FINANCIAL LITERACY AND ADVICE SEEKING;
COMPLEMENTS OR SUBSTITUTES?
T. KLEIN WASSINK
1869922
University of Groningen
Faculty of Economics and Business
Nettelbosje 2
9747 AE Groningen
Tel: 050-3634624
FINANCIAL LITERACY AND ADVICE SEEKING;
COMPLEMENTS OR SUBSTITUTES?
ABSTRACT
1. INTRODUCTION
Investment mistakes made by households has been the topic of many studies in the past. There are significant welfare cost that stem from several sources of inefficiency, including underdiversification and non-participation in risky asset markets (Calvet, Campbell and Sodini, 2007). There is no obvious reason why people make these sub-optimal investment decisions, but the growing literature on financial literacy suggests that consumers lack knowledge of even basic financial principles and products (Calcagno and Monticone, 2014).
The ability of individuals to make good financial decisions about managing income, financial products and their investments is receiving growing attention. Individuals are taking responsibility for an increasing number of financial decision even though financial literacy present in households nowadays might not be sufficient to make good financial decisions (Hung, Parker and Yoong, 2009). The recent financial crisis has shown the danger of a large number of individuals making poorly informed decisions about complex financial products (Schmeiser and Seligman, 2013). A lack of financial literacy can be, to some extent, compensated by acquiring financial advice. Financial advisers may benefit financial decision making in several ways. Collins (2012), for example, indicates that hiring an adviser may lower information acquisition costs compared to searching without an adviser through economies of scale. Also, financial advisers can identify and correct cognitive errors that may lead to sub-optimal investment decisions.
According to Grable and Joo (2001), not enough is known about why people seek advice and even less is known about the factors that influence the decision to seek advice from a professional adviser compared to a non-professional source of advice. They observe that little is known about who choose to rely on professional advisers other than that these people typically exhibit higher financial well-being. Existing literature suggests a relationship between financial literacy and advice seeking. How financial literacy and advice seeking relate to each other, however, has only recently become a topic of research and the results so far are quite limited and mixed. This paper aims to address this question.
invest in stocks. They argue that this is mainly because having financial literacy makes it easier to collect and process financial information.
Von Gaudecker (2014) studies the return loss that results from underdiversified portfolio’s and the role of financial literacy and advice seeking. His most significant finding is that the largest return losses from underdiversification are incurred by those who neither have good skills in basic financial concepts nor turn to external help with their investment decision i.e. seek advice. Acquiring professional advice can prevent underdiversification, Von Gaudecker (2014) however shows that not everyone with low financial literacy turns to a financial adviser.
It is clear that financial literacy and the use of financial advice both play an important role in the financial decisions made by households, but there are mixed results on how these two factors relate to each other. Robb, Babiarz and Woodvard (2012) find a positive relationship between financial literacy and advice seeking. Their results indicate that more knowledgeable individuals better understand the potential benefits of good financial advice. These individuals would be more likely to use advice because they are aware that this might help them avoid costs associated with poor financial decisions. There are also opposite results however. Finke, Huston and Winchester (2011), for example, find that greater financial knowledge is
negatively associated with the decision to pay for financial advice. As a possible explanation they state that a lack of a trust-based relationship between the client and adviser might give rise to agency costs. If people with high financial knowledge are better able to estimate these potential agency costs they will be less inclined to purchase financial advice.
This study aims to provide further insight into the relationship between financial literacy and advice seeking. Where most studies have only focused on professional advice, I will distinguish between professional and non-professional sources of advice (advice from parents, friends and acquaintances). To the best of my knowledge, this distinction has never been made before. It would be a valuable addition to the existing literature to also get some more insight into the relationship between financial literacy and non-professional advice seeking.
2. LITERATURE
2.1 Theory on financial advice seeking behavior
The field of behavioral finance has taught us that people don’t always act rationally and are bound to make, at least to some extent, investment mistakes. Examples of these investment mistakes are the disposition effect (holding on to losing stocks for too long) and the tendency to invest in local stocks (which may lead to underdiversification). Acquiring (professional) financial advice might avoid these investment mistakes and help individuals in making decisions that support their long-term financial security (Collins, 2012).
Bluethgen, Gintschel, Hackethal and Mueller (2008) believe that cognitive errors and costly information acquisition offer a reasonable basis for a theory of financial advice. According to them, individuals make mistakes the most in more complex situations because of cognitive limitations. In these cases, financial advisers can be valuable by helping investors to avoid such mistakes. The costly information explanation states that although people are deciding optimally based on the information available to them, they make sub-optimal choices because they lack information. Financial advisers can then add value by exploiting economies of scale in information acquisition (Bluethgen, Gintschel, Hackethal and Mueller, 2008). These economies of scale arise from the fact that an adviser can spread the costs of acquiring knowledge across multiple clients, who may only need the information once. This could be an indication that financial advice is most valuable to individuals who have high information acquisition costs and suffer from cognitive biases and errors the most.
Haslem (2008) states that in the light of the financial crisis, financial advisers can help clients avoid panicking and selling investments under distress, i.e. act on their emotions instead of rational thinking. According to Engelmann, Capra, Noussair and Bems (2009) during times of economic downturn people seek advice the most. Because of an increased perception of risk people may feel unfit to predict the consequences of their choices and may seek the advice of experts to reduce this perceived risk.
decisions. The probability of seeking advice outside one’s social network tends to increase as decisions are more complex or there is more need for specialized knowledge (Chang, 2005).
Professional financial advice is not the only source of advice people make use of. In addition to paying for financial advice, there is also the option to collect advice from one’s social network. Brown, Ivkovi, Smith and Weisbenner (2008) study what they call “community effects”, which occur when people rely on whatever information they have obtained through word-of-mouth communication. Financial decision making may be influenced by these social interactions. Brown, Ivkovi, Smith and Weisbenner (2008), for example, find a causal relationship between this word-of-mouth communication and stock market participation. One rationale behind relying on these social interactions is that the decisions of others may reflect information that they have and others do not. These peer effects may lead to people doing what others are doing rather than relying on their own information and knowledge (Banerjee, 1992).
