias, which is also known as the ‘knew-it-all-along’ bias (Kramer, 2016). Meaning that individuals believe they understand the past, which means they can also predict what will happen in the future (Kahneman, 2011). Self-attribution is the phenomenon that individuals address good outcomes to their own ability and skills. The opposite is true for bad outcomes, which is attributed to bad luck (Daniel & Hirshleifer, 2015). Confirmation bias occurs when individuals are looking for information which confirms prior ideas rather than looking for new contradicting information (Kramer, 2016).
Following the results of Cobb-Clark & Schurer (2012), concluded after using the HILDA survey from 2003-2007 on Australian Households (N= 7682 households), the Big Five personality traits, the non-cognitive skills, can be seen as stable inputs into a wide range of financial decisions. The financial decisions that are influenced by personality traits are discussed in the next section.
1.2 Personality traits and Financial Decisions
This research tries to capture the influence of personality traits on the attitude towards household financial decisions. First, we need to know what these financial decisions are. “Personal finance defines all financial decisions and activities of an individual or household, including budgeting, insurance, mortgage planning, savings and retirement
planning.”(Investopedia, 2017). In this paper we focus on the attitude towards saving money and risk aversion. The influence of personality traits on both dependent variables is measured separately and therefore discussed individually throughout this paper. First, the two variables will be more explored individually.
1.2.1 Attitude towards saving money
have sufficient income to save money while this is not important for them. This research will look at the personality traits that influence the attitude towards saving money. Having a negative (positive) attitude towards saving money which ultimately result in having less (more) savings.
1.2.2 Risk Aversion
The second dependent variable that this research will focus on is risk aversion. Risk aversion is the behavior of humans, in this case households, to reduce uncertainty when they are exposed to uncertainty (Thomas, 2016). Risky assets are stocks, options and mutual funds (Donkers & Soest, 1999). For example, stocks are more riskier than bonds and bonds are more riskier than having a savings account. In return for higher risks individuals, or
households, gain higher returns on their investments. Risk-free examples are Bonds and safe mutual funds that only invest in bonds.
These two attitudes towards financial decisions are used in this paper. Earlier mentioned personality traits, the big five and overconfidence, are used to explain the attitude towards saving and risk aversion. Literature is used in order to provide hypotheses which combined will answer the research question: What is the influence of personality traits on Dutch Household’s attitude towards financial decisions? The literature on the relation between personality traits and the attitude towards financial decisions are described in the next part.
1.3 Personality traits on financial decisions
Following the results of Nyhus & Webley (2011), Brown & Taylor (2014) and Bucciol & Zarri (2015) this research will empirically research the relationship between the personality traits and the attitudes towards financial investment decisions. The personality traits are all described and hypotheses are stated. The hypotheses indicate which relation is expected between personality traits and the attitude towards financial decisions based on existing literature.
1.3.1 Extraversion on financial decisions
It is argued that “extroverted investors arelikely to have good relationships with other people in the market; thus, they can access more information for investment decisions. (Zhang et al., 2014).” Nyhus & Webley (2001) found that introverted individuals save more and borrow less. They used data from the Center Saving Survey from 1996 and 1997. The conclusion of their research results in the following hypotheses. The hypotheses for this research is:
Anic (2007) found that extraversion is significantly correlated with risk measures. This can also be explained by looking at the traits that are part of extraversion. One of the traits that can be used in order to describe extraverted people is sensation seeking (McCrae & Costa, 1995). Bibby & Ferguson (2011) concluded that sensation seeking is associated with higher risk tolerance. For this research the following hypothesis is formed: The hypothesis is therefore:
Hypothesis 2: Extraversion has a negative effect on risk aversion.
1.3.2 Agreeableness on financial decisions
Nyhus & Webley (2001) concluded that individuals scoring high on agreeableness borrow more and save less. Which is supported by Kausel, Hansen, Tapia (2015) which suggested that agreeableness is negatively effecting the pension savings of individuals. Kausel et al. (2015) based their results on a nationally representative survey from Chile. The hypothesis for this research are therefore:
Hypothesis 3: Agreeableness has a positive effect on the attitude towards saving.
Bucciol & Zarri (2015) found, using data from the US Health and Retirement Study from 2006-2012, that investors scoring low on agreeableness are more willing to take up risk. This is easily explained by the fact that investors scoring low on Agreeableness are more self-centered and uncooperative (Bucciol & Zarri, 2015). Bucciol & Zarri (2015), focused on individuals with an age between 50 and 80 and financial wealth being higher than 1000 US Dollars. They concluded that investors scoring high on agreeableness are considered to be more risk averse. Borhans et al. (2009) also concluded that agreeableness is positively
associated with risk aversion. This is supported by Anic (2007) who indicated that low scores on agreeableness were associated with higher scores on risk-taking. The following hypothesis is formed.
Hypothesis 4: Agreeableness has a positive effect on risk aversion.
1.3.3 Conscientiousness on financial decisions
As mentioned before, conscientiousness individuals are organized, disciplined and acting in an effective manner. This results in individuals not making impulsive investing decisions, which results in individuals scoring high on conscientiousness are almost never in debt (Nyhus & Webley, 2001). This is supported by Altunbüken (2016) which found that
by the UK government, while also holding household savings in this account. The hypothesis for this research is therefore:
Hypothesis 5: Conscientiousness has a positive effect on the attitude towards saving.
