s2542676
Nuria Dominguez Enfedaque MSc Finance
Prof. Dr Auke Plantinga
Optimism and household financial decisions
Final VersionAbstract
In this study we follow Puri and Robinson (2007) paper and investigate whether optimism influences financial decisions. We construct three measures of optimism using data gathered from the Dutch Household Survey from 2003 until 2013. After testing the hypotheses using cross-sectional panel data model we find that households’ financial decision-making is related to optimism as optimistic households choose riskier portfolios. Further, moderate and extreme optimists both hold a higher amount of equity as a fraction of their total assets, whereas extreme optimists have taken higher risks in the past years.
1. Introduction
Traditionally, finance theory assumes that individuals are rational and make unbiased and consistent optimal investment decisions without being influenced by emotions (Ackert and Deaves, 2010; Byrne and Utkus, 2013). However, there is empirical evidence revealing the imperfection of financial markets (Byrne and Utkus, 2013). Behavioral finance attributes this imperfection to psychological biases, which impede individuals from making rational financial decisions (Shefrin, 2005; Ackert and Deaves, 2010). For example, mental accounting and regret aversion lead to sell winning investments too soon and hold losing ones too long (Odean, 1998).
Several studies reveal the effects of being optimistic in sciences and psychology. Empirical evidence shows that in general “people tend to be unrealistically optimistic” (Weinstein, 1980; Sharot, 2011; Ackert and Deaves, 2010). They tend to underestimate the chances of having a cancer, getting fired, or getting a divorce, and they tend to overestimate their life expectancy and expect their children to achieve more than other kids (Sharot, 2011; Weinstein, 1980; Ariely, 2009).
The literature distinguishes two main types of optimism. On one hand, dispositional optimism refers to generalized expectancies that positive outcomes have a higher probability of occurrence than negative or undesirable ones (Scheier and Carver, 1985; Puri and Robinson, 2007). On the other hand, optimism can vary from one context to another, and it is known as optimism bias (Weinstein, 1980). For instance, students expect to get higher scores on future exams than they will actually get (Buehler et al., 1994). In the case of medicine, evidences show that optimistic patients suffering from cancer have more chances to recover from it than pessimistic ones (Schulz et al., 1996). Also, optimists have a better capacity to overcome stressful situations in life (Friedman et al., 1995; Puri and Robinson, 2007).
Optimism is also the cause of some economic and financial events. Personality traits, such as overconfidence and optimism, have been recognized as driving financial behavior (Chen and Lin, 2012; Barber and Odean, 2001; Puri and Robinson, 2007). The literature mainly focuses on the influence of optimism and overconfidence on individual and corporate financial decisions (Miller, 1977; Brunnermeier and Parker, 2005; Heaton, 2002). For example, optimistic people underestimate risk and overestimate the likelihood of success, which can be helpful for society as human optimism drives economy cycles (Keynes, 1936). Moreover, we know that optimistic people are more eager to work to fulfill their goals (Kenny, 2015; Puri and Robinson, 2007); and overconfidence and optimism have a strong influence on corporate managers and investors, which in turn affect stock prices (Barberis et al., 1998; Barber and Odean, 2001).
and banks, which were optimistic about the future financial situation, provided enormous quantities of loans, which ended-up in tremendous losses in 2008 (Stockman, 2009). Moreover, Barber and Odean (2000) find that overconfident investors are very optimistic about the future stock returns and trade in excess. They find that the 20% of the most active investors have a yearly turnover of 250%.
Puri and Robinson’s paper (2007) is the most important study of the influence of optimism in household portfolio choice. These authors use life expectancy miscalibration to measure optimism. In this paper we follow Puri and Robinson (2007) and examine whether optimism influences households financial decisions.
For our study, we construct three measures of optimism instead of one in order to capture financial optimism better since life expectancy beliefs might be unrelated to the financial situation (Balasuriya et al., 2010). We first use the Dutch Household Survey of Consumer Finance to calculate the three measures of optimism, which are the following: financial expectation, financial situation compared to others of the household environment, and life expectancy miscalibration. Secondly, we wonder and check whether optimism is related to riskier portfolio choices using panel data methodology. Afterwards, we differentiate between extreme optimistic individuals and moderate ones in order to check whether the first ones take higher financial risks. Finally, we run some robustness tests in order to check the validity of the results.