It is questionable whether relying on non-professional sources of advice when making financial decisions is beneficial. According to the FPA (Financial Planning Association of Australia) this source of advice, although well-meant, is unlikely to provide the best investment strategy. Where advice provided by professional advisers is individually tailored to each client based on their specific information, advice from one’s social network is more likely to be based on their financial circumstances. Banerjee (1992) indicates that there is evidence that the average U.S. citizen has limited investment knowledge and that this might cause them to make financial decisions based on what they have learned from social interactions. This lack of financial knowledge is not restricted to the U.S., Van Rooij, Lusardi and Alessie (2011b) have observed that most Dutch households also lack knowledge of fundamental financial concepts. It would be interesting to see how this effect of relying on social interactions when making financial decisions relates to financial literacy. That is why, in addition to the relationship between financial literacy and professional advice seeking, I will study the relationship between financial literacy and non-professional advice seeking to see to what extend these relationships are similar or different.
2.2 Defining and measuring financial literacy
skills and financial attitudes (Hung, Parker and Yoong, 2009). I adopt the definition of Hung, Parker and Yoong (2009) who provide an overarching concept of financial literacy based on existing conceptual definitions. They define financial literacy as “the knowledge of basic economic and financial concepts, as well as the ability to use that knowledge and other financial skills to manage financial resources effectively for a lifetime of financial well-being”. This definition states that not only knowledge about financial matters is important, but also the ability to use that knowledge. Having financial knowledge on its own influences the ability to use financial knowledge, this ability is however also determined by other factors like social networks and access to resources (Hung, Parker and Yoong, 2009). Perhaps that is why most measures of financial literacy seem to focus more on the financial knowledge itself than the ability to use this knowledge.
Similar to the diversity in definitions of financial literacy, researchers have used a variety of methods to measure financial literacy. Hung, Parker and Yoong (2009) distinguish between performance tests (measured knowledge) and self-reported financial literacy (perceived knowledge). They state that since there is only a moderate correlation between actual and perceived knowledge, care should be taken when using self-reported financial literacy as a proxy for actual knowledge. Especially since research has shown that individuals think they know more than they actually do, i.e. are overconfident (Alba & Hutchinson, 2000), a finding that is generally most present amongst men (Barber & Odean, 2001). According to Hung, Parker and Yoong (2009) there is added value in looking at both self-perceived literacy and measured literacy. Given the fact that the correlation between the two is only moderate they state that perceived literacy may have predictive ability on its own as well as add predictive validity.
2.3 Theoretic link between financial literacy and professional advice seeking
Calcagno and Monticone (2014) and Collins (2012) find that indeed investors with high financial literacy are more likely to consult financial advisers than investors with low financial literacy. In the case of Calcagno and Monticone (2014) this result is conditional on whether people were investing in risky assets. This is important since my study is not limited to individuals who are investing in risky assets. I will, however, perform a regression only on this group of individuals as a robustness check.
Lusardi and Mitchell (2011) provide further indications that the relationship between financial literacy and professional advice seeking is positive. They find that respondents who scored highest on financial literacy were in general more likely to use formal retirement planning tools (retirement seminars, financial experts etc.) over informal tools (friends, acquaintances etc.). This finding already provides an indication that the source of advice matters and that there might be different results for professional and non-professional advice.
There is also a line of reasoning that supports the negative relationship between financial literacy and professional financial advice seeking. As mentioned previously, according to Robb, Babiarz and Woodvard (2012), there is a trade-off between individual decision making and professional advice. Following this line of reasoning, individual decision making will become costly as decisions become more complicated since this requires more financial knowledge. In this view, individuals who score high on financial literacy will be able to deal with these complex issues better and will have lower individual decision making costs. This means that for them the need for financial advice is lower and the trade-off between individual decision making and professional advice will tilt more towards individual decision making.
existing literature on financial literacy has shown that in general female and older individuals are also the least financially literate Viera (2012).
2.4 Theoretic link between financial literacy and non-professional advice seeking
Although, for instance, Bluethgen, Gintschel, Hackethal and Mueller (2008) find that in Germany more than 80% of investors make use of professional financial advisers, Grable, Cantrell and Maddux (2004) find that overall only 20% of households uses professional advice. Professional advice seems most attractive to people investing in risky assets, the majority of households, however, relies on non-professional sources of advice due to cost and trust factors. There should be a significant difference in the quality of advice between professional and non-professional advice. A professional financial advisers invests a lot of time to develop a thorough understanding of income levels, family situation and risk profile of each client, which is different for each person. Non-professional advice is less tailored to the specific situation of each person and thus will not be appropriate in all situations. In addition, non-professional advisers are unlikely to have the same amount of experience with and knowledge about financial markets and products. It is safe to say that people are aware of this difference in quality since professional advice comes at a price while non-professional advice comes at no cost, but is acquired through social networks and “community effects” (Brown, Ivkovi, Smith and Weisbenner (2008). The decision to seek advice may be influenced by demographic and socioeconomic elements, as well as financial knowledge and attitudes. Those who are more knowledgeable in matters of financial nature and exhibit better financial behaviors than others may feel less urge to seek advice from non-professional sources (Grable and Joo, 1999).
interactions. This would imply that people with higher financial literacy base their decisions less on these “community effects” and rely more on their own judgment.
3. DATA AND METHODOLOGY
3.1 Data on Dutch households
I use data from the DNB Household Survey (DHS), this survey is administered via the Internet by CentERdata and contains data on over 2000 Dutch households. It is a reflection of the Dutch speaking population and it allows researchers to study the psychological and economic aspects of financial behaviors. Any selection problems due to a lack of Internet access are resolved by equipping respondents without a computer with a box that enables them to respond via their television. DHS data contains detailed information on financial matters and has been used quite a lot in the past, for example by Van Rooij, Lusardi, and Alessie (2010, 2011a, 2011b) and von Gaudecker (2014). I use the DHS 2005 wave to get information on advice seeking behavior and all the other variables I need. Measured financial literacy information was also obtained in 2005 for 1508 individuals by Van Rooij, Lusardi, and Alessie. From the DHS 2005 wave, not all information was available for these 1508 individuals. After matching the financial literacy module and the DHS modules, 1115 individuals remain in the sample. From these 1115 individuals, only 1034 answered the questions on advice seeking behavior and self-perceived literacy. This means that measured financial literacy data on 1115 individuals and self-perceived data and advice seeking data on 1034 individuals remain. After including the control variables in the main regression the number of respondents included drops to 865. This is because there is some missing data on the control variables as well and this missing data is not the same for all variables.