Almlund et al. (2011) indicated that there is a positive correlation between risk aversion and conscientiousness. This result is supported by the fact that individuals scoring high on Conscientiousness do not make impulsive investing decisions (Nyhus & Webley, 2001). Conscientiousness people are considered to be thorough and reliable therefore, the following hypothesis is also formed:
Hypothesis 6: Conscientiousness has a positive effect on risk aversion. 1.3.4 Neuroticism on financial decisions
Emotional stable individuals save more and borrow less (Nyhus & Webley, 2001). This is supported by Harrison & Chudry (2011) who found that emotional unstable individuals find themselves bad money managers, and have anxiety towards financial related decisions. The hypothesis therefore is:
Hypothesis 7: Neuroticism has a negative effect on the attitude towards saving.
Using data from Dutch High School students (N=347), Borghans et al. (2009), concluded that risk aversion is positively associated with neuroticism. This can be explained by the fact that individuals scoring high on neuroticism experience anxiety, depression and overreact to bad feelings. Such individuals are not likely to take up risk (Borghans et al. 2009). This is supported by Anic (2007) who concluded that lower scores in neuroticism were associated with higher scores on risk-taking. The hypothesis for this research is therefore:
Hypothesis 8: Neuroticism has a positive effect on risk aversion.
1.3.5 Openness to Experience on financial decisions
Nyhus & Webley (2001) found high level openness individuals are more independent and critical thinkers and therefore have less debt and more savings. Kausel et al. (2015) support this finding by suggesting that individuals scoring high on openness to experience are more likely to save, where individuals scoring low on openness to experience are saving less. This is explained by the fact that individuals scoring high on openness to experience think
critically before making decisions, which would result in more savings (Nyhus & Webley, 2001). The hypothesis is:
Hypothesis 9: Openness to Experience has a positive effect on the attitude towards saving.
more openminded. Beyer et al. (2015) state that individuals scoring high on openness to experience tend to increase risk taking and are therefore less risk averse. This can be
explained by the fact that more specific traits that are captured within openness to experience are curiosity and open-mindedness (Goldberg, 1993).
Hypothesis 10: Openness to Experience has a negative effect on risk aversion.
1.3.6 Overconfidence on financial decisions.
Overconfidence causes investors to be too certain about their own abilities and not to weight the opinion of others sufficiently. Furthermore, overconfident investors underreact to new information, or overweight the value of information, but they also hold unrealistic beliefs about how high their returns will be (Barber & Odean, 2000). Overtrading, and therefore taking more risk, is easily explained by Overconfidence. Individuals believe they have more knowledge then other individuals while all individuals have excess to the same information (Daniel & Hirschleifer, 2015). The illusion of knowledge is present and therefore individuals start trading more. Peón et al. (2015) also found a relation between overconfidence and risk seeking. The hypothesis therefore is:
Hypothesis 11: Overconfidence has a negative effect on risk aversion.
Gross (2015) states that most Americans feel confident about their finances while most of them have no reason of seeing a bright financial future. Gross (2015) based his findings on a survey on the National Foundation for credit Counselling. Being overconfident results in seeing a brighter financial future which results in overconfident individuals not saving money. The hypothesis therefore is:
Hypothesis 12: Overconfidence has a negative effect on the attitude towards saving.
2 Data
In this section, the method is described. First the data used will be described by giving general information on the data. In the operationalization section, the variables formed will be
described and how they are constructed. Reliability analysis are performed and the
Cronbach’s Alpha is given for each construct. In order to answer the hypotheses a multiple regression is used, this is discussed in the Plan of Analysis. Again, it is worth mentioning that this research focusses on two dependent variables separately. This paper tries to explain the dependent variables, both attitude towards saving and risk aversion, separately. This will be explained into more detail after the data is described.
2.1 Data information
The dataset used is generated by CentERdata and is called the DNB Household Survey 2015. Since 1993, the CentERdata collects data through a panel that consists of Dutch households. Data is collected from more than two thousand different households. The DNB household Survey is studying the economic and psychological determinants of the saving behaviour of households. Households are provided with a computer and access to the internet in order to fill in the questionnaires. It reflects the composition of the Dutch-speaking population in the Netherlands. The data used for this research is acquired from the 23rd wave. The data was collected between April 2015 and October 2015. Individuals aged 16 or over were all interviewed and used for this research. When individuals had no access to a computer, the computer was provided by CentERdata.
The dataset is divided into several questionnaires. These questionnaires are focusing on different topics. These topics are, General information on the household, household and work, accommodation and mortgages, health and income, assets and liabilities, economic and psychological concepts. For this research, general information on the household and economic and psychological are used in order to create different variables.
2.2 Preparation of Data:
these financial decisions are not relevant. Therefore, not all the data of each individual is selected. With the use of selected cases on the variable ACCOUNT, which asks if the individual is responsible for the financial decisions, the relevant individuals are selected. Individuals can answer this question with the values 0 (=no) and 1 (=yes). By selecting cases with a value equal to 1 on the variable account, relevant individuals are selected. All persons aged 16 or over were interviewed, while the data showed Year of Births of 2015. In order to overcome this problem cases were selected in which the Year of Birth was equal or smaller than 1999.
2.3 Descriptive Statistics
A total of 5133 individuals filled in the General information on the household, these individuals are part of 2128 households. 496 of these individuals represent an one person household, 1844 individuals live in a household of two persons and the rest has 3 or more members with a maximum of 8 people within a household. The oldest person was born in 1919 and the youngest in 2015. The average person was born in 1971. In total, 2523 of the individuals are men and 2610 individuals are women.
