THE INFLUENCE OF PARENTAL FINANCIAL TEACHING OF CHILDREN ON THEIR FUTURE RISK AVERSION

(1)

Master thesis Finance

By Daniëlle Æbelina van Dalfsen

THE INFLUENCE OF PARENTAL

FINANCIAL TEACHING OF

CHILDREN ON THEIR FUTURE

(2)

THE INFLUENCE OF PARENTAL FINANCIAL

TEACHING OF CHILDREN ON THEIR FUTURE

RISK AVERSION

Master thesis Finance

By Daniëlle Æbelina van Dalfsen

University of Groningen

Faculty of Economics and Business

June, 2014

Msc Finance

Supervisor: prof. dr. C.L.M. Hermes

Wielewaalplein 136

9713 BR Groningen

d.c.van.dalfsen@student.rug.nl

(3)

ABSTRACT

This study investigates the influence of parental financial teaching in childhood on the risk aversion of an individual during adulthood. Three ways of parental teaching are distinguished: pocket money, advice and control of parents in money usage. In this study the dataset of the Dutch DNB Household survey (year 2012) is used. The final sample in this study consists of 2,019 observations over 2012. This study distinguishes itself of other literature by including parental teaching as an independent variable in relation to risk aversion, which is not investigated before. Overall, we can conclude that parental teaching in childhood has an effect on risk aversion at an adult age. However, not all parental teaching forms work in the same direction. Results show that receiving pocket money in childhood leads to less risk averse individuals at an adult age. Results on parental control show that a higher level of parental control perceived in childhood leads to more risk averse individuals. This is not in line with the expectation on forehand. Receiving pocket money and parental control in childhood show the strongest relation, a strong conclusion about receiving advice in childhood and risk aversion cannot be made. However, the results slightly indicate receiving advice in childhood shows a negative relation with risk aversion.

Keywords: parental teaching, risk tolerance, risk aversion, parental control, pocket money, parental advice

(4)

1. INTRODUCTION

Financial decision making is an important household topic. A vast amount of research has been conducted to distinguish what drives peoples financial decision making. It is often acknowledged that people are not always rational beings. According to Benartzi & Thaler (2007) most people cope with financial behaviour by adopting simple heuristics, or rules of thumb. However, Benartzi & Thaler (2007) describe that these simple heuristics or rules of thumb can lead to systematic biases, therefore households can make the wrong financial decisions.

There is a large amount of research on the determinants of financial behaviour, for example the effect of financial literacy. Studies like Jappelli & Padula (2013), van Rooij, Lusardi & Alessie (2011) and Ameriks, Caplin & Leahy (2003) provide evidence of a link between financial literacy and saving behaviour. They discovered that higher financial sophistication is associated with greater wealth. Also Bayer, Bernheim & Scholz (2008) found significantly higher participation in saving plans when employers offered retirement seminars to their employees. The employees investigated in this study were educated and therefore demonstrated a higher participation than their uneducated colleagues. Clearly, there is a broad range of literature focusing on the effect of financial literacy on financial behaviour. However, more factors could determine the financial behaviour of people, such as the financial education given by parents in their childhood. The parental financial teaching topic is related to the topic of financial literacy: both refer to some kind of education. Bucciol & Veronesi (2013) studied whether parental teaching at a young age supports higher savings at an adult age. Their finding is that parental financial teaching has a significant positive effect on the saving behaviour of adults. According to their study providing pocket money and parental control seem to be the most effective strategies to achieve this. Norvilitis and MacLean (2010) find that college students have less credit card debt when mentored by parents with financial skills. Mentoring by parents leads to a lower impulse driven credit card use. Also Webley and Nyhus (2006) demonstrated parental behaviour and parental orientations have a clear impact on both a child’s and an adults economic behaviour. These studies present a correlation between parental financial education in people’s childhood and their financial behaviour at an adult age.

(6)

points at the potentially significant role of financial education early in life (Ryack, 2011). However, parental financial teaching has never been investigated in combination with risk tolerance. Therefore, this paper investigates whether parental financial education provided to a child influences the child’s risk aversion at an adult age. Three ways of parental teaching are distinguished: pocket money, control and advice. The main question of this study is: Will an adult’s risk aversion change when received financial education from its parents at childhood? Parental financial teaching and financial education received of parents are terms that are used interchangeably in this paper. This study contributes to existing literature by investigating a relationship that has not been investigated previously.

To answer the main question the dataset of the DNB Household Survey (DHS) is used. The final sample consists of 2,019 observations made in 2012. This survey includes several general socio-demographic questions and several questions on parental methods to teach children how to cope with financial issues. Furthermore, questions on an individual’s risk tolerance are addressed. Therefore, this dataset supplies the necessary data and enables the investigation into the research topic.

Out of this research there can be concluded that parental financial teaching has an influence on risk aversion. However, different models led to different results. People that received more pocket money in childhood are less risk averse at adult age. However, not in line with the expectation, receiving a high level of control leads to more risk averse individuals at an adult age. The receiving of pocket money and parental control gave the most significant results. The advice model did not show very strong results in relation to risk aversion.

(7)

2. LITERATURE AND HYPOTHESIS DEVELOPMENT

This section is the theoretical backbone of this thesis in which existing literature on risk tolerance and parental financial teaching is described. The literature review is used to formulate the hypotheses.

2.1 LITERATURE REVIEW

Economists have studied the relationship between education and wages for decades now. During that time the focus was on the relationship between education and work skills. However, work skills are not the only economic objective of education. It is also important to develop decision-making skills. It may enhance individuals to weigh alternatives, exploit opportunities and achieve personal objectives (Bayer et al., 2009). In recent years more focus has been applied to the relationship between education and decision-making skills. In Finance literature the focus is especially on financial education (literacy) and financial decisions and behaviour. Education of financial decisions may influence a wide range of behavioural aspects, including pension plan participation, voluntary contributions, portfolio mix, and the individuals overall saving rate (Bernheim & Garrett, 2003). In this paper we are curious to see whether parental financial education at childhood also influences the behaviour regarding financial risk aversion. In this paper the topics of education, financial knowledge and parental teaching at childhood are related to each other, because all three topics relate to ‘learning’. Therefore, the literature on education and financial knowledge are also discussed.

(8)

cognitive abilities is significantly more willing to accept risks in the lottery experiments performed in this study. This relationship between cognitive ability and risk aversion is demonstrated by people having all ages and both genders, although the relationship is somewhat weaker for females and younger individuals (Dohmen et al, 2010). Furthermore, Donkers, Melenberg and van Soest (2001) found highly educated respondents to have a higher risk tolerance in hypothetical gambling. In contrast to former literature, Benjamin et al. (2005) show higher test scores achieved by Chilean high school students to be related with risk neutrality.