3.2 Variable construction and data description
3.2.1 Financial advice
Table 1: Most important source of advice among respondents
Source of advice N
1. Parents, friends or acquaintances 2. Professional financial advisers 3. Information from the newspapers 4. Financial magazines, guides, books
5. Brochures from my bank or mortgage adviser.
6. Advertisements on TV, in the papers, or in other media 7. Financial computer programs
8. Financial information on the Internet 9. Other 239 277 95 102 90 33 8 104 86 Note: Total number of respondents is 1034
When people consult for example brochures or the Internet, it can be argued that they are actively looking for information to form their own opinion or increase their own knowledge base. Therefore, I split the answers up into three categories; professional advice, non-professional advice and relying on one’s own judgment. I create three dummy variables, one for professional advice (answer 2), non-professional advice (answer 1) and own judgment (all other answers). Table 2 shows the advice seeking behavior of the respondents across several demographics. There is no clear indication that education has an influence on advice seeking behavior, although people with a higher vocational or university degree rely slightly less on non-professional sources of advice and more on their own judgment.
Table 2: Advice seeking across demographics
Most important source of advice Education Professional (N=277) Non-professional (N=239) Own judgment (N=518) N < Intermediate vocational Intermediate vocational Secondary pre-university Higher vocational University 21.5% 23.8% 28.8% 27.9% 22.7% 24.8% 22.9% 21.8% 17.3% 19.9% 53.7% 53.3% 49.4% 54.8% 57.4% 288 199 154 265 128 Age 21-30 years 31-40 years 41-50 years 51-60 years 61-70 years 71 years and older
Looking at age, it seems that people that are between 31 and 50 years old rely the most on professional advisers. People up to 40 years old rely the most on non-professional advice and people older than 50 years rely most on their own judgment. Gender seems to have an influence on whether one relies on non-professional sources of advice or on their own judgment; women seem to rely more on non-professional sources of advice and men more on their own judgment.
3.2.2. Financial literacy
I measure financial literacy in three different ways; (measured) basic literacy, (measured) advanced literacy and self-perceived literacy. The measured financial literacy data consists of two modules. The first module aims to asses basic financial literacy and consists of five questions. These questions cover topics as interest compounding, inflation and the time value of money. The second module covers more advanced financial topics such as risk diversification and the difference between stocks and bonds. This set of questions is aimed towards measuring more advanced financial knowledge. Self-perceived literacy is measured by a single question that assesses how confident an individual is towards his financial knowledge. Households were instructed to answer the basic and advanced literacy questions without consulting any information sources or using a calculator. The full list of basic and advanced literacy questions can be found in appendix A.
Basic literacy
Table 3: Correct, incorrect and do not know answers for basic and advanced questions Panel A: Basic literacy – weighted percentages of total number of respondents (N=1115)
Question Correct Incorrect Do not know
(1) Numeracy 94.6% 4.0% 1.4%
(2) Interest compounding 81.5% 17.2% 1.3%
(3) Inflation 87.4% 8.7% 4.9%
(4) Time value of money 77.3% 20.6% 2.1%
(5) Money illusion 71.5% 26.3% 2.2%
Panel B: Advanced literacy – weighted percentages of total number of respondents (N=1115)
Question Correct Incorrect Do not know
(6) Function of stock market 70.9% 13.4% 15.7%
(7) Stock ownership 66.5% 24.6% 8.9% (8) Mutual funds 72.8% 10.6% 16.6% (9) Bond ownership 62.5% 16.2% 21.3% (10) Asset returns 53.0% 28.8% 18.2% (11) Risk of assets 72.5% 12.5% 15.0% (12) Risk diversification 68.4% 16.9% 14.7%
(13) Selling bonds before maturity 35.6% 28.3% 36.1%
(14) Stocks vs. bonds 39.5% 40.3% 20.2%
(15) Company stock vs. mutual fund 53.2% 25.3% 21.5%
(16) Bond pricing 33.9% 34.6% 31.5%
Advanced literacy
Table 4: Summary of responses
Panel A: Basic literacy – weighted percentages of total number of respondents (N=1115) Number of correct, incorrect and do not know answers (out of 5 questions)
None 1 2 3 4 All Mean
Correct 0.6% 1.9% 4.7% 14.5% 33.9% 44.4% 4.12
Incorrect 47.0% 35.6% 13.0% 3.4% 1.0% 0.0% 0.76
Do not know 92.7% 4.8% 1.2% 0.8% 0.4% 0.2% 0.12
Note: The responses may not sum up to 100% and the means not to the total number of questions due to rounding
Basic and advanced literacy indexes
With the financial literacy information obtained from the two modules (basic and advanced), I construct two separate financial literacy indexes. In a similar vein as Van Rooij, Lusardi, and Alessie (2011b), the first step is to perform a factor analysis on all 16 questions from both modules. The outcome of this factor analysis is consistent with the division of the financial literacy questions; it indicates that there are two main factors with different factor loadings. On the one hand there is a factor containing the 5 basis literacy questions, and on the other hand a factor containing the 11 advanced literacy questions. The next step is to perform a factor analysis on the two sets of questions separately. By doing this I can construct two separate indexes: a basic literacy index and an advanced literacy index. In line with Van Rooij, Lusardi, and Alessie (2011b), I take into account the differences between “incorrect” and “do not know” answers. According to them, in order to differentiate among degrees of financial knowledge it is important to exploit this information. I do this by including the “do not know” answers separately in the factor analysis, the factor loadings are reported in appendix B. This also means that in the measured financial literacy indexes, “do not know” answers are considered as incorrect. As a robustness check, I use another way to construct the literacy indexes in which one difference is that the “do not know” answers are treated as a random guess.
Basic and advanced literacy across demographics are reported in table 5. Education seems to be an important factor in how financially literate someone is. For both basic and advanced literacy, the mean literacy scores increase along with the education level. For basic literacy, it can be seen that people between the age of 31 and 60 years old score the highest. Panel B: Advanced literacy – weighted percentages of total number of respondents (N=1115)
Number of correct, incorrect and do not know answers (out of 11 questions)
None 1 2 3 4 5 6 7 8 9 10 All Mean
For advanced literacy this finding is different; the mean advanced literacy score increases with age. Gender differences can also be observed; the mean score on basic and advanced literacy is 4.26 and 2.85 respectively for men and 3.94 and 2.05 respectively for women indicating that in line with existing literature men are more financially literate than women.