This research tries to explain the relationship between personality traits and financial
decisions therefore, not all data form the DHS will be used. In total 1138 individuals, with an age of 16 or older, are financial decision takers of their household and filled in the economic and psychological questions and can therefore be used for this research. This questionnaire on economic and psychological concepts has been split into two parts since 2000 and merged into one part since 2009. Questions for the attitude towards saving are only filled in by individuals in position 1, 2 or 3 on question. These individuals are, 1) head of the household, 2) spouse, 3) permanent partner. In total, 1530 are head of the household, 512 are spouse and 68 are permanent partner which in total gives 2110 individuals. The construct to measure the attitude towards money uses 16 variables which are not all filled in. Therefore, the construct has only 466 valid responses. Overconfidence and risk aversion are filled in by 1073 and 1054 individuals. The difference between overconfidence, risk aversion and the Big Five
personality traits can easily be explained due to the fact that they are part of different
questionnaires and therefore not filled in by the same amount of individuals. More descriptive statistics on the data for both dependent variables are given in the Result section.
2.4 Control Variables
for the results found and to increase the significance (Becker, 2008). The decision to save is influenced by demographic factors like age, gender and education (Popovici, 2012). Also the number of children present in the household as an impact on the savings (Popovici, 2012). 2.5 Operationalization
The variables used for this research are all constructed using a multitude of questions from the questionnaires. In total, six independent variables are used in order to predict the individual dependent variables. A total of five control variables is added. These variables are discussed in table 1, which shows how the variables are measured. The personality traits and dependent variables are constructed using multiple variables, all questions are shown in the Appendix. In order to construct the variables used, the Cronbach’s Alpha should be sufficient, this is
discussed after table 1 is presented.
Table 1: Operationalization of variables used
Variable Operationalization
Independent Variables
Extraversion* Measured using 7 questions, answered on a 5 item Likert scale. Example, I am the life of the party, I start conversations and I do not talk a lot (reversed).
Conscientiousness* Measured using 7 questions, answered on a 5 item Likert scale. Example, I like order, I do chores right away and I often make a mess of things (reversed).
Openness to Experience* Measured using 8 questions, answered on a 5 item Likert scale. Example, I have excellent ideas, I am full of ideas, I do not have a good imagination (reversed).
(reversed).
Neuroticism* Measured using 9 questions, answered on a 5 item Likert scale. Example, I get stressed out easily, I often feel blue, I am relaxed most of the times (reversed).
Overconfidence* Measured using 3 questions, answered on a 7 item Likert scale. Example, compared to my friends I am better off.
Dependent Variables
Attitude towards saving* Measured using 16 questions, answered on a 7 item Likert scale. Questions are all part of the question: How important is it to you to save money?
Risk aversion* Measured using 2 questions, answered on a 7 item Likert scale. Example, If I want to improve my financial position, I should take financial risk (reversed)
Control Variables
Gender 1=male and 2=female
Year of Birth Any answer possible.
Education Level Measured on a 9 item scale. 1=special education and 7= University
Region 1= three largest cities, 2=west, 3=north,
Number of Children 9 point scale with 0= 0 children and 9= 9 children within the household.
*average scores were calculated (total score/number of questions)
Personality traits and both dependent variables are constructed using multiple questions. Some questions can be used in order to measure the personality traits or dependent variables, but are measured using a reversed scale. For example “I insult people”, which can be used to measure the level of agreeableness, but in a reversed way. Such questions, which is indicated per question, are reversed and new variables are created. These reversed variables are used when constructing the personality traits or dependent variables. Constructing variables using a multitude of other variables can only be done when these variables measure the same thing. This reliability is tested by measuring the Cronbach’s Alpha. This is discussed in the next section.
2.6 Cronbach’s Alpha’s:
All variables used for the regression are constructed by using multiple variables. In order to come up with reliable constructs the Cronbach’s Alpha’s are measured. The Cronbach’s Alpha is used to measure the internal consistency (Tavakol & Dennick, 2011). It is a way to see of a multitude of items can form one variable together. The Cronbach’s Alpha gives an indication of the reliability of the formed variable using different items.
The independent variables are the Big Five personality traits and the added personality trait overconfidence. For each personality the variables used are stated in the appendix.
Cronbach’s Alphas are used to measure the reliability of the constructed variables and are listed in table 2. The Cronbach Alpha should be more than .7 in order for the construct to be reliable (Tavakol & Dennick, 2011). As can be seen in table 2, all of the constructs are therefore reliable.
Table 2: Cronbach’s Alpha
Variable Cronbach’s Alpha # of items
Extraversion .813 7
Conscientiousness .724 7
Openness to experience .728 8
Agreeableness .825 9
2.7 Plan of analysis
In order to test in which way both dependent variables are predicted using the set of
independent variables, a linear multiple regression is performed. The multiple regression is performed in order to look at the influence of the 6 independent variables on both the dependent variables. A multiple regression is used when the relationship between several independent variables and one dependent variable is being explained.
The multiple regression is done for both dependent variables separately, since they are not related to each other. Doing this will give insights on the way both dependent variables are explained by the set of personality traits separately. The independent variables are
extraversion, conscientiousness, openness to experience, agreeableness, neuroticism and overconfidence. The dependent variables are the level of attitude towards saving money and the level of risk aversion. Control variables are added for age, gender, education, number of children and region. The regression will be performed in order to see if the independent variables can significantly predict the dependent variables.