Wang (2009) found in his research that objective knowledge, subjective knowledge, and risk taking are highly correlated. Objective knowledge refers to what an individual actually knows, whereas subjective knowledge is the knowledge that the person believes he possesses. Subjective knowledge is considered to include an individual’s degree of confidence in their knowledge (Wang, 2009). Objective knowledge may enhance subjective knowledge. In other words, factual knowledge enhances self-confidence. This can lead to a higher willingness to accept risks. Consistent with the notion of objective knowledge enhancing subjective knowledge, male investors tend to have higher subjective knowledge as well as objective knowledge than female investors (Wang, 2009). Gender is therefore another important factor that differentiates investors’ level of knowledge and risk taking. There is evidence explaining women in general are more risk averse than men (Agnew, Anderson & Gerlach, 2008), so they may take less riskier decisions. However, generally speaking it is also proven women to be less financially educated than men (Agnew et al., 2008) which could lead to women’s risk aversion being higher.

Other factors such as age, income and wealth play an important role for individuals risk tolerance for a sample of US households (Riley & Chow, 1992). Grable (2000) finds that risk tolerance is associated with being male, of older age, married, receiving a higher income, having completed a higher educational level and having a higher financial knowledge. Jacobs-Lawson & Hershey (2005) also provide consistent evidence for this topic. Age is a widely studied determinant for risk tolerance. However, empirical studies find mixed results. Grable (2000) finds that older age is associated with risk tolerance. However, Riley & Chow (1992) find that risk tolerance increases with age up to a point (65 years) and then decreases. The majority of research finds that risk tolerance decreases with age (Hallahan et al., 2004; Gilliam, Chatterjee, & Zhu, 2010). According to Hallahan et al. (2004) this can be intuitively explained by the fact that younger individuals have more years to recover from losses that may be incurred with risky investments. The reason for this inconsistency between research on age and risk tolerance is not clear and may be caused by differences in both the methodologies being applied and the sample demographics (Ryack, 2011).

(9)

‘learning’. Jappelli & Padula (2013) assume people to have an initial stock of financial literacy from their childhood. They actually found evidence that improving mathematical skills early in life will eventually raise households financial literacy and wealth. However, there is no evidence yet that parental financial education at childhood has a relationship with risk tolerance at an adult age. Parental teaching in this study can be viewed as the educational role played by parents in teaching children how to deal with financial issues. This study distinguishes three kinds of parental teaching: giving pocket money, giving advice to children about financial matters and parents trying to control their children’s financial behaviour. When children receive pocket money or advice from their parents, it is found that these children are more knowledgeable in respect to money issues (Lewis & Scott, 2000). These children have the knowledge and experience how to deal with their money. Furthermore, they feel the responsibility to deal with their own finances. This responsibility is not felt when children receive a lot of parental control on the way the children spend their money. Therefore, they are less able to develop self-control. Risk-avoiding behaviour requires self-control, less self-controlled individuals leads to being less able to recognize, forecast and manage risks (Fischer, Kastenmüller & Asal, 2012).

Continuities between parents behaviour and children’s behaviour are found in research. Webley & Nyhus (2006) focused on economic socialization of parents to children. They provide evidence of a link between the parental approaches to economic matters and that of their children. The future orientation of parents is correlated with that of their children, which is referred to as evidence for economic matters to be transmitted from parents to children (Webley & Nyhus, 2006). Morrongiello, Corbett & Bellissimo (2008) find that both parental teaching on safety and how parents approach risks each uniquely contribute to children’s safety and risk practices, either in the short-term or long-term (in adulthood). Furthermore, Bucciol & Veronesi (2013) illustrated parental teaching to have a positive influence on the saving attitude at an adult age. Also Norvilitis & MacLean (2010) investigated how parental teaching affected students’ financial behaviour and attitude. More concrete, they tied parental teaching to the credit card debt of students. They found that more parental teaching is significantly related to a lower level of credit card debt of students. This implies that parental teaching is persistent throughout adult age. However, these are the only studies (Bucciol & Veronesi, 2013; Norvilitus & MacLean, 2010) focusing on parental teaching and the influence on financial behaviour at an adult age.

(10)

2.2 HYPOTHESES DEVELOPMENT

The objective of this study is to relate an individual’s risk aversion with parental financial teaching during the individual’s childhood. The study tests three hypotheses. The three different strategies of parental teaching are tested in three separate models, to see whether and if so, what the effects of the different strategies of parental teaching are on risk aversion.

Pocket money

Hypothesis 1: There is a negative relationship between receiving pocket money in an individual’s childhood and the financial risk aversion at an individual’s adult age.

When children received pocket money from their parents, children are more knowledgeable with respect to money issues (Lewis & Scott, 2000). These children get more experience in dealing with money and therefore learn about money issues. This study expects receiving pocket money from parents will lead children to have a higher understanding of financial issues. As described above, people who finished a higher educational level or have a higher financial knowledge are less risk averse. According to Jappelli & Padula (2013) people receive an initial stock of financial literacy from their childhood. Improving financial skills early in life will raise households financial literacy in the long run. Therefore, in this study we expect a negative relation between providing pocket money and the financial risk aversion at an adult age of an individual.

Advice

Hypothesis 2: There is a negative relationship between received advice from parents on financial issues in a person’s childhood and the financial risk aversion at an adult age of that individual. A negative relationship is expected between the advice that children receive from their parents and their financial risk aversion at an adult age. This is in line with the reasoning of hypothesis 1. When children are given advice on financial matters, they also become more financially educated and receive more financial knowledge. Literature states financial literacy to have a positive influence on financial decisions. Moreover, the higher the level of financial education, the less risk averse a person may be. When given advice on financial matters children obtain more knowledge on financial issues which causes us to suspect them being less risk averse at an adult age.

Control

Hypothesis 3: There is a negative relationship between a high level of parental control on the way a child spends its money in an individual’s childhood and the financial risk aversion of an individual at an adult age.