Table 5: Basic and advanced literacy across demographics Panel A: Basic literacy across demographics
Number of correct basic literacy answers (out of 5 questions)
Education None 1 2 3 4 All Mean N
< Intermediate vocational Intermediate vocational Secondary pre-university Higher vocational University 1.6% 0.5% 0.0% 0.4% 0.0% 3.3% 3.3% 0.6% 1.1% 0.0% 8.8% 5.6% 2.9% 1.8% 2.1% 19.2% 17.3% 15.3% 12.4% 3.5% 36.8% 37.4% 31.2% 35.3% 22.7% 30.3% 36.0% 50.0% 49.1% 71.6% 3.77 3.96 4.27 4.29 4.64 307 214 170 283 141 Age 21-30 years 31-40 years 41-50 years 51-60 years 61-70 years 71 years and older
1.0% 0.5% 1.2% 0.0% 0.6% 0.7% 1.9% 3.5% 0.8% 1.2% 2.3% 2.2% 3.8% 4.0% 2.8% 4.9% 6.2% 7.2% 23.1% 11.4% 15.5% 13.5% 14.1% 13.0% 22.1% 31.8% 35.5% 36.9% 33.9% 37.7% 48.1% 48.8% 44.2% 43.4% 42.9% 39.1% 4.08 4.17 4.16 4.16 4.07 4.02 104 201 251 244 177 138 Gender Male Female 0.5% 0.8% 1.3% 2.7% 4.3% 5.1% 10.8% 19.3% 31.8% 36.6% 51.4% 35.4% 4.26 3.94 629 486
Note: The responses may not sum up to 100% and the means not to the total number of questions due to rounding Total number of respondents is 1115
Self-perceived literacy
Self-perceived literacy is measured by the question “How knowledgeable do you consider yourself with respect to financial matters?”. Respondents can choose one out of four Panel B: Advanced literacy across demographics
Advanced literacy quartiles
Education 1 (low) 2 3 4 (high) Mean N < Intermediate vocational Intermediate vocational Secondary pre-university Higher vocational University 40.4% 27.1% 20.6% 16.3% 10.6% 27.4% 29.0% 23.5% 24.0% 17.7% 18.6% 22.4% 30.6% 28.3% 29.8% 13.6% 21.5% 25.3% 31.4% 41.9% 2.05 2.38 2.61 2.75 3.03 307 214 170 283 141 Age 21-30 years 31-40 years 41-50 years 51-60 years 61-70 years 71 years and older
options; not knowledgeable, more or less knowledgeable, knowledgeable and very knowledgeable. Table 6 shows how many times each option was chosen and the overall self-perceived literacy across demographics.
Table 6: Self-perceived literacy across demographics
Knowledge of financial matters (1 = not knowledgeable, 4 = very knowledgeable)
Education 1 (N=144) 2 (N=621) 3 (N=238) 4 (N=31) Mean N < Intermediate vocational Intermediate vocational Secondary pre-university Higher vocational University 17.0% 11.1% 20.1% 9.1% 14.1% 61.8% 66.8% 53.9% 60.0% 53.1% 19.8% 21.6% 24.0% 25.7% 25.8% 1.4% 0.5% 2.0% 5.2% 7.0% 2.06 2.12 2.08 2.27 2.26 288 199 154 265 128 Age 21-30 years 31-40 years 41-50 years 51-60 years 61-70 years 71 years and older
8.5% 14.8% 13.7% 13.0% 15.6% 16.3% 59.6% 59.9% 64.2% 53.5% 61.1% 63.7% 27.7% 19.8% 19.5% 30.4% 21.6% 19.3% 4.2% 5.5% 2.6% 3.1% 1.7% 0.7% 2.28 2.16 2.11 2.24 2.09 2.04 94 182 226 230 167 135 Gender Male Female 11.8% 16.7% 55.9% 65.5% 27.7% 16.9% 4.6% 0.9% 2.25 2.02 585 449 Note: The responses may not sum up to 100% and the means not to the total number of questions due to rounding Note: Total number of respondents is 1034
Over 60% of the respondents answered that they are only more or less knowledgeable with respect to financial matters while only 3% of the respondents consider themselves very knowledgeable. For self-perceived literacy, just as for basic and advanced literacy, education seems to matter. People with at least a higher vocational degree have the most confidence in their knowledge of financial matters. When we look at age, no clear pattern is visible although it seems that people who are older than 60 have the least confidence in their knowledge of financial matters. The same gender difference as for basic and advanced literacy can be observed; men rate themselves higher on knowledge of financial matters. This difference, however, is smaller than for the basic and advanced literacy measures.
3.2.3 Control variables
Demographic factors influencing advice seeking
wealthy households might rely on more informal social networks (Chang, 2005). In addition to wealth, Bluethgen, Gintschel, Hackethal and Mueller (2008) find that older individuals and women are more likely to seek professional financial advice. The literature review by Viera (2012) on financial literacy confirms that female investors possess in general less financial literacy. When age is concerned, both younger and older investors possess less financial literacy (Van Rooij, Lusardi and Alessie, 2011b). In the case of non-professional advice seeking, the FPA found that young people are relying on financial advice from friends and family rather than seeking the services of a qualified independent financial adviser.
Finke, Huston and Winchester (2011) find that higher income levels makes it more effective to outsource some of the burden of financial planning to a financial professional. Higher levels of education is also associated with higher financial literacy (Viera, 2012). Being self-employed can also have an influence on the decision to pay for financial advice. Self-employed individuals have a greater expected benefit from making more tax efficient choices and improving portfolio composition by hiring financial professionals than those who are not self-employed (Finke, Huston and Winchester (2011).
Psychological factors influencing advice seeking
Table 7: Descriptive statistics of control variables
Control variable N Mean Median Min Max St. dev.
Age 1115 51.1 51 22 90 14.89 Education 1115 2.76 3.0 1.0 5.0 1.41 Male 1115 0.56 1.0 0.0 1.0 0.50 Wealth 965 165896 107411 -752920 2219358 228215 Number of children 1115 0.65 0.0 0.0 6.0 1.04 Retired 1115 0.22 0.0 0.0 1.0 0.41 Self-employed 1115 0.04 0.0 0.0 1.0 0.19 Married 1115 0.56 1.0 0.0 1.0 0.50 Household income 1115 3.54 4.0 1.0 5.0 1.05 Risk tolerance 1041 0.0 -0.05 -1.72 2.98 1.0 Financial satisfaction 1115 0.0 0.13 -2.72 3.01 1.0 3.3 Correlations
Table 8 shows the correlations between the advice seeking variables (professional advice, non-professional advice and own judgment) and financial literacy measures (basic, advanced and self-perceived). What stands out is that none of the literacy measures is significantly correlated to professional advice seeking. They are, however, all significantly correlated with non-professional advice seeking and relying on one’s own judgment. Hung, Parker and Yoong (2009) found that in earlier studies, correlations between self-perceived and measured financial literacy are ranging from .10 to .78 (median correlation was .49). The low correlations I find (.147 for basic and .256 for advanced literacy) may be an indication that in this study self-perceived literacy has some predictive ability on its own.