2. Literature Review 2.1 Optimism
Scheier et al. (1994) highlight the power of dispositional optimism in medicine and psychology. They show that optimistic people, who face difficulties persevere in pursuing their goals by breaking bad habits, so their chances of achieving the desired results increase. In order to evaluate dispositional optimism those authors create a Life Orientation Test (LOT), which measures optimism directly by asking people whether they expect more good or bad things to happen to them in their future lives. This questionnaire consists of a variety of statements related to optimism such as “Overall, I expect more good things to happen to me than bad”, “In uncertain times, I usually expect the best”. Participants are asked to rate each statement using a scale from 0 to 4, where 0 means “strongly disagree” and 4 means “strongly agree”. They find that optimism is related to healthier life and optimistic people recover faster after medical interventions for heart disease or cancer.
Inventory (MBHU) that assesses attitudes toward stress and interpersonal traits such as confidence and pessimism.
Additionally, optimistic people are associated with better sport and professional performance; i.e. success in diverse occupations such as sales, they work more hours, and are more successful in relationships and they are less likely to divorce (Seligman, 2000; Gillham, 2000; Puri and Robinson, 2007). Optimistic individuals are also more likely to better overcome problematic or stressful situations as they are goal committed, which requires reaching the desire result (Carr, 2011; Zang et al., 2007), and have better health habits (Sheier and Carver, 1992).
But is optimism always advantageous? If optimistic individuals have positive anticipations about the future they do not worry about the negative consequences of taking risky decisions (Seligman, 2000). Hence, some authors have pointed out that a possible drawback to optimism is a greater tendency to exhibit riskier behavior and take less preventive measures (Tennen and Affleck, 1987; Weinstein, 1980). In the next section we discuss the negative aspects of optimism.
The optimism bias refers to the overestimation of the odds that a positive outcome will happen, and/or underestimating the chances a negative outcome will occur (Seligman, 2000). Optimistic individuals tend to overestimate predictions for success and underestimate the likelihood of negative events happening to them (Helweg-Larsen and Shepperd, 2001; Ackert and Deaves, 2010).
The most widely method used in detecting optimism bias entails having people evaluate their chances of living an experience compared to other colleagues, who have similar individual characteristics such as gender or age (Helweg-Larsen and Shepperd, 2001). There are two ways of analyzing these estimates: directly or indirectly (Weinstein, 1980; Scheier et al., 1994).
Measuring optimism bias directly involves an individual estimating his or her likelihood of achieving bad and good things relative to other peers in the same way as the LOT questionnaire explained earlier (Weinstein, 1980). On the other hand, in order to measure optimism bias indirectly, the participant estimates his or her own chances of experiencing a future event compared to real tables expectancies, i.e. their estimates of their life expectancies compared to the life tables (Puri and Robinson, 2007).
Weinstein (1980), pioneer of the optimism bias literature research, investigates the individual tendency to overestimate future life events. In his study he measures optimism bias directly by using a sample of 258 college students who estimate their chances of experiencing 42 situations in comparison to their peers’ chances of experiencing the same situations. By using multiple regression analysis he finds that most college students think they are less likely than their peers to experience a negative event such as a divorce or contracting a disease, and their chances of positive events are higher than of the other students; e.g. they believe they have more chances to live longer than their colleagues and that their chances of getting a job offer before graduation are higher than those of their peers.
being infected by AIDS or suffering from cancer, or of being part of a car accident, and overestimate their own abilities for success (Nofsinger, 2005). This can be dangerous since it makes people careless from taking preventive measures, and increases their chances of experiencing worse outcomes in the future than those who acknowledge risk (McKenna, 1993; Weinstein and Klein, 1996). Further, Chesterman et al. (1990) find that optimistic women suffer from more birth complications than older pregnant women.
Optimism bias effects go beyond psychology and medicine, as financial decisions may be affected by optimism (Barber and Odean, 2001; Puri and Robinson, 2007). In the next section we discuss the effects of optimism in finance.