(11)

By perceiving a high level of parental control in childhood on money issues, we expect these children do not learn how to develop self-control. They do not experience the feeling of controlling themselves, because their parents are doing that for them. Therefore, we expect that when parents decide on all their children’s financial issues, the children do not learn how to discipline themselves from a financial perspective. Therefore children will be less risk averse at an adult age. At first sight this argumentation contradicts the above two argumentations of hypothesis 1 and 2. However, control is a somewhat different kind of parental teaching. People most effectively learn by doing and getting experience out of it, when children are completely controlled by their parents they do not receive this experience (Hallahan et al., 2004). Therefore, a negative relationship is expected between a high level of parental control on the way a child spends its money and the financial risk aversion that child has at an adult age.

(12)

3. METHODOLOGY

This section contains the data and methodology used in this study. First, the sample selection method and the data source is addressed. Furthermore, the empirical model used in this study is shown. Finally, the different variables in the research are described.

3.1 SAMPLE SELECTION AND DATA SOURCES

The data used in this study have been retrieved from a dataset which contains approximately 2,000 households within the Netherlands. The dataset results from the DNB Household survey (formerly known as CentER Savings study). Since 1993, CentER data uses an internet panel to collect economic data. In case a respondent does not have an internet access the person will receive the appropriate device and support to complete the survey. The data is available in the period of 1993 to 2012. In each participating household, every person older than 16 years is interviewed. So multiple persons in one household can be part of this survey. The goal of the DNB Household survey is primarily to study the psychological and economic aspects of financial behaviour. Additionally, it includes information about peoples work, pensions, housing, mortgages, income, assets & debts, health, as well as their demographic information. Using this data the objective of this study, i.e. to investigate the relationship between parental teaching in childhood and risk tolerance, can be achieved. Specification of the data sample are shown in Table 1.

The final sample consists of 2,019 observations made in 2012. The original dataset consists of 4,102 observations, a reservation has to be made because not all respondents answered all required questions in this research. This study uses data from the year 2012. The parental teaching questions are only asked once to every respondent in all of the years, meaning that it is not panel data. Therefore, the decision was made to take the most recent cross-sectional dataset.

TABLE 1. Summary of sample selection process

All respondents included in the dataset for the year 2012 4,102

Less: respondents who did not fill in their education level 9 Less: respondents who did not fill in the risk tolerance questions 2,074

Number of observations available for the year 2012 2,019

3.2 EMPIRICAL MODEL

The objective of this study can be tested with the following regression equation for the household head i in year 2012.

(13)

where

ε

i,t is the error term and risk aversioni,t is the dependent variable. The model is estimated three times, in each model a different type of parental teaching (pocket money, control, advice) is used. 3.3 MEASUREMENT

Table 2 shows the variables used in the analysis. A distinction is made between the dependent variable, independent variables and the control variables.

Dependent variable

In this study the dependent variable is the risk aversion of the adult individual. The risk aversion is measured by the following statements which are scaled from 1 to 7 (totally disagree to totally agree) in the DNB Household survey:

TABLE 2. Description variables Variables used

in research

Description

Age The age of the respondent, which is calculated by: 2012 less year of birth respondent

Gender Gender of the respondent (dummy: 1 = female, 0 = female)

Education Highest level of education completed of the respondent

Highest level of education completed of the respondent divided in three categories: low, moderate and high education

Risk aversion factor 1

Risk3

Risk5

Risk6

‘If I think an investment will be profitable, I am prepared to borrow money to make this investment’

‘I get more and more convinced that I should take greater financial risks to improve my financial position’

‘I am prepared to take the risk to lose money, when there is also a chance to gain money’

Risk1

Risk2

Risk4

‘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 would never consider investments in shares because I find this too risky’

‘I want to be certain that my investments are safe’

Parental teaching

Pocket money ‘When you were between 8 and 12 years of age, did you receive an allowance from your parents then? By allowance we mean a fixed amount received on a regular basis’

Control ‘When you were between 8 and 12 years of age, could you spend

your money as you pleased?’

(14)

1. 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.

2. I would never consider investments in shares because I find this too risky.

3. If I think an investment will be profitable, I am prepared to borrow money to make this investment.

4. I want to be certain that my investments are safe.

5. I get more and more convinced that I should take greater financial risks to improve my financial position.

6. I am prepared to take the risk to lose money, when there is also a chance to gain money. Given the nature of the dependent variable - a limited dependent variable - suitable models for the analysis belong to the classes of ordered probit and ordered logit models. The dependent variable is a limited dependent variable due to the scale and the answers that belong to the possible scores. An ordered logit is performed. Not all statements are consistent with risk aversion. Therefore, the answers are converted in such a way that answer option 1 is always directed towards risk tolerance and answer option 7 towards risk aversion. The questions marked by ‘R’ (see Table 3) are the statements whose score have been reversed to ensure all statements are consistent. Statements 3, 5 and 6 have been asked to the respondent to determine the willingness of taking risks. Statements 1, 2 and 4 are questions that aim to determine the willingness of safety.

In order to determine the factor structure of the six questions of risk aversion a factor analysis is applied. Factor analysis is based on a principal components analysis (PCA), with a varimax rotation. First, the appropriateness of the PCA is investigated. This is performed by applying the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy and the Bartlett’s Test of Sphericity. The KMO measure has to exceed 0.6 to indicate that the data is appropriate for PCA and the Bartlett’s test must be significant at a 5% level. In this case the KMO measure is 0.640 and the Bartlett’s test is significant at a 1% level. This indicates that the data is suitable for a PCA.

(15)

TABLE 3. Factor analysis of the risk aversion using PCA with a varimax rotation Factors Factor 1 (Riskaversion1) Factor 2 (Riskaversion2)

Risk1 -0.041 0.837 Risk2 0.414 0.553 Risk3 (R) 0.644 -0.009 Risk4 0.003 0.861 Risk5 (R) 0.834 -0.020 Risk6 (R) 0.845 0.189

– (R) denotes the statements whose score has been reversed to make all the statements consistent

– The largest factor loadings in each column are highlighted in bold

As can be seen, the questions that are asked in the direction of taking risks (see Table 3) load the most on the first factor. The questions which are more in the direction of safety load on the second factor.

Independent variable

The independent variables in the model include measures of parental financial teaching, which is the main focus of this paper, as well as measures that have been used in previous research on determinants of risk aversion. In particular there are three groups of independent variables in the model:

- Parental financial teaching at childhood. Parental teachingi is a variable that informs on

whether children received teaching from their parents. For each individual it is known whether he or she regularly received pocket money in the age of 12, if the individual in the age of 8-12 was free to decide on how to use the money, and whether the individual received advice about this in the age of 12-16. To investigate which of the three parental teaching methods (pocket money, control and advice) has the strongest influence on risk tolerance, they are investigated separately.