The Big Five personality traits are added in model 1, overconfidence is added in model 2 and finally Control variables are added in model 3. The control variables and Big Five personality traits are added in the same model since there is no order in which they need to be added. The following estimating formula is formed for personality traits on attitude towards saving:
𝑦 = β0 + β1(𝑥1) + β2(𝑥2) + β3(𝑥3) + 𝑒
Where 𝑦 is the dependent variable, the attitude towards saving or the level of risk aversion. β0
is the constant, 𝑥1 is a vector including the Big Five Personality traits, 𝑥2 is overconfidence and 𝑥3 is a vector of control variables and 𝑒 is the error term. The results of the multiple regression is given in the next section.
Overconfidence .837 3
Attitude towards saving money .860 16
3 Results
After exploring the literature and introducing the data used, the results will be given in this section. First, the variables used are discussed in the univariate section, all variables are explained by looking at the mean, std. deviation and number of participants. Second, the bivariate section in which correlations between all variables are explored and discussed. Third, the results from the multiple regression is given and discussed. This is done for both the models separately, since this research tries to explain two different dependent variables from the same set of independent variables.
3.1 Descriptive statistics:
In this section, the statistics of the variables used will be stated. Where earlier descriptives, part 2.3, focussed on the descriptive of the whole dataset, this descriptive will be based on the data that is actually used in order to explain the effect of personality traits on both dependent variables. Since this research focusses on 2 dependent variables, the descriptive statistics will be reported individually. The personality variables and dependent variables are all continuous variables, of which the mean and standard deviation are interesting to look at. The control variables are all categorical variables, of which statistics are given below table 3 and 4. 3.1.1 Descriptive statistics Attitude towards Saving
First, the descriptive statistics for the regression on the attitude towards saving are given in table 3.
Table 3: Descriptive Statistics
Mean Std. Deviation N
Attitude Towards Saving Money 4.49 .93 458
Extraversion 3.14 .66 458 Conscientiousness 3.72 .53 458 Openness To Experience 3.40 .56 458 Agreeableness 3.87 .54 458 Neuroticism 2.45 .68 458 Overconfidence 3.85 .99 458
high standard deviation means that scores within this variable deviate a lot. Overconfidence is measured on a 7 point scale, so therefore scores are automatically higher. Taking this into account doesn’t gives an average score compared to the other personality traits. High scores are also present for the attitude towards saving, 4.49 with a standard deviation of .93, meaning the average person find saving important.
The control variables are all measured on a categorical scale and results are shown in the appendix. The descriptive statistics show that 254 individuals are male, 55.5% vs. 204 females, 44.5%. Most of these individuals live in the three largest cities or in the west, a total of 201 or 43.9%. Only 50 individuals are situated in the north, which is 10.9%. East and South are presented almost equally with 106 vs 101 individuals, or 23.1% vs 22.1%.
Individuals finished vocational colleges the most, 116 or 25.3% and only 14 individuals only finished kindergarten/primary education. Most households had no children present within the household, 64.2%, and households with only one or two children present accounted for 28.6% of the 458 households used. The median for the year of birth was 1957, which makes
individuals 58 years old at the time of the interview. The oldest individual was born in 1925 and the youngest in 1988.
3.1.2 Descriptive statistics Risk Aversion
Second, descriptive statistics are given for the regression on risk aversion. Results are shown in table 4.
Table 4: Descriptive Statistics
Mean Std. Deviation N Risk Aversion 5.1333 1.48855 975 Extraversion 3.1213 .68213 975 Conscientiousness 3.7468 .54568 975 Openness To Experience 3.3913 .56816 975 Agreeableness 3.9300 .54416 975 Neuroticism 2.3787 .70569 975 Overconfidence 3.8147 1.07365 975
A total of 975 respondents can be used from the dataset. These 975 individuals show high means for conscientiousness and agreeableness, 3.78 and 3.93. Meaning that individuals used in order to explain risk aversion saving are organized, disciplined, naïve and easily
aversion has a high standard deviation, 1.48, meaning that scores show a high variance. The control variables are all measured on a categorical scale and results are shown in the
appendix. More males are present in this dataset, 557 males are used in total vs 418 females, but this is not a large variance. Most individuals are situated in the three big cities or in the west, in total 416 individuals are living in these regions. The least individuals are living in the north, only 113, and the east and south is almost the same with 208 living in the east and 238 living in the south. Most individuals finished pre-vocational education, 221, senior vocational training, 216, or vocational college, 248. In total, 27 individuals only finished
kindergarten/primary education. A total of 663, or 68%, of the households had no children present in the household. Households with one or two children made up for 254, or 26.1%, of the, in total, 975 individuals. The mean for year of birth was 1956, indicating that individuals were on average 59 years old when interviewed. Four individuals were born in 1925 and the youngest was born in 1991.
These are the descriptive statistics on the datasets used for both dependent variables separately. In order to see how the variables interact with each other, the correlations are calculated. The correlations are stated and discussed in the next section.
3.2 Bivariate Regression
The correlations are given in table 6 and table 7. This is done to see how the variables interact with each other and if this relation is significant. The variables used are extraversion,
conscientiousness, openness to experience, agreeableness, neuroticism and overconfidence, The attitude towards saving money and the level of risk aversion. Table 6 shows the
Table 6: Correlation IV on Reasons on Saving Money
Extraversion Conscientiousness Openness to Exp.