Table 8: Correlation table
Mean Median St.dev
Prof. advice Non prof. advice Own judgment Basic literacy Advanced literacy Self-perceived literacy Prof. advice .268 .000 .443 1.000 Non prof. advice .231 .000 .422 -.332 (.000)*** 1.000 Own judgment .501 1.000 .500 -.606 (.000)*** -.549 (.000)*** 1.000 Basic literacy .000 .325 1.00 .043 (.170) -.115 (.000)*** .059 (.056)* 1.000 Advanced literacy .000 .190 1.00 .028 (.369) -.206 (.000)*** .149 (.000)*** .427 (.000)*** 1.000 Self-perceived literacy 2.150 2.000 .683 -.015 (.623) -.175 (.000)*** .149 (.000)*** .147 (.000)*** .256 (.000)*** 1.000
Note: p-values are reported in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1
3.4 Methodology
The findings from existing literature result in the following hypotheses:
H1: Self-assessed, basic and advanced financial literacy are positively related to professional advice seeking
H2: Self-assessed, basic and advanced financial literacy are negatively related to non-professional advice seeking
H3: Self-assessed, basic and advanced financial literacy are positively related to relying on own judgment
These hypotheses will be tested using a linear probability model. Since the error terms cannot be assumed to be normally distributed, I use heteroscedasticity-robust standard errors. OLS assumption tests are reported in appendix D. The following equation represents the probability that someone relies on professional advice, non-professional advice or own judgment. This probability (P) is partially determined by the financial literacy measures and the control variables. The constant term is represented by βi and the disturbance term by ui.
P(yi = 1) = βi + Xi Financial literacy + Zi Control variables + ui
4. RESULTS
4.1 Main regressions
Table 9: multivariate analysis of advice seeking
(1) Control variables only (2) Including literacy measures
Professional advice Non-professional advice Own judgment Professional advice Non-professional advice Own judgment Basic literacy Advanced literacy Self-assessed literacy .008 (.663) -.006 (.751) -.084 (.383) -.014 (.383) -.047 (.006)*** -.270 (.002)*** .006 (.753) .053 (.011)** .355 (.001)***
Age dummies (base group: age ≤ 30) • 30 < age ≤ 40 • 40 < age ≤ 50 • 50 < age ≤ 60 • 60 < age ≤ 70 • age > 70 .092 (.163) .091 (.171) -.030 (.655) .047 (.544) .055 (.529) .010 (.860) -.199 .(001)*** -.218 (.000)*** -.246 (.000)*** -.291 (.000)*** -.102 (.154) .109 (.132) .247 (.001)*** .199 (.020)** .236 (.014)** .090 (.170) .088 (.184) -.031 (.641) .047 (.544) .056 (.522) .011 (.855) -.204 (.001)*** -.219 (.000)*** -.245 (.000)*** -.296 (.000)*** -.101 (.158) .116 (.109) .250 (.001)*** .197 (.021)** .240 (.012)**
Education dummies (base group: < int. vocational)
• Intermediate vocational • Secondary pre-university • Higher vocational • University .005 (.913) .088 (.075)* .059 (.168) -.022 (.693) -.066 (.113) -.081 (.067)* -.053 (.164) -.031 (.529) .061 (.230) -.007 (.899) -.006 (.905) .053 (.380) .006 (.904) .088 (.086)* .058 (.180) -.023 (.692) -.058 (.161) -.048 (.295) -.032 (.406) .005 (.926) .052 (.299) -.040 (.469) -.026 (.580) .018 (.773) Financial satisfaction Risk tolerance Male Married Number of children Retired Self-employed Household income -.015 (.404) .008 (.637) .002 (.952) .031 (.393) .029 (.124) -.088 (.142) .095 (.296) .006 (.783) -.014 (.376) -.028 (.054)* -.068 (.029)** -.053 (.103) -.016 (.344) .042 (.431) -.025 (.763) -.034 (.079)* .029 (.136) .020 (.249) .065 (.082)* .022 (.580) -.013 (.525) .046 (.484) -.070 (.478) .028 (.234) -.015 (.411) .009 (.599) .003 (.936) .030 (.403) .029 (.121) -.088 (.144) .096 (.291) .006 (.789) -.007 (.658) -.025 (.084)* -.041 (.195) -.053 (.103) -.017 (.311) .043 (.425) -.004 (.956) -.033 (.089)* .022 (.263) .017 (.351) .038 (.323) .022 (.571) -.012 (.550) .045 (.488) -.092 (.354) .027 (.251)
Wealth dummies (base group: wealth < 10.000) • 10.000 < wealth ≤ 50.000 • 50.000 < wealth ≤ 150.000 • 150.000 < wealth ≤ 375.000 • wealth > 375.000 Constant R-squared Observations .058 (.266) .108 (.029)** .092 (.061)* .151 (.020)** .116 (.299) .051 865 -.097 (.038)** -.097 (.029)** -.074 (.095)* -.067 (.247) .903 (.000)*** .151 865 .039 (.492) -.012 (.830) -.019 (.728) -.083 (.237) -.020 (.871) .120 865 .057 (.277) .109 (.029)** .093 (.060)* .151 (.021)** .117 (.306) .051 865 -.102 (.028)** -.080 (.071)* -.058 (.189) -.046 (.430) .843 (.000)*** .162 865 .045 (.423) -.029 (.593) -.035 (.514) -.105 (.139) .040 (.745) .128 865 Note: p-values are reported in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1
basic literacy, in contrast to the other literacy measures, does not yield any significant results might be an indication that the questions are too basic and are more of a cognitive ability measure.
4.2 Robustness checks
In this section I will perform additional regressions to see if (1) a different method of constructing the basic and advanced literacy indexes and (2) including only risky-asset holders yields different results as the main regression. Table 10 shows the results of both robustness checks, the control variables are included in the regressions but are not listed for reasons of space and relevance.