2.2 Optimism and corporate finance
The efficient market hypothesis (EMH) is based on the assumption of perfect markets and rational individuals (Fama and French, 1972). This traditional theory assumes that investors are rational individuals and hence are not influenced by their emotions (Ackert and Deaves, 2010; Nofsinger, 2005). Nevertheless, in reality, markets and individuals are far from being perfect, and individuals have several biases, which affect their decision-making (Byrne and Utkus, 2013; Ackert and Daves, 2010). It can be shown how behavioral finance can explain some of the anomalies derived from the EMH (Ackert and Deaves, 2010; Odean, 1999).
For instance, the optimism bias is a twist in how people perceive reality while thinking they are acting in a rational way (Ariely, 2009). Survey results show that analysts and business managers tend to be very optimistic about the chances of success of the company they work for (Ackert and Deaves, 2010). This bias can lead managers to overestimate the likelihood of success and underestimate the risks (Nofsinger, 2005). For example, when taking financial decisions, managers and individuals who participate in financial market usually make mistakes, as they tend to overestimate the likelihood of good outcomes, which leads them to overvalue their own projects and they might even invest in negative net present value (NPV) projects (Heaton, 2002).
since it leads managers to take less rational decisions that can be very costly (Gervais et al., 2010; Barber and Odean, 2000).
Furthermore, there is a relationship between managerial optimism and investment level (Chen and Lin, 2012). Ben-David and al. (2007), after calculating optimism and overconfidence measures, find that firms with overly optimistic managers tend to invest more. Also, optimistic biased managers, who think that the market underestimates their risky securities, may pass-up positive net present value projects, and may invest in negative net present value projects because of biased returns forecasts (Heaton, 2002; Kahneman, 2011). Likewise, mergers and acquisitions (M&A) agreements can be seen as another example of optimism bias, as optimistic managers overestimate the future value of the new firm (Malmendier and Tate, 2005). Malmendier and Tate (2005) find that optimistic CEOs are more likely to overpay target firms and even take on value-destroying M&A compared to rational ones. In addition, they overestimate the returns of such investment and, as a consequence, they over-invest and tend to choose higher leverage (Kahneman, 2011; Baker et al., 2007). However, they can be disappointed as the stocks of the buying firm drop in about two-thirds of the cases (Mauboussin, 2009).
We can conclude that there is enough evidence showing that individuals often overestimate the chances of a good financial outcome when taking financial decisions. This is the case for managers who are optimistic biased about the future performance of their firm and take on risky corporate financial decisions (Malmendier and Tate, 2005; Kahneman, 2011; Baker et al., 2007; Ben-David et al., 2007). As managers are influenced by optimism in their financial decision-making, optimistic households also tend to take riskier financial decisions (Balasuriya et al., 2010; Puri and Robinson, 2007).
2.3 Optimism and household finance
The household sector plays an important role in the economy and financial markets due to its large size (Campbell et al., 2009; Balasuriya et al., 2010). Therefore, we analyze the relationship between optimism and household financial decisions.
An individual who is optimistic about his future might also be optimistic about the stock market and might take riskier financial decisions (Balasuriya et al., 2010; Nofsinger, 2005). It therefore leads investors not to diversify their portfolios appropriately (Baker and Riccardi, 2014; Barber and Odean, 2000). Barber and Odean (2000) find that overconfident investors take excessive risk as they underdiversify their portfolios and hold on average only 4 stocks. Puri and Robison’s (2007) paper is the most important published paper that addresses the influence of optimism on households’ economic choice.
individual characteristics with the respondents’ self-reported life expectancy. They use the answers to the question: “About how long do you think you will live?” After collecting all the answers, they construct a regression where the dependent variable is the optimism index and the independent variables are some individual characteristics. They find that differences in gender, race and level of education among others; e.g. men are more optimistic than women and white people are more optimistic when compared to other races. In order to check whether optimism influences risk taking, Puri and Robinson (2007) study the correlation between these two variables. They measure risk taking using people’s financial risk tolerance. They gather the information from the respondents’ answer to the following question “How much financial risk are you willing to take?” Although optimism is highly related to risky behavior they find a low correlation (9.40%) between optimism and risk taking.
Puri and Robinson (2007) also study the relationship between optimism and portfolio allocation. To measure portfolio allocation, they create two variables of stock ownership; i.e. the quantity of stock wealth as a proportion of total equity, and the proportion of equity wealth to total financial wealth. They find that optimism is significantly correlated with participation in the equity market. However, they find an insignificant effect of optimism on the equity assets to total financial wealth ratio. They conclude that optimism is correlated with equity wealth allocation between equity assets, but not among equity to debt extent. Moreover, they find that optimistic people are more likely to invest in individual stocks than in mutual funds or other equity instruments.