Agreeableness Neuroticism Overconfidence Att. Tow. saving Gender Year of Birth Level of Education Region #of Children Extraversion 1 Conscientiousness .144** 1 Open. to Exp. .317** .218** 1 Agreeableness .328** .364** .291** 1 Neuroticism -.204** -.254** -.125** -.190** 1 Overconfidence .127** .052 .112** -.024 -.077* 1
Att. Tow. Saving .061 .118* .076 .198** .143** .186** 1
Gender .004 .008 -.034 -.040 -.003 -.015 -.020 1 Year of Birth .025 .026 .045 .051 .029 -.040 .043 .153** 1 Level Education .033 -.035 .043 .011 .004 -.011 .044 -.091** .276** 1 Region .042 -.041 .034 .009 .053 .050 .117* .029 -.004 -.022 1 # of Children .018 .030 .032 .044 .003 -.027 .096* .056* .462** .129** .041 1
Table 7: Correlation IV on Risk Aversion
Extraversion Conscientiousness Openness to Exp.
Agreeableness Neuroticism Overconfidence Risk Aversion Gender Year of Birth Level of Education Region #of Children Extraversion 1 Conscientiousness .144** 1 Open. To Exp. .317** .218** 1 Agreeableness .328** .364** .291** 1 Neuroticism -.204** -.254** -.125** -.190** 1 Overconfidence .127** .052 .112** -.024 -.077* 1 Risk Aversion .024 .025 -.094** .083** -.029 -.158** 1 Gender .004 .008 -.034 -.040 -.003 -.015 .035 1 Year of Birth .025 .026 .045 .051 .029 -.040 .004 .153** 1 Level Education .033 -.035 .043 .011 .004 -.011 -.009 -.091** .276** 1 Region .042 -.041 .034 .009 .053 .050 .005 .029 -.004 -.022 1 # of Children .018 .030 .032 .044 .003 -.027 .001 .056* .462** .129** .041 1
3.2.1 Dependent Variable Attitude towards saving money:
The dependent variable attitude towards saving money is significantly correlated with
conscientiousness, agreeableness, neuroticism, overconfidence, region and number of children in the household. Meaning that, with each point scored for conscientiousness, agreeableness, neuroticism, overconfidence, region or number of children, the attitude towards saving money increases with .118 (p<.05), .198 (p<0.01), .143 (p<0.01), .186 (p<0.01), .117 (p<0.05) and .096 (p<0.05).
From this result it shows that high scores on conscientiousness correlates positive with the attitude towards saving (r=.118; p<0.05). Meaning that more organized and disciplined people have a more positive attitude towards saving. The highest correlation between a variable and the attitude towards saving can be found for agreeableness (r=.198; p<0.01). Meaning that individuals which are more generous, trustworthy and tolerant contributes with a more positive attitude towards saving. The same positive correlation can be found for individuals which are more emotional unstable (neuroticism), individuals which are misjudging their ability or opinion (overconfidence), households with more children present and the region in which individuals are living. Accept for neuroticism and overconfidence, these correlations were all expected from the literature. The relations are more explored in the section on multiple regression and are discussed in the conclusion.
3.2.2 Dependent Variable Risk Aversion:
The dependent variable risk aversion is significantly correlated with openness to experience, agreeableness and overconfidence. Meaning that with each point scored on openness to experience, agreeableness and overconfidence, the score of risk aversion increases with .094 (p<0.01), .083 (p<0.01) and -.158 (p<0.01).
explained in the section on multiple regression, which is discussed after the multicollinearity is given.
3.3. Multicollinearity:
Multicollinearity occurs when more than one variable is used trying to predict one dependent variable, and the independent variables are highly correlated (Allison, 2012). This means that one independent variable can be linearly predicted by the other variable. In order to test for Multicollinearity, the Variance Inflation Factors are being explored. Each predictor has a VIF, this shows how much the influence is of that variable on the other variables used. A VIF is considered good with a score lower than 2.5 (Allison, 2012). The scores are calculated using 1/(1-R2) (Allison, 2012). The scores are shown in the Appendix. All VIF scores are around 1, so multicollinearity is not present in both models. Meaning that none of the independent variables are predicted by other variables. Multiple regression can now be performed and discussed. First, assumptions need to be met in order to perform a multiple regression. This is discussed in the next section.
3.4 Assumptions:
For this research a linear regression is used. When using this regression the data must first pass several checks. This makes sure that the data that is used can actually be analyzed using the linear regression. There are six assumptions that needs to be passed in order for the data to be sufficient. Variables should be measured on an interval level, linear, no outliers, no
autocorrelation, homoscedastic and errors should be normally distributed. All assumptions are met and supporting figures are shown in the Appendix. Meaning that the multiple regression can be performed, which is discussed in the next section.
3.5 Multiple Regression
Since this research tries to predict to dependent variables from the same set of independent variables, two multiple regressions are performed and results are discussed individually. The results from the linear regression are shown in table 8 and table 9. Table 8 shows the results from the multiple regression of the households for the attitude towards saving money. Table 9 shows the results from the multiple regression of the households for the level of risk aversion. The prediction for both models was based on the personality traits and control variables. Three models were formed, where the first model was a vector of the Big Five personality traits. Overconfidence was added in the second model and in the third model, control
saving, the models are first analyzed and discussed. This is done by looking at the R2, the adjusted R2 and the significance of the F change. R2 indicates how much of the dependent variable is explained (Frost, 2013). The adjusted R2 which gives an indication on how much of the dependent variable would be explained if a sample from the population was used, in contrast with R2 which looks at the explained variance using the dataset(Frost, 2013). The F change indicates if the model is significantly improved when adding new variables. For instance, model 2 does not significantly improve the model if a value of p>.05 is found. Meaning that model 2 would not be the best fitting model in order to predict the dependent variable. In order to come up with the best fitting model for this paper, the results are discussed.