Table 10: multivariate analysis of advice seeking with different specifications
(1) Different literacy measures (2) Risky asset holders only
Professional advice Non-professional advice Own judgment Professional advice Non-professional advice Own judgment Basic literacy Advanced literacy Self-assessed literacy .026 (.761) -.020 (.773) -.088 (.366) -.118 (.121) -.163 (.010)** -.263 (.002)*** .092 (.321) .184 (.018)** .351 (.001)*** .048 (.272) .004 (.945) -.447 (.013)** -.002 (.954) -.077 (.039)** -.123 (.344) -.047 (.324) .074 (.179) .570 (.003)*** Controls included Age dummies Education dummies Financial satisfaction Risk tolerance Male Married Number of children Retired Self-employed Household income Wealth dummies Constant R-squared Observations Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes .088 (.478) .051 865 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes 1.036 (.000)*** .163 865 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes -.124 (.359) .128 865 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes .042 (.272) .100 268 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes .667 (.134) .157 268 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes .290 (.657) .165 268 Note: p-values are reported in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1
This means that if there are for example four answer possibilities, someone who answered “do not know” will receive 0.25 points (out of 1) on that question. The results are similar to the main regression, but there are some differences. The relationship between basic literacy and non-professional advice seeking (p-value is .121 versus .383) and relying on own judgment (p-value is .321 versus.753) becomes more significant. The relationship between advanced literacy and all forms of advice seeking, on the other hand, becomes slightly less significant (although still significant for non-professional advice seeking and relying on own judgment).
In the second regression I only include respondents who have money invested in growth funds, mutual funds, bonds and/or stocks. Respondents are included when they have money invested in one or more of these risky asset classes. Because of this threshold, the number of observations included in the regression drops to 268. The results from the regression on risky asset holders only are in some ways different from the main regression. When we focus on professional advice seeking, self-assessed literacy is in this case negatively related to professional advice seeking (p-value is .013). For non-professional advice seeking, advanced literacy becomes slightly less significant (p-value is .039 versus .006) and self-assessed literacy becomes non-significant. Advanced literacy was positively related to relying on own judgment in the main regression, in this regression however this relationship is not significant. Comparing again the different literacy measures, it can be seen that when I include only the risky asset holders the advanced and self-assessed literacy measures yield significantly different results. This provides an indication that the way in which financial literacy is measured does matter and has an impact on the results.
Performing two separate regressions on these subsamples yields the following results. For basic literacy, the relationship with non-professional advice seeking when having acquaintances with a higher education has a coefficient of -.008 (p-value is .745) in comparison to -.018 (p-value is .426) for acquaintances with the same or a lower education. For advanced literacy, the relationship with non-professional advice seeking when having acquaintances with a higher education has a coefficient of -.029 (p-value is .275) in comparison to -.057 (p-value is .011) for acquaintances with the same or a lower education. For self-assessed literacy, the relationship with non-professional advice seeking when having acquaintances with a higher education has a coefficient of -.268 (p-value is .075) in comparison to -.269 (p-value is .012) for acquaintances with the same or a lower education. These results show that the education level of acquaintances does matter; the relationship between financial literacy and non-professional advice seeking is stronger when acquaintances have the same or a lower education. The main regression indicated that people with high financial literacy have a lower probability of relying on non-professional advice. This result, however, becomes weaker when someone has acquaintances that are higher educated than he/she is.
4.4 Overconfidence or underconfidence?
5. CONCLUSION
5.1 Summary of main results
In this study, I analyze the relationship between financial literacy and different forms of financial advice seeking. The data set allows me to construct three different measures of financial literacy; basic financial literacy, advanced financial literacy and self-perceived financial literacy. The empirical findings show that both advanced and self-perceived literacy are negatively associated with relying on non-professional advice. Furthermore, they are both positively related to relying on own judgment. Wealth, as expected, is positively related to relying on professional advice. In addition, it is negatively related to relying on non-professional advice. The robustness checks show that a different manner in which the measured financial literacy indexes are constructed does not yield significantly different results. Including only asset holders, however, does have an impact. When only risky-asset holders are included, self-perceived literacy becomes negatively related to professional advice seeking. The association with non-professional advice seeking on the other hand becomes non-significant. For advanced literacy, the association with relying on own judgment becomes non-significant. The reduced number of respondents might be a determining factor in this case.
For relying on non-professional advice, the education level of acquaintances seems to matter. When someone has acquaintances that are higher educated than he/she this seems to weaken the negative association with advanced literacy. Furthermore, I show that in my sample there is an equal amount of underconfidence and overconfidence amongst women. Amongst men however, there is a clear indication of underconfidence, a finding that contradicts existing literature.
5.2 Limitations and future directions
only state their most important source of advice and not all forms of advice they acquire. I believe, however, that when people pay for professional advice they would consider that their most important source of advice, else why pay for it? An important limitation is that my results are not robust to endogeneity. This means that the results do not necessarily provide indications on the direction of the causality. For the results I find, however, it seems unlikely that for example relying on non-professional sources of advice will lower someone’s financial literacy. It seems more likely that someone who is financially literate will rely less on non-professional sources of advice, but this can’t be stated with any certainty.
Future studies on the topic might focus more on non-professional sources of advice. Little is currently known about the quality of financial decisions people make when relying on non-professional advice versus professional advice. What might also be an interesting question is whether the financial crisis has changed the way in which people view professional financial advisers. It could, for instance, be the case that the financial crisis has shown the necessity to acquire financial advice. The opposite could also be the case, financial advisers might have lost credit because most of them did not see the financial crisis coming and were unable to prevent their clients from losing money.
5.4 Conclusion
REFERENCES
Alba, J.W. & Hutchinson, J.W. 2000. Knowledge Calibration: What Consumers
Know and What They Think They Know. Journal of Consumer Research, vol. 27: 123-156.
Banerjee, A.V. 1992. A Simple Model of Herd Behavior. Quarterly Journal of Economics, vol. 107 (3): 797-817
Barber, B.M. & Odean, T. 2001. Boys Will be Boys: Gender, Overconfidence, and Common Stock Investment. Quarterly Journal of Economics, vol. 116 (1): 262-292
Bluethgen, R., Gintschel, A., Hackethal, A., Mueller, A. 2008. Financial advice and Individual Investors’ Portfolios. Working Paper Series. Social Science Research Network, (acquired at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=968197)
Brown, J.R., Ivkovi, Z., Smith, P.A., Weisbenner, S. 2008. Neighbors Matter: Causal Community Effects and Stock Market Participation. Journal of Finance, vol. 63 (3): 1509-1531
Calcagno, R. & Monticone, C. 2014. Financial Literacy and the Demand for Financial Advice. Journal of Banking & Finance (forthcoming)
Calvet, L.E., Campbell, J.Y., Sodini, P. 2007. Down or Out: Assessing the Welfare Costs of Household Investment Mistakes. Journal of Political Economy, vol. 115 (5): 707-747
Chang, M. (2005). With a little help from my friends (and my financial planner). Social Forces, vol. 83(4): 1469-1497.