The questions in the DNB Household survey used to measure parental teaching are:

1. Pocket money: When you were between 8 and 12 years of age, did you receive an allowance from your parents then? By allowance we mean a fixed amount received on a regular basis.

2. Control: When you were between 8 and 12 years of age, could you spend your money as you pleased?

3. Advice: Did your (grand) parents try to teach you how to budget when you were between 12 and 16 years of age?

(16)

and advice. It is helpful to create dummies, because the distance between the answer possibilities of the question are not equal.

Control variables

- Age. Ai is a control variable that measures the age of the respondent. The age of a respondent

can influence his or her risk aversion. Risk aversion is associated with age (Riley & Chow, 1992; Grable, 2000) explaining why age is taken as a control variable. The question used in the DNB Household survey for this variable is:

1. Year of birth of the respondent

This variable is converted into the age of the respondent instead of the year of birth. That way the age of the respondents can be categorized. The following three categories have been determined: 1: 16 – 45 years, 2: 46 – 65 years, 3: 65 years and older. Dummies are created for these categories.

- Education. Ei is a control variable that measures the respondent’s highest level of fulfilled

education. Literature has shown that education has a positive influence on risk tolerance. This is the reason why the education level of the respondent is included as a control variable. The question used in the DNB Household survey for this parameter is:

1. Highest level of education completed

This variable is categorized in a low, moderate and high education level. The categorization has been performed as following: Low: Special education, kindergarten/primary education, no education, other sort of education; Moderate: Pre-vocational education, pre-university education, senior vocational training; High: Vocational colleges, university education. Dummies are created for the three different categories.

- Gender. Gi is a control variable for gender differences. Literature suggests that women are

more risk averse than men. Women are also less financially educated, so this could influence the dependent variable risk tolerance. Because women are less financially educated, in general they have less financial knowledge. In line with the other control variable, education, a lower level of education also has an influence on risk tolerance. The question used in the DNB Household survey for this parameter is:

1. What is your gender?

(17)

4. EMPIRICAL RESULTS

In this section the empirical results are presented. First the descriptive statistics are discussed, including a correlation matrix for each kind of parental teaching model (pocket money, control & advice). Next, for each model the results of the regression analysis (ordered logit) are described. 4.1 DESCRIPTIVE STATISTICS

Table 4 shows the descriptive statistics. For each included variable the mean, median, standard deviation, minimum and maximum value are displayed.

TABLE 4. Descriptive statistics

Variable Mean Std. Deviation Minimum Median Maximum Risk aversion Factor 1 16.77 3.644 3 17 21 Factor 2 15.46 4.228 3 16 21 Risk1 5.18 1.874 1 6 7 Risk2 4.90 2.078 1 6 7 Risk3 6.05 1.412 1 7 7 Risk4 5.37 1.591 1 6 7 Risk5 5.10 1.680 1 6 7 Risk6 5.62 1.498 1 6 7 Parental teaching Pocket money 2.38 1.355 1 2 4 Yes 0.45 0.498 0 0 1 Yes, forgotten 0.05 0.225 0 0 1 Occasionally 0.15 0.345 0 0 1 No 0.34 0.475 0 0 1 Control 2.58 1.330 1 3 5 Parents decided 0.29 0.455 0 0 1

Parents decided most 0.21 0.404 0 0 1

Parents decided partly 0.23 0.423 0 0 1

Mostly own decision 0.17 0.374 0 0 1

Own decision 0.10 0.301 0 0 1

Advice 2.47 1.073 1 2 4

Yes 0.23 0.422 0 0 1

Yes, some advice 0.28 0.449 0 0 1

Yes, to certain extent 0.27 0.444 0 0 1

(18)

From the descriptive statistics in Table 4 it can be concluded that most people tend to be risk averse. The mean and median of all risk aversion variables are very close to the maximum. All questions about risk aversion have a high mean (higher than 4) and a high median (higher than 6). Furthermore, the parental teaching questions are all quite evenly answered. The majority of people in the sample always received pocket money (45%). Parents have quite some control over their children’s money usage, in fifty percent of the cases the parents control money usage. The distribution between female and male respondents is quite even (56% male). In this dataset the education level is skewed towards the middle and high level of education. Only 5% of the respondents has a low level of education. The categories of age are quite even distributed.

Table 5, 6 and 7 show the correlation matrices of each parental teaching model. For each variable the Pearson Correlation Coefficient is displayed.

TABLE 5. Correlation matrix Pocket money Variable Risk aversion factor 1 Risk aversion factor 2 Pocket money

Age Gender Education

Risk aversion factor 1 1 .219** .094** .121** .203** -.107** Risk aversion factor 2 1 .036 .052* .118** -.006 Pocket money 1 .363** -.004 -153** Age 1 -.106** -.110** Gender 1 -.102** Education 1

(19)

TABLE 6. Correlation matrix Control Variable Risk aversion factor 1 Risk aversion factor 2

Control Age Gender Education

Risk aversion factor 1 1 .219** -.088** .121** .203** -.107** Risk aversion factor 2 1 -.032 .052* .118** -.006 Control 1 -.252** -.009 .097** Age 1 -.106** -.110** Gender 1 -.102** Education 1

** Correlation is significant at the 0.01 level (2-tailed) * Correlation is significant at the 0.05 level (2-tailed)

A significant negative correlation is shown in Table 6 between risk aversion factor 1 and the control of parents. The risk aversion factor 2 and the control of parents also demonstrates a negative correlation, however this is not significant. The correlation between risk aversion and the control variables are the same as described at Table 5.

TABLE 7. Correlation matrix Advice Variable

Risk

aversion

factor 1

Risk aversion

factor 2

Advice

Age

Gender

Education

Risk aversion factor 1 1 .219** -.004 .121** .203** -.107** Risk aversion factor 2 1 .006 .052* .118** -.006 Advice 1 .225** -.058** -.068** Age 1 -.106** -.110** Gender 1 -.102** Education 1

** Correlation is significant at the 0.01 level (2-tailed) * Correlation is significant at the 0.05 level (2-tailed) There is no significant correlation between giving parental advice and risk aversion. Both factors have no significant correlation and have an opposite direction in the correlation. The correlation between risk aversion and control variables are the same as described at Table 5.