3.5.1 Multiple regression for the Attitude towards saving money
The first dependent variable regressed was the attitude towards saving money. Table 8 shows the results for all models. The three models will be first explored in order to indicate the best fitting model in predicting the attitude towards saving money.
When looking at model 1 a significant regression equation was found with an R2 of .081. Meaning that the vector of Big Five Personality traits accounts for 8.1% of the variation in the attitude towards saving. The Adjusted R2 is 7.1% which is 1% lower than the R2. This means that if the model was derived from the population rather than the data used, it would explain 1% less, or 7.1% in total, of the attitude towards saving. The F change of 7.978 is significant with p=.000. Model 2 gives an adjusted R2 of .113 with an R2 of .124, which is again 1% lower meaning that a sample from the population will explain 11.3% of the variation in the attitude towards saving. The significant F change of model 2 is 22.197 (p=.000). The
significance of the F change means that adding overconfidence in the regression significantly improves the prediction. Compared to model 2, model 3 shows a R2 of .143 and an adjusted R2 of .120. The F change, 1.810, is not significant (p=.109), meaning that adding the vector of control variables to the regression does not improve the prediction. The best fitting model for predicting the attitude towards saving is therefore model 2.
When looking at model 2, a significant regression equation was found (F(6,451)=10.659, p<.000), with an R2 of .124. The F-ratio is 10.659 with p=.000, meaning that the Big Five personality traits and overconfidence significantly improves the ability to predict the attitude towards saving. Households predicted attitude towards saving money is equal to:
The attitude towards saving money increased with .173 (p=.043) with each point conscientiousness increased, increased with .349 (p=.000) points with each point
agreeableness increased, increased with .282(p=.000) points with each point neuroticism increased and increased .200(p=.000) points with each point overconfidence increased. Overall, 12.4% of the attitude towards saving can be explained by using the Big Five personality traits and overconfidence. Low R2 values are often the case for psychology studies since individuals are unpredictable (Frost, 2014). Low R2 does not mean that no significant results can be withdrawn from the regression. Based on the results of table 8, the conclusion can be made that individuals scoring high on conscientiousness, agreeableness, neuroticism and overconfidence have a more positive attitude towards saving money. agreeableness was found to be the best predictor of the attitude towards saving money (b= .349, p=.000). The results are reflected on the hypotheses which were formed after literature was explored.
The hypothesis for conscientiousness is supported by the findings using the DNB Household data and are in line with findings from existing literature. The hypothesis used in this paper was: Conscientiousness has a positive effect on the attitude towards saving (b=.173, p<0.05). Nyhus & Webley (2001) found that individuals scoring high on conscientiousness save more. For their research, Nyhus & Webley (2001) used multiple forms of savings in an amount of euro’s. This paper looks at the attitude towards saving, which suggest that individuals scoring high on Conscientiousness have a positive attitude towards saving and, suggested by Nyhus & Webley (2001), end up saving more.
When looking at the hypothesis stated for agreeableness: Agreeableness has a negative effect on the attitude towards saving(b=.349, p=.000), table 8 shows that agreeableness has a positive relationship with the attitude towards saving. The positive relationship is not in line with the hypothesis. Individuals scoring high on agreeableness are trusting, cooperative and forgiving (Brown and Taylor, 2011). Meaning that, more trusting, cooperative and forgiving individuals have a more positive attitude towards saving, this contradicting result will be discussed in more detail in the next section.
When looking at the hypothesis formed for neuroticism: Neuroticism has a negative effect on the attitude towards saving (b=.200, p=.000), the hypothesis is not supported by the results. Where a negative effect was suspected from the literature, a positive effect was found using the DNB Dutch Household data. Nyhus & Webley (2001) explain the positive relation between emotionally stable individuals and saving by looking at traits that are part of
& Webley, 2001). Therefore, emotional stable individuals are more likely to be able to follow their own plans and budgets in contract with emotional unstable individuals. Since this paper focusses on the attitude towards saving money and Nyhus & Webley (2001) used the amount of savings, contradicting results can be expected. Based on the results in table 8, it is argued that emotional unstable individuals, scoring high on neuroticism, are willing to save and have a positive attitude towards saving money. This is an interesting result, which will be discussed in more detail later.
Overconfidence also shows an contradicting result based on suggestions from the literature used. Based on the literature the following hypothesis was formed: Overconfidence has a negative effect on the attitude towards saving (b=.200, p=.000). Where a negative relation was suspected, a positive relation was found between overconfidence and the attitude towards saving. Results based on the DNB Household Data suggest that overconfident individuals have a more positive attitude towards saving money.
Based on the results it can be concluded that the attitude towards saving money is not
significantly predicted by extraversion and openness to experience. The following hypothesis are therefore not supported by the data:
Hypothesis 1: Extraversion has a negative effect on the attitude towards saving (b=.002, p>0.05).
Hypothesis 9: Openness to experience has a positive effect on the attitude towards saving (b=-.007, p>0.05).