Collins, M.J. 2012. Financial Advice: A Substitute for Financial Literacy? Financial Services
Review, vol. 21 (4): 307-322
Finke, M.S., Huston, S.J., Winchester, D.D. 2011. Financial Advice: Who Pays? Journal of
Financial Counseling & Planning, vol. 22 (1): 18-26
Georgarakos, D. & Pasini, G., 2011. Trust, Sociability, and Stock Market Participation.
Review of Finance. 15 (4): 693-725.
Grable, J.E. & Joo, S. 1999. Financial Help-seeking Behavior: Theory and Implications.
Financial Counseling and Planning, vol. 10 (1): 13-24.
Grable, J., Cantrell, J., Maddux, E. 2004. Seeking Financial Help from Friends and Relatives.
Proceedings of the Association for Financial Counseling and Planning Education, Denver.
Grable, J.E. & Joo, S. 2001. A Further Examination of Financial Help-Seeking Behavior.
Financial Counseling and Planning, vol. 12 (1): 55-74
Haslem, J.A. 2008. Why do Mutual Fund Investors Employ Financial Advisers? Journal of
Investing, vol. 17: 91-94
Hung, A.A., Parker, A.M., Yoong, J.K. 2009. Defining and Measuring Financial Literacy. Working Paper WR-708, RAND Labor and Population
Hung, A.A., Yoong, J.K. 2010. Asking for Help – Survey and Experimental Evidence on Financial Advice and Behavior Change. Working Paper WR-714-1, RAND Labor and
Population
Lusardi, A., Mitchell, O., 2011. Financial Literacy and Planning: Implications for
Retirement Wellbeing. In: Lusardi, A., Mitchell, O. (Eds.), Financial Literacy: Implications for Retirement Security and the Financial Marketplace. Oxford University Press, Oxford.
Robb, C.A., Babiarz, P., Woodvard, A. 2012. The Demand for Financial Professionals’ Advice: The Role of Financial Knowledge, Satisfaction and Confidence. Financial Services
Schmeiser, M.D. & Seligman, J.S. 2013. Using the Right Yardstick: Assessing Financial Literacy by Way of Financial Well-Being. Journal of Consumer Affairs, vol. 47 (2): 243-262
Van Rooij, M.C.J., Lusardi, A., Alessie, R.J.M. 2011a. Financial Literacy and Retirement Planning in the Netherlands. Journal of Economic Psychology, vol. 32 (4): 593-608
Van Rooij, M.C.J., Lusardi, A., Alessie, R.J.M. 2011b. Financial Literacy and Stock Market Participation. Journal of Financial Economics, vol. 101 (2): 449-472
Viera, A. 2012. What do we know about Financial Literacy? A Literature Review. Marmara
Journal of European Studies. vol. 20 (2): 23-38
APPENDIX
Appendix A: Basic and advanced literacy questions
Table A1 and A2 list the basic and advanced literacy questions as provided to the respondents.
Table A1: Basic literacy questions
(1) Numeracy Suppose you had €100 in a savings account and the interest rate was 2% per year. After 5 years, how much do you think you would have in the account if you left the money grow? (1) More than €102; (2) Exactly €102; (3) Less than €102; (4) Do not know; (5) Refusal (2) Interest
compounding
Suppose you had €100 in a savings account and the interest rate is 20% per year and you never withdraw money or interest payments. After 5 years, how much would you have on this account in total?
(1) More than €200; (2) Exactly €200; (3) Less than €200; (4) Do not know; (5) Refusal (3) Inflation Imagine that the interest rate on your savings account was 1% per year and inflation was 2%
per year. After 1 year, how much would you be able to buy with the money in this account? (1) More than today; (2) Exactly the same; (3) Less than today; (4) Do not know; (5) Refusal (4) Time value
of money
Assume a friend inherits €10.000 today and his sibling inherits €10.000 3 years from now. Who is richer because of the inheritance?
(1) My friend; (2) His sibling; (3) They are equally rich; (4) Do not know; (5) Refusal (5) Money
illusion
Suppose that in the year 2010 your income has doubled and prices of all goods have doubled too. In 2010, how much will you be able to buy with your income?
(1) More than today; (2) Exactly the same; (3) Less than today; (4) Do not know; (5) Refusal
Table A2: Advanced literacy questions
(6) Function of stock market
Which of the following statements describes best the main function of the stock market? (1) Stock market helps to predict stock earnings; (2) Stock market results in an increase in the price of stocks; (3) Stock market brings people who want to buy stocks together with people who want to sell stocks; (4) None of the above (5) Do not know; (6) Refusal
(7) Stock ownership
Which of the following statements is correct? If somebody buys the stock of firm B in the stock market:
(1) He owns a part of firm B; (2) He has lent money to firm B; (3) He is liable for firm B’s debt; (4) None of the above (5) Do not know; (6) Refusal
(8) Mutual funds
Which of the following statements is correct?
(1) Once one invests in a mutual fund, one cannot withdraw the money in the first year; (2) Mutual funds can invest in several assets, for example in stocks and bonds; (3) Mutual funds pay a guaranteed rate of return which depend on past performance; (4) None of the above (5) Do not know; (6) Refusal
(9) Bond ownership
Which of the following statements is correct? If somebody buys a bond of firm B: (1) He owns a part of firm B; (2) He has lent money to firm B; (3) He is liable for firm B’s debt; (4) None of the above (5) Do not know; (6) Refusal
(10) Asset returns
Considering a long time period (10 or 20 years for example), which asset normally gives the highest return?
(11) Risk of assets
Normally, which asset displays the highest fluctuation over time? (1) Savings accounts; (2) Bonds; (3) Stocks; (4) Do not know; (5) Refusal (12) Risk
Diversification
When an investor spreads his money among different assets, the risk of losing money: (1) increases; (2) decreases; (3) Stays the same; (4) Do not know; (5) Refusal (13) Selling
bonds before maturity
If you buy a 10-year bond, it means you cannot sell it after 5 years without incurring a major penalty. True or false?