4.2 REGRESSION ANALYSIS

(20)

TABLE 8. Regression results pocket money and risk tolerance Risk aversion

factor 1

Risk1 Risk2 Risk3 Risk4 Risk5 Risk6

Ordered logit Ordered logit Ordered logit Ordered logit Ordered logit Ordered logit Ordered logit Ordered logit Expl. variables (std. err.) (std. err.) (std. err.) (std. err.) (std. err.) (std. err.) (std. err.) (std. err.) Pocket money Yes 0.165 -0.065 0.186 -0.177 0.362* -0.020 0.070 -0.023 (0.179) (0.172) (0.175) (0.173) (0.193) (0.176) (0.181) (0.182) Occasionally 0.218 -0.061 0.087 -0.120 -0.006 -0.072 0.160 0.312 (0.199) (0.191) (0.195) (0.194) (0.213) (0.197) (0.201) (0.204) No 0.326* 0.105 0.326* -0.010 0.274 0.084 0.247 0.309* (0.185) (0.177) (0.182) (0.179) (0.199) (0.182) (0.188) (0.189) Female 0.757*** 0.454*** 0.317*** 0.540*** 0.606*** 0.290*** 0.561*** 0.742*** (0.081) (0.079) (0.082) (0.082) (0.091) (0.081) (0.081) (0.084) Age: 46-64 0.354*** 0.123 0.237** -0.001 0.400*** 0.187* 0.284*** 0.263*** (0.098) (0.096) (0.098) (0.098) (0.106) (0.098) (0.098) (0.100) Age: > 64 0.710*** 0.209* 0.209* 0.143 0.850*** 0.448*** 0.492*** 0.646*** (0.111) (0.109) (0.111) (0.112) (0.124) (0.112) (0.112) (0.115) Middle education 0.159 0.470** 0.329* 0.159 0.294 0.418** -0.052 0.184 (0.199) (0.193) (0.196) (0.198) (0.215) (0.205) (0.195) (0.200) High education -0.147 0.348* 0.495*** -0.361* 0.108 0.429** -0.281 -0.216 (0.200) (0.194) (0.197) (0.200) (0.216) (0.206) (0.197) (0.202) (Pseudo) R2 0.016 0.005 0.005 0.013 0.019 0.006 0.013 0.027 Log likelihood -4805.202 -5281.530 -3460.784 -3585.400 -2572.675 -3292.349 -3454.700 -3063.963

*Denote significance at the 10% level. **Denote significance at the 5% level. ***Denote significance at the 1% level.

(21)

(22)

However, one model (risk3) results in a significant positive relation between receiving pocket money and risk aversion. This would imply that receiving pocket money in childhood leads to more risk aversion. This is not matching our expectations. Risk3 is part of the first factor of risk aversion, which gives a different result described earlier in this paragraph. Risk3 is the variable that loaded least on the first factor of risk aversion. As such, the results are ambiguous in this model. Because of the three models that show results in the direction that receiving pocket money leads to a decreasing risk aversion, we can prudently suggest that this relation is stronger.

All models show females being significantly more risk averse than men. This is in line with literature. Furthermore, most of the models show age is significantly positively related to risk aversion. The older the respondent’s age the more risk averse the respondent is. What also can be concluded out of these models is the relation becomes stronger when older in age. A few models give significant results between education and risk aversion. Most models report a positive relation between middle and high education and risk aversion. This means that people with a middle or high education are more risk averse than people with a low level of education. This is not in line with the expectation of the literature. However, one model (risk2) gives a significant negative relation between high education and risk aversion. These differences may result from the different questions. Only the variables that load on the second factor of risk aversion (variables in the direction of safety) and the second factor of risk aversion itself show significant results with regards to education. This may suggest that the questions asked in the direction of safety lead to a higher number of risk averse answers. Furthermore, the sample is not evenly distributed. Only five percent of the sample has a low education. Therefore, the results regarding education may be flawed.

The pseudo R-squared in all the models is low. Pseudo R-squared has a minimum value of zero and a maximum value of 1. The higher the pseudo R-squared the better the model fits (Brooks, 2008).

(23)

TABLE 9. Regression results control and risk tolerance Risk aversion

factor 1

Ordered logit Ordered logit Ordered logit Ordered logit Ordered logit Ordered logit Ordered logit Ordered logit Expl. variables (std. err.) (std. err.) (std. err.) (std. err.) (std. err.) (std. err.) (std. err.) (std. err.) Control

Parents decided all 0.119 0.131 0.132 0.032 0.086 0.294** 0.129 0.154

(0.115) (0.112) (0.116) (0.117) (0.128) (0.116) (0.116) (0.120)

Parents decided most -0.066 -0.055 -0.060 -0.091 -0.036 0.073 -0.018 -0.123

(0.118) (0.116) (0.120) (0.119) (0.130) (0.118) (0.119) (0.122)

Mostly own decision -0.173 -0.117 -0.070 -0.310** -0.073 0.139 -0.110 -0.254*

(0.124) (0.123) (0.123) (0.125) (0.136) (0.124) (0.124) (0.127) Own decision -0.214 -0.000 0.104 -0.196 0.151 0.241 -0.157 -0.290* (0.148) (0.148) (0.150) (0.151) (0.169) (0.150) (0.151) (0.154) Female 0.757*** 0.447*** 0.318*** 0.531*** 0.600*** 0.284*** 0.560*** 0.730*** (0.081) (0.080) (0.082) (0.082) (0.091) (0.082) (0.081) (0.084) Age: 46-64 0.357*** 0.121 0.229** -0.008 0.356*** 0.172* 0.293*** 0.289*** (0.097) (0.096) (0.098) (0.098) (0.106) (0.097) (0.097) (0.099) Age: > 64 0.710*** 0.202* 0.199* 0.141 0.754*** 0.399*** 0.504*** 0.701*** (0.110) (0.108) (0.111) (0.111) (0.123) (0.111) (0.111) (0.114) Middle education 0.170 0.474** 0.352* 0.136 0.318 0.434** -0.050 0.190 (0.198) (0.193) (0.196) (0.198) (0.215) (0.205) (0.194) (0.199) High education -0.144 0.345* 0.518*** -0.392** 0.147 0.449** -0.288 -0.235 (0.200) (0.194) (0.197) (0.200) (0.216) (0.206) (0.196) (0.201) (Pseudo) R2 0.017 0.005 0.005 0.014 0.018 0.006 0.013 0.027 Log likelihood -4803.300 -5281.093 -3461.598 -3582.411 -2576.432 -3289.463 -3453.992 -3063.417

With respect to the control variables gender, age and education the results are equal to those found in the first model. To clarify: - Females are more risk averse than men.

- The older the age of the respondent the more risk averse the person is.