3.5.2 Multiple regression on the Level of Risk Aversion
The second dependent variable regressed was the level of risk aversion. The results are shown in table 9. The three models will be first explored in order to indicate the best fitting model in predicting the attitude towards saving money.
When looking at model 1 a significant regression equation was found with an R2 of .021 and a significant F change of 4.094 (p=0.001). Meaning that the vector of Big Five Personality traits accounts for 2.1% of the variation in the risk aversion. The Adjusted R2 is 1.6% which is .5% lower than the R2. This means that if the model was derived from the population rather than the data used, it would explain .5% less, or 1.6% in total, of the attitude towards saving. Compared to model 1, model 2 gives a R2 of .044 and a adjusted R2 of .039. Meaning that a sample of the population would explain 3.9% of the variation in risk aversion. The significant F change is 24.117 (p=0.000). The significance of the F change means that adding
2, model 3 shows an adjusted R2 of .036. The F change, .547, is not significant (p=.740), meaning that adding the vector of control variables does not significantly improve the prediction. The best fitting model for predicting the level of risk aversion is therefore model 2. When looking at model 2 a significant regression equation was found (F(6,968)=7.512 p<.000), with an R2 of .044. The F-ratio is 7.512 with a p=.000, meaning that the Big Five personality traits and Overconfidence significantly improves the ability to predict Risk Aversion. Households predicted Risk Aversion is equal to:
5.696 + .106 (Extraversion) + .059 (Conscientiousness) - .306 (Openness To Experience) + .209 (Agreeableness) - .029 (Neuroticism) - .218 (Overconfidence).
The level of risk aversion decreased with .306 (p=.001) points with each point openness to Experience increased, increased with .212 (p=.031) points with each point agreeableness increased and decreased .219 (p=.000) points with each point overconfidence increased. Meaning that individuals scoring high on overconfidence, openness to experience and low on agreeableness are more risk loving.
First, openness to experience showed to be the best predictor of risk aversion (b=-.306, p=0.001). When looking at the hypothesis formed for openness to experience: Openness to experience has a negative effect on risk aversion, it can be concluded that this hypothesis is confirmed using the DNB Household data. Individuals scoring high on openness to
experience are more open minded, knowledgeable and curious which could explain the negative relation on risk aversion.
Second, when looking at the hypothesis formed for agreeableness: Agreeableness has a positive effect on risk aversion (b=.212, p<0.05), it can be concluded that this hypothesis is confirmed by the DNB Household data. Individuals scoring low on agreeableness are more self-centred and uncooperative, on which Bucciol & Zarri (2015) explained their result that low agreeableness is related to risk loving.
Third, when looking at the hypothesis formed for overconfidence: Overconfidence has a negative effect on risk aversion (b=-.219, p=.000), it can be concluded that this hypothesis is confirmed by the DNB Household data. Which makes sense since, overconfident individuals believe in their selves and that they are better off than others, due to their own abilities and skills resulting in confident individuals taking up more risk.
Hypothesis 2: Extraversion has a negative effect on risk aversion (b=.106, p>0.05). Hypothesis 6: Conscientiousness has a positive effect on risk aversion (b=.059, p>.05). Hypothesis 8: Neuroticism has a positive effect on risk aversion (b=-.029, p>0.05). The conclusions based on these results are reviewed in the next section.
Table 8: Linear Regression IV on Saving Money
Model 1 Model 2 Model 3
R2 .081 .124 .142
F Change 7.978 22.197 1.810
Sig. F change .000 .000 .109
* Correlation is significant at the .05 level ** Correlation is significant at the .01 level
Table 9: Linear Regression IV on Risk Aversion
Model 1 Model 2 Model 3
R2 .021 .044 .047
F Change 4.094 24.117 .547
Sig. F Change .001 .000 .740
4 Conclusion
This section reviews the main findings of this paper and how this relates to the hypotheses which were formed after reviewing existing literature. After these main findings are reviewed, the limitations are discussed and suggestions for further research are given.
This research tried to explain the effects of personality traits on the attitude towards financial decisions. From the literature several personality traits are obtained. These personality traits are extraversion, openness to experience, agreeableness, neuroticism, conscientiousness and overconfidence. These personality traits were used to explain two dependent variables. These dependent variables are used in order to capture the attitudes towards saving money at the broadest level. The dependent variables are the attitude towards saving money and risk aversion. The DNB Household Survey 2015 is used in order to explain the effects of personality traits on the attitudes towards financial decisions. Results are discussed in the previous section and the conclusions will be reviewed next. Conclusions are made for each dependent variable separately. First, the conclusions for the effects of personality traits on the attitude towards saving money is reviewed.
4.1 Personality traits and Attitude towards saving money
Based on the literature it was hypothesized that the attitude towards saving money was positively influenced by agreeableness, conscientiousness and openness to experience. Negative influence on the attitude towards saving was expected from extraversion, neuroticism and overconfidence.
Based on the DNB Household survey, the conclusion can be made that the attitude towards saving money is explained by several personality traits. The most important personality trait positively influencing the attitude towards saving money is agreeableness. This is not in line with the findings of Nyhus & Webley (2001) which suggested that individuals scoring high on agreeableness borrow more and save less. The results from Nyhus & Webley (2001) were based on the Dutch data from Center Saving Survey from 1996 and 1997 and looked at the amount of money saved. This unexpected result is explained by looking closer at the
personality traits which are part of agreeableness. It is suggested that individuals scoring high on agreeableness are more likely to donate money (Nyhus & Webley, 2001), which could explain the differences between the positive attitude towards saving but ultimately saving less money. The attitude towards saving is measured using 16 variables, of which 5 variables were related to saving money in order to provide for others. The positive relation between
concerning saving money for children or other relatives.. Agreeable individuals could be guided in saving more money, since they are willing to, by looking at the factors influencing the non-saving behavior. This can be done by looking at the high likelihood of agreeable individuals to donate money (Brown & Taylor, 2011).