(1) True; (2) False; (3) Do not know; (4) Refusal (14) Stocks
vs. Bonds
Stocks are normally riskier than bonds. True or false? (1) True; (2) False; (3) Do not know; (4) Refusal (15) Company
stock vs. mutual fund
Buying a company stock usually provides a safer return than a stock mutual fund. True or false?
(1) True; (2) False; (3) Do not know; (4) Refusal (16) Bond
pricing
If the interest rate falls, what should happen to bond prices?
(1) Rise; (2) Fall; (3) Stay the same; (4) None of the above (5) Do not know; (6) Refusal
Appendix B: Construction of financial literacy indexes
Table B1 shows the result of the factor analysis on all 16 basic and advanced literacy questions. Two separate factors can be distinguished; basic and advanced literacy.
Table B1: Factor analysis on both financial literacy modules
Factor
Question 1 2
(1) Numeracy .547 -.077
(2) Interest compounding .461 -.044
(3) Inflation .338 .252
(4) Time value of money .283 .173
(5) Money illusion .150 .049
(9) Bond ownership .041 .575
(14) Stocks vs. Bonds .096 .565
(8) Mutual funds .169 .543
(6) Function of stock market .203 .508
(16) Bond pricing .023 .476
(11) Risk of assets -.134 .452
(13) Selling bonds before maturity .193 .449
(7) Stock ownership .004 .437
(15) Company stock vs. mutual fund .292 .436
(10) Asset returns -.082 .384
(12) Risk Diversification -.021 .239
Table B2: Factor analysis on basic literacy questions
Question Factor loading
(1) Numeracy .686
(3) Inflation .623
(4) Time value of money .583
(2) Interest compounding .575
(5) Money illusion .293
In constructing the advanced literacy index, the “do not know” answered are included separately in the factor analysis. Table B2 reports the factor loadings for the advanced literacy questions. Again, these factor loadings are then used to construct the advanced literacy index.
Table B3: Factor analysis on advanced literacy questions
Question Factor loading
(6) Function of stock market Correct .597
Do not know -.727
(7) Stock ownership Correct .440
Do not know -.648
(8) Mutual funds Correct .650
Do not know -.779
(9) Bond ownership Correct .613
Do not know -.746
(10) Asset returns Correct .378
Do not know -.708
(11) Risk of assets Correct .448
Do not know -.756
(12) Risk Diversification Correct .297
Do not know -.740
(13) Selling bonds before maturity Correct .526
Do not know -.747
(14) Stocks vs. Bonds Correct .648
Do not know -.788
(15) Company stock vs. mutual fund Correct .567
Do not know -.751
(16) Bond pricing Correct .504
Appendix C: Control variables
Demographic control variables
- Age; several dummies are constructed. The base group is age < 30, the dummies are 30 < age ≤ 40, 40 < age ≤ 50, 50 < age ≤ 60, 60 < age ≤ 70 and age > 70
- Gender; dummy coded, 0 = female, 1 = male
- Married (yes/no); dummy coded, 0 = not married, 1 = married - Number of children; ranging from 0 to 6
- Retired (yes/no); dummy coded, 0 = not retired, 1 = retired
- Self-employed (yes/no); dummy coded, 0 = not self-employed, 1= self-employed
- Household income; categorical variable, defined as last year’s total household income. 1 = < 14000, 2 = 14000 – 22000, 3 = 22000 – 40000, 4 = 40000 – 75000, 5 = > 75000
- Education; defined as highest level of education completed. Several dummies are constructed. The base group is education < intermediate vocational, the dummies are intermediate vocational, secondary pre-university, higher vocational and university.
- Wealth; defined as total wealth of the household. This was calculated using aggregate data on the households including 26 asset and 11 liability classes. Here also several dummies are constructed. The base group is wealth < 10.000, the dummies are 10.000 < wealth ≤ 50.000, 50.000 < wealth ≤ 150.000, 150.000 < wealth ≤ 375.000 and wealth > 375.000.
Psychological control variables
The psychological control variables are constructed by performing a factor analysis on several questions that measure (1) attitudes towards risk and (2) financial satisfaction,
measured by comparing one’s financial situation to others. The questions and factor loadings are listed below.
Risk-tolerance factor
component analysis on the questions. Text box C1 shows the questions, the factor loadings are listed next to the questions. The factor has a Cronbach’s alpha of 0.657.
Box C1: Factor loadings risk-tolerance factor
Financial satisfaction factor
The financial satisfaction factor is composed of 6 questions (see table below) that measure a person’s financial situation in comparison to others. A person should be more satisfied with its current financial situation when he states that he is better off than others around him.
Box C2: Factor loadings financial satisfaction factor
The questions are answered on a 7-point Likert scale ranging from strongly disagree to strongly agree. As can be observed, question 3 had to be recoded. The factor was composed by performing a principle component analysis on the questions. Text box C2 shows the questions, the factor loadings are listed next to the questions. The factor has a Cronbach’s alpha of 0.815.
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 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 0.580 0.674 0.495 0.574 0,564 0.783
1. Compared to others in my environment, I am better off 2. I think I have more assets than others in my environment
Appendix D: OLS assumptions
As previously mentioned, when using the linear probability model the assumption of normal distribution of error terms will not hold. To correct for this I have used heteroscedasticity-robust standard errors. In addition to heteroscedasticity, there are also other assumptions of OLS that can be violated. Below I list the results of autocorrelation and multicollinearity tests.
Autocorrelation
To statistically test for autocorrelation, I use the Durbin-Watson test. The null-hypothesis that there is no autocorrelation should be rejected when the value of the Durbin-Watson test is in the rejection region (upper or lower critical level). The outcome of the test is a value of 2.0 exactly. This means that there is no evidence of autocorrelation and I cannot reject the null-hypothesis.
Multicollinearity
To test for multicollinearity I use collineary statistics that show tolerance and the variance inflation factor (VIF). Rule of thumb is that a tolerance value below 0.1 might be an indication of multicollinearity. As can be seen in table D1, no independent variable falls below that threshold.
Table D1: Multicollinearity statistics
Variable Tolerance VIF
Basic literacy Advanced literacy Self-assessed literacy
Age dummies (base group: age ≤ 30) • 30 < age ≤ 40
• 40 < age ≤ 50 • 50 < age ≤ 60 • 60 < age ≤ 70 • age > 70
Married
Number of children Retired
Self-employed Household income