(24)

TABLE 10. Regression results advice and risk tolerance Risk aversion

factor 1

Ordered logit Ordered logit Ordered logit Ordered logit Ordered logit Ordered logit Ordered logit Ordered logit Expl. variables (std. err.) (std. err.) (std. err.) (std. err.) (std. err.) (std. err.) (std. err.) (std. err.) Advice Yes 0.081 -0.054 -0.023 0.052 0.225* -0.041 -0.005 0.069 (0.111) (0.110) (0.112) (0.112) (0.125) (0.112) (0.113) (0.116) Yes, some -0.126 -0.015 -0.052 0.080 0.027 0.008 -0.116 -0.122 (0.105) (0.104) (0.106) (0.107) (0.118) (0.106) (0.107) (0.109) No -0.056 -0.034 -0.102 0.182 -0.007 0.018 -0.067 0.065 (0.115) (0.114) (0.116) (0.116) (0.187) (0.117) (0.117) (0.120) Female 0.758*** 0.459*** 0.325*** 0.548*** 0.601*** 0.296*** 0.563*** 0.743*** (0.081) (0.079) (0.082) (0.082) (0.091) (0.082) (0.081) (0.084) Age: 46-64 0.388*** 0.142 0.257*** 0.019 0.3997*** 0.195** 0.312*** 0.319*** (0.097) (0.096) (0.098) (0.098) (0.106) (0.097) (0.097) (0.099) Age: > 64 0.782*** 0.267** 0.267** 0.186* 0.828*** 0.473*** 0.561*** 0.774*** (0.107) (0.104) (0.107) (0.107) (0.120) (0.107) (0.107) (0.110) Middle education 0.177 0.479** 0.354* 0.148 0.321 0.424** -0.044 0.203 (0.199) (0.193) (0.196) (0.198) (0.215) (0.205) (0.194) (0.200) High education -0.143 0.343* 0.510*** -0.381** 0.143 0.429** -0.287 -0.225 (0.200) (0.194) (0.197) (0.199) (0.216) (0.206) (0.144) (0.201) (Pseudo) R2 0.016 0.005 0.004 0.013 0.018 0.005 0.012 0.025 Log likelihood -4805.543 -5283.269 -3463.224 -3585.874 -2575.528 -3293.163 -3455.926 -3068.456

(25)

5. FURTHER ANALYSIS

The former section describes the empirical results. The advice model gave no significant results regarding the independent variable. Furthermore, the pocket money model gave ambiguous results. In this chapter two other methods, ordinary least squares (OLS) and binary probit, are used to determine whether the results are consistent. The first method described here is using the OLS. In the first analysis the ordered logit is performed. However, on the basis of a factor analysis two new variables can be generated automatically, those variables being continuous variables. These two new variables are called factored risk aversion 1 and factored risk aversion 2. An OLS regression is used to estimate the three models with the factored risk tolerance. OLS is a proper method because the dependent variable is a continuous variable. Table 11, 12 and 13 show the results of this first method.

TABLE 11. OLS Regression pocket money and risk tolerance

Factored risk aversion 1 Factored risk aversion 2

OLS OLS

Expl. variables (std. err.) (std. err.)

Pocket money Yes 0.032 -0.035 (0.098) (0.102) Occasionally 0.098 -0.074 (0.109) (0.113) No 0.131 0.015 (0.101) (0.104) Female 0.433*** 0.160*** (0.044) (0.045) Age: 46-64 0.163*** 0.068 (0.054) (0.056) Age: > 64 0.316*** 0.075 (0.060) (0.062) Middle education 0.099 0.264** (0.103) (0.107) High education -0.122 0.313*** (0.104) (0.108) Constant -0.433*** -0.371*** (0.135) (0.140) (Pseudo) R2 0.078 0.012 Log likelihood -2782.067 -2852.264

(26)

Table 12 shows significant negative relationships between less parental control and risk aversion. This means the less parental control in childhood is given, the less risk averse an adult is. In Table 9 no significant relationships were found in both the risk aversion factors. The coefficients of the risk aversion factors in Table 9 are in the same direction as in Table 12. Furthermore, in the separate risk models in Table 9, significant relationships were revealed in the same direction as the ones in here. So this confirms findings of the first analysis when more parental control is given in childhood, a person is more risk averse. Additionally, the control variables give the same relations as found earlier, so these are confirmed as well. In this case this model also seems to have a better fit, a higher R-squared is demonstrated.

Table 13 presents no significant relationships between receiving parental advice in childhood and risk aversion at adult age. The direction of the coefficients are the same as in Table 10. It can be seen that most coefficients show a negative relation. This indicates that when there is more parental advice at childhood, these individuals are less risk averse in their adulthood. This matches the expectations. However, these are not significant results and therefore we cannot reach a conclusion based on this TABLE 12. OLS Regression control and risk tolerance

Factored risk tolerance 1 Factored risk tolerance 2

OLS OLS

Control

Parents decided all 0.025 0.046

(0.062) (0.065)

Parents decided most -0.054 -0.040

(0.065) (0.068)

Mostly own decision -0.136** -0.032

(27)

information. The control variables show the same relations as in Table 10, this result confirms these relations.

TABLE 13. OLS Regression advice and risk tolerance

Factored risk tolerance 1 Factored risk tolerance 2

OLS OLS

Advice Yes 0.057 -0.061 (0.061) (0.063) Yes, some -0.049 -0.020 (0.058) (0.060) No -0.031 -0.052 (0.062) (0.064) Female 0.432*** 0.162*** (0.044) (0.045) Age: 46-64 0.186*** 0.070 (0.054) (0.056) Age: > 64 0.365*** 0.089 (0.058) (0.060) Middle education 0.102 0.263*** (0.103) (0.108) High education -0.127 0.308*** (0.104) (0.108) Constant -0.376** -0.366*** (0.114) (0.118) (Pseudo) R2 0.078 0.012 Log likelihood -2782.531 -2852.603

The second method is based on the fact that the sample is skewed in the direction of risk aversion. The mean and median of the risk aversion variables are high. Therefore, in this further analysis two new dummies are created for risk aversion factor 1 and 2. The 0 is awarded to the respondents below the median and an 1 is awarded to the respondents above the median. This hopefully results in a higher distinction between the ‘risk tolerant’ group (dummy = 0) and the ‘risk averse’ group (dummy = 1). Table 14, 15 and 16 present the results of this method.