Nyhus & Webley (2001) also found a negative relation between saving and neuroticism, which formed the basis of the hypothesis for this paper. From table 8, a positive relation can be found between neuroticism and the attitude towards saving. This is an important result, since previous literature states that neuroticism is negatively related to saving money. This research suggests that emotional unstable individuals (high neuroticism) have a positive attitude towards saving, and therefore are willing to save. This can be explained by looking closer the traits of an individual scoring high on neuroticism. Emotional unstable individuals lack planning and self-control (Brown & Taylor, 2011). This could be the reason why emotional unstable individuals end up not saving money whilst having a positive attitude towards saving. Where other researchers suggested to improve savings of emotional unstable individuals by giving guidance, this research suggest that this guidance is most effective when looking at the reasons why savings are not realized, such as lack of self-control.
Nyhus & Webley (2001) also found that conscientiousness has a positive effect on saving, which is supported by result from this research. Meaning that individuals that are more realistic and disciplined have a positive effect on the attitude towards saving. Suggesting that individuals with low scores on conscientiousness are less positive towards saving. This is explained by individuals scoring high on conscientiousness realistically looking at the future and understanding the importance of saving money. This finding can be used in order to make individuals scoring low on conscientiousness, who are unrealistic and undisciplined, more aware of the importance of saving money.
This research also found a positive significant relation between overconfidence and the attitude towards saving money. Which is an interesting finding since this research is the first to explain the attitude towards saving by looking at overconfidence. Conclusions can be drawn with regards to individuals scoring low on overconfidence, which would result in lower scores for the attitude towards saving. This finding can be used in order to make individuals scoring low on overconfident more aware of the importance of saving money.
be used in more effectively making individuals aware of the importance of saving money, and in more effective guidance on how to save money.
4.2 Personality traits and Risk Aversion
In order to best understand the attitude towards financial decisions, risk aversion is also explained using personality traits. Based on the literature it was hypothesized that the level of risk aversion was positively influenced by agreeableness, conscientiousness and neuroticism. Negative influence on the level of risk aversion was expected from extraversion, openness to experience and overconfidence.
The strongest significant negative relation can be found between openness to experience and risk aversion. This is in line with the hypothesis and in line with the findings by Beyer et al. (2015). The negative relation between openness to experience and risk aversion is explained by the fact that individuals scoring high on openness to experience are found to be more curious (Goldberg, 1993). These individuals are also more open minded and therefore open to new experiences and ideas (Almlund, 2011). These findings can be used in order to reduce unnecessary risk taking by high scoring individuals on openness to experience by supressing the curiosity and willingness to try new things such as buying risky stocks or options just because they are curious.
A negative relation can also be found between overconfidence and risk aversion, which is in line with the hypothesis stated. This is explained by the fact that overconfident individuals mistakenly believe they control outcomes based on their own ability and opinion (Barber & Odean, 2000). It is logical that these individuals end up taking more risk, since they believe the outcome of an action is based on their own, misjudged, ability. This insight can be used in order to prevent individuals from misjudging risks, by letting them think about how certain actions and outcomes are outside the influence of themselves.
The only significantly positive relation that could be found was agreeableness on risk
aversion. This is in line with the hypothesis stated and in line with the results from Bucciol & Zarri (2015) which based their results on the US Health and Retirement Study from 2006-2012. Individuals scoring low on agreeableness are selfish and distrust other individuals (Goldberg, 1993). This can be used in order to prevent individuals scoring low on agreeableness taking up to much risk.
money. Openness to experience, agreeableness and overconfidence can be used in order to explain the level of risk aversion. Limitations and suggestions for further research are discussed in the next section.
5 Limitations and Further Research
The conclusion was reviewed in the previous section, suggesting some interesting and useful insight on the effects of personality traits on financial decisions. The limitations and
suggestions for further research are discussed in this section.
First, when looking at the data that is used for this research, the number of observations seems useful. Over 5000 individuals were interviewed and obtained in the dataset. But when looking at the different sections and variables used, only a small subsample can be used in order to predict the dependent variables. When looking at the attitude towards saving, with a N of 458, the response rate was very low.
Second, is the way in which household’s financial decisions are made. Partners have influence on the decisions as well, even though individuals state that they are the financial decision taker in a household doesn’t automatically mean that they individually make these decisions. Popovici (2012) even indicated that married couples save less. For this research the decision was made to only take up those individuals who make financial decisions, since they are the ones which personality matters most when making these decisions. So, when
individuals lived with their partner, the financial decision maker was only included. When including all individuals, also the ones that do not take part in the decision making, a biased result would have been suspected. When individuals, which were not part of the financial decision making process, were included in the dataset the opposite limitation was formed. With only looking at the financial decision maker, the influence of personality traits on the attitude towards financial decisions were thought to best be captured. Future research could overcome this by looking at one person households in which you can be certain that the financial decisions are made by one person, differences between individual decision makers and ones with a partner present could be analysed.