(28)

TABLE 14. Binary probit regression pocket money and risk tolerance

Dummy factor 1 risk aversion Dummy factor 2 risk aversion

Binary probit Binary probit

Pocket money Yes 0.184 -0.090 (0.130) (0.128) Occasionally 0.162 -0.049 (0.144) (0.142) No 0.282** -0.032 (0.134) (0.132) Female 0.481*** 0.291*** (0.058) (0.057) Age: 46-64 0.226*** 0.081 (0.072) (0.071) Age: > 64 0.362*** 0.088 (0.080) (0.079) Middle education 0.049 0.316** (0.139) (0.136) High education -0.192 0.215 (0.140) (0.138) Constant -0.412** -0.380** (0.181) (0.178) (Pseudo) R2 0.044 0.013 Log likelihood -1324.772 -1380.703

Table 15 shows the results of the control model. This model does not reveal significant results between parental control and risk aversion. Therefore, this model cannot confirm the findings found earlier. However, the direction of the coefficients are the same as those found in the former analysis.

(29)

*Denote significance at 10% level. **Denote significance at 5% level. ***Denote significance at 1% level. TABLE 16. Binary probit regression advice and risk tolerance

Advice Yes 0.035 -0.132* (0.082) (0.080) Yes, some -0.109 -0.022 (0.077) (0.076) No -0.097 -0.060 (0.083) (0.081) Female 0.481*** 0.294*** (0.059) (0.057) Age: 46-64 0.247*** 0.081 (0.071) (0.071) Age: > 64 0.410*** 0.103 (0.077) (0.076) Middle education 0.061 0.311** (0.138) (0.136) High education -0.189 0.204 (0.140) (0.138) Constant -0.195 -0.386** (0.153) (0.151) (Pseudo) R2 0.043 0.014 Log likelihood -1325.234 -1379.670

TABLE 15. Binary probit regression control and risk tolerance

Control

Parents decided all 0.079 0.020

(0.083) (0.082)

Parents decided most -0.013 -0.064

(0.087) (0.086)

Mostly own decision -0.110 -0.024

(30)

6. DISCUSSION

Table 17 summarizes the results of the research, with regards to the defined hypotheses. The first hypothesis is partially confirmed by the ordered logit (Table 8), resulting from ambiguous results in both directions. The OLS gave no significant results on this hypothesis. The binary probit (Table 14) confirmed the first hypothesis. However, pocket money only gave results on the first factor of risk aversion, which is the factor towards taking risks. This indicates that individuals that did not receive pocket money in their childhood, gave stronger risk averse answers on the ‘risk-taking’ questions. Overall, it can be concluded that there is evidence for hypothesis 1 being partially confirmed. There is evidence showing that receiving pocket money at childhood leads to less risk aversion at an adult age on one factor of risk aversion. The second hypothesis gave no results in the first two analyses. However, in the third model there was a small significant (at the 10% level) negative relation between receiving parental advice in childhood and risk aversion. Furthermore, only the second factor of risk aversion led to results. Definitive conclusions cannot be drawn from this result. Further research is required because the results to certain extent suggest a relationship between parental advice in childhood and risk aversion exist. The third hypothesis shows significant results on both the ordered logit and the OLS. There is a significant positive relation found between parental control in childhood and risk aversion at an adult age. However, the ordered logit (Table 9) led to no results on both risk aversion factors, only on the separate models of risk. The OLS (Table 12) generated results on the first factor of risk aversion. The results indicate a positive relation between parental control in childhood and risk aversion at an adult age. Therefore, the hypothesis is rejected. The results are not in line with the expectations set in hypothesis 3.

TABLE 17. Overview results research

Hypothesis Ordered logit OLS Binary probit

1 Pocket money Partly confirmed - Partly confirmed

2 Advice - - Partly confirmed

3 Control Rejected Rejected -

(31)

Bucciol & Veronesi (2013) paper gave results on this regarding saving behaviours. This difference can be explained to the different dependent variable and the giving of advice providing a stronger relation to saving attitudes.

Because parental teaching, financial knowledge and education are all related to ‘learning’, the results can be compared to those found in literature. Especially receiving pocket money and advice at childhood relates to education, because of the experience children gain on money issues. To a certain extent the results of this thesis correspond to the research of Grable (2000). Grable (2000) found that people with higher levels of financial knowledge are found to be less risk averse. We find that people receiving more parental teaching in childhood (and therefore have more financial knowledge) are found to be less risk averse at an adult age. However, we also found one model that showed receiving pocket money results in a significant positive relationship to risk aversion. The coefficients are not consistent. However, the negative relationship is consistent throughout the further analysis, so we can prudently conclude that this is the strongest relationship. Also in line with this is literature stating higher educated people tend to be less risk averse (Riley & Chow, 1992; Donkers, Melenberg & van Soest, 2001; Carducci & Wong, 1998). This thesis distinguishes itself from these findings by including parental teaching as a form of ‘learning’ and education. To our knowledge, this has not been investigated yet.

(32)

7. CONCLUSION

In this study we analysed whether parental financial teaching in childhood influences the risk aversion at an adult age. We analysed the DNB Household Survey for the year 2012 to study this effect. The research question of this study was: Will an adult’s risk aversion change when received financial education from its parents at childhood? We demonstrated relations to exist between receiving of pocket money in childhood and the risk aversion at an adult age. Receiving pocket money at childhood on average leads to less risk averse persons. Furthermore, more parental control at childhood leads to more risk averse persons at an adult age. The influence of parental advice at childhood led to the weakest results in this study. Overall, the conclusion can be drawn that there is a relation between parental teaching and risk aversion. However, not all forms of parental teaching showed relations in the same direction. As such, this indicates that an adult’s risk aversion changes when received financial education from its parents at childhood. The kind of parental teaching the parent gives influences the risk aversion in a different manner.

This thesis contributes to behavioural finance literature by focusing on parental teaching, especially the three forms of parental teaching, in combination with risk aversion. To date parental teaching in combination with risk aversion has received little attention within behavioural finance research. Our analysis therefore suggests parental financial teaching received during childhood is important to stimulate financial literacy and knowledge of individuals. This could lead them to take better financial decisions and improved risk evaluation. This would lead households to making better decisions for themselves and their families. Furthermore, research suggests that a decreased risk aversion leads to larger wealth accumulation before retirement (Hariharan, Chapman & Domian, 2000; Yuh & DeVaney, 1996).

LIMITATIONS AND RECCOMENDATIONS FOR FUTURE RESEARCH

An limitation of this research is the dataset, it only included Dutch households. Caution should therefore be exercised in generalization of the results. Cultural differences could alter the results. Besides, the sample regarding the education level of the respondents is not even distributed. Only five percent of the sample received a low education. Therefore, the results regarding this control variable education can be flawed. Furthermore, on the basis of Bucciol & Veronesi (2013) three forms of parental teaching were chosen in this study. However, other parental teaching activities could influence the relationship with risk aversion. It may be helpful to be further investigated in future.

(33)

demonstrated. Future research could generate more knowledge on this matter as well as strengthen the results that were not strong enough to draw any conclusions from.

Future research is necessary to distinct the two types of risk aversion (safety and risk taking). This research suggests the questions asked in the direction of taking risks provide more significant results than questions in the direction of safety. No direct explanation of a link between these two types of risk aversion can be given in this study. Further investigation on these two concepts can create deeper insights. Furthermore, additional research is needed to investigate the effects of objective knowledge and subjective knowledge. In this study these effects have not been revealed. However, factual knowledge enhancing confidence with knowledge seems rather logical. Risk-taking self-confidence is important for people to even dare to take a risk, and is therefore important in future research on this matter.

(34)

REFERENCES

Agnew, J.R., Anderson, L.R., & Gerlach, J.R. (2008). Who chooses annuities? An experimental investigation of the role of gender, framing, and defaults. American Economic Review, 98(2) :418-422.

Ameriks, J., Caplin, A., & Leahy, J. (2003). Wealth accumulation and the propensity to plan. Quarterly Journal of Economics, 118 (3): 1007-1047.

Bayer, P.J., Bernheim, B.D., & Scholz, J.K. (2008). The effects of financial education in the workplace: evidence from a survey of employers. Economic Inquiry, 47(4): 605-624.

Benartzi, S., & Thaler, R.H. (2007). Heuristics and biases in retirement savings behavior. Journal of Economic Perspectives, 21(3): 81-104.

Benjamin, D., Brown, S., & Shapiro, J. (2005). Who is ‘Behavioral’? Unpublished.

Bernheim, B.D., & Garrett, D.M. (2003). The effects of financial education in the workplace: evidence from a survey of households. Journal of Public Economics, 87: 1487-1519.

Brooks, C. (2008). Introductory econometrics for finance. New York, NY: Cambridge University Press.

Bucciol, A., & Veronesi, M. (2013). Teaching children to save: What is the best strategy for lifetime savings? Network for Studies on Pensions, Aging and Retirement, DP 10/2013-051.

Carducci, B.J., & Wong, A.S. (1998). Type A and risk taking in every day money matters. Journal of Business and Psychology, 12(3): 355-359.

Dohmen, T., Falk, A., Huffman, D., & Sunde, U. (2010). Are risk aversion and impatience related to cognitive ability? American Economic Review, 100: 1238-1260.

Donkers, B., Melenberg, B., & van Soest, A. (2001). Estimating risk attitudes using lotteries: a large sample approach. Journal of Risk and Uncertainty, 22(2): 165-195

(35)

Frederick, S. (2005). Cognitive reflection and decision Making. Journal of Economic perspectives, 19(4): 25–42.

Gilliam, J. E., Chatterjee, S., & Zhu, D. (2010). Determinants of risk tolerance in the baby boomer cohort. Journal of Business & Economics Research, 8: 79-87.

Grable, J.E. (2000). Financial risk tolerance and additional factors that affect risk taking in everyday money matters. Journal of Business and Psychology, 14(4): 625-630.

Hallahan, T., Faff, R., & McKenzie, M. (2004). An empirical investigation of personal financial risk tolerance. Financial Services Review, 13: 57-78.

Hariharan, G., Chapman, K., & Domian, D. (2000). Risk tolerance and asset allocation for investors nearing retirement. Financial Services Review, 9: 159-170.

Jacob-Lawson, J.M., & Hershey, D.A. (2005). Influence of future time perspective, financial knowledge, and financial risk tolerance on retirement saving behaviors. Financial Services Review, 14: 331-344.

Jappelli, T., & Padula, M. (2013). Investment in financial literacy and saving decisions. Journal of Banking & Finance, 37: 2779-2792.

Lewis, A., & Scott, A. (2000). The economic awareness, knowledge and pocket money practices of a sample of UK Adolescents: a study of economic socialization and economic psychology. Children’s Social and Economic Education, 4: 34-46.

Morrongiello, B.A., Corbett, M., & Bellissimo, A. (2008). “Do as I say, Not as I do”: family influences on children’s safety and risk behaviors. Health Psychology, 27(4): 498-503.

Norvilitis, J.M., & MacLean, M.G. (2010). The role of parents in college students’ financial behaviors and attitudes. Journal of Economic Psychology, 31: 55-63.

Riley, W., & Chow, K.V. (1992). Asset allocation and individual risk aversion. Financial Analysts Journal, 48: 32–37.

(36)

Ryack, K. (2011). The impact of family relationships and financial education on financial risk tolerance. Financial Services Review, 20: 181-193.

Wang, A. (2009). Interplay of Investors’ Financial Knowledge and Risk Taking. The Journal of Behavioral Finance, 10: 204-213.

Webley, P., & Nyhus, E.K. (2006). Parents’ influence on children’s future orientation and saving. Journal of Economic Psychology, 27: 140-164.

(37)

THE INFLUENCE OF PARENTAL FINANCIAL TEACHING OF CHILDREN ON THEIR FUTURE RISK AVERSION

Master thesis Finance

By Daniëlle Æbelina van Dalfsen

THE INFLUENCE OF PARENTAL

FINANCIAL TEACHING OF

CHILDREN ON THEIR FUTURE

THE INFLUENCE OF PARENTAL FINANCIAL

TEACHING OF CHILDREN ON THEIR FUTURE

RISK AVERSION

Master thesis Finance

By Daniëlle Æbelina van Dalfsen

University of Groningen

Faculty of Economics and Business

June, 2014

Msc Finance

Supervisor: prof. dr. C.L.M. Hermes

Wielewaalplein 136

9713 BR Groningen

d.c.van.dalfsen@student.rug.nl

ABSTRACT

TABLE OF CONTENTS

1. INTRODUCTION

2. LITERATURE AND HYPOTHESIS DEVELOPMENT

3. METHODOLOGY

ε

4. EMPIRICAL RESULTS

Risk

aversion

factor 1

Risk aversion

factor 2

Advice

Age

Gender

Education

5. FURTHER ANALYSIS

6. DISCUSSION

7. CONCLUSION

REFERENCES

APPENDIX 1