Further, in order to differentiate good optimism from bad optimism they consider a degree of life expectancy miscalibration. They find two types of optimistic behavior, the moderate optimists and the extreme optimists. They define extreme optimism as the 5% on the right tail of the distribution and conclude that the first ones have more prudent financial behaviors than the second ones, since extremely optimistic individuals take riskier financial decisions as they tend to focus on shorter horizons saving less and holding more stocks than less optimistic individuals.
2.4 Individual characteristics in risk taking
Many studies show that individual characteristics influence individuals’ portfolio choices (Riley and Chow, 1992; Rooij et al., 2011; Heaton and Lucas, 2000; Barber and Odean, 2001). We take a look at the following main characteristics: gender, age, education level, marital status, occupation and wealth.
Overall, literature research states that there is a positive relationship between investing in risky assets and wealth, education level and being a male, and a negative relationship between risky asset investment and age, being married and being a woman (Riley and Chow, 1992). We discuss the influence of each demographic variable on investment decisions. These demographics are used later on in this study as control variables.
Generally, women are more risk averse than men and participate less in the stock market (Byrnes et al., 1999; Rooij et al., 2011; Barber and Odean, 2001). Byrnes et al. (1999) use the SCF to estimate the coefficient of risk aversion and find that men invest more in stocks than women. Further, women are less optimistic and overconfident than men (Barber and Odean, 2001; Ackert and Deaves, 2010).
These differences are due to socio-cultural reasons; i.e. different gender roles (Beyer and Bowden, 1997). Women investors invest less in risky assets since they focus more on the chances of a loss (Olsen and Cox, 2001; Barber and Odean, 2001; Baker and Riccardi, 2014). Thus, women place more importance to safety than men (Byrnes et al., 1999). Education and literacy also play an important role in this gender gap since women usually lack of financial literacy, and men are more financially literate (Lusardi and Mitchell, 2014; Rooij et al., 2011).
Age plays an important role in individuals’ financial decisions due to the life-cycle (Lusardi and Mitchell, 2014; Agarwal et al., 2009). For example, people who just started working and need to invest their money in the acquisition of a house and in primary necessities invest less than people in a later stage who have less expenses (Agarwal et al., 2009). Therefore, age has an influence on individual investment choice. Riley and Chow (1992) find that risk aversion declines with age, and increases again in a later age; people in the age range between 55 and 64 tend to be more risk tolerant than younger people or people after the age of 65. We then expect younger individuals and retired people to hold a lower amount of risky assets in their portfolios. We also expect optimism to increase with age (Puri and Robinson, 2007).
Marital status can play an important role in portfolio allocation. There is a negative relationship between participation in stock market and marital status as, all other things equal, married people are less risk tolerant than single ones (Riley and Chow, 1992). However, married women tend to invest more in risky stocks than single women (Barber and Odean, 2001).
Occupation can also affect individuals’ portfolio decisions (Riley and Chow, 1992). People owning their own business or those who are self-employed are less likely to invest in stocks, they are more risk averse and choose safer portfolios in order to guarantee the business continuity and diversify their business idiosyncratic risk (Heaton and Lucas, 2000; Riley and Chow, 1992).
Wealth and income play very important roles in portfolio allocation. For instance, risk tolerance may change due to a person’s increase or decrease in his or her wealth level or income (Riley and Chow, 1992; Plantinga, 2012). High-income people are believed to be more risk taking, as the proportion of risky assets held increases as income increases (Riley and Chow, 1992; Cohn et al., 1975). However, the inverse can also happens, as wealthier people may want to secure their profits (Plantinga, 2012). Further, there is evidence that as income increases, people become more optimistic (Puri and Robinson, 2007).
3. Methodology 3.1 Hypotheses
As previously mentioned, evidence from literature show that optimism influences individuals decision-making. Optimistic individuals overestimate the probability of positive outcomes and hence take on higher risks (Ariely, 2009; Ackert and Deaves, 2010); i.e. they invest in riskier assets or take on riskier business strategies (Heaton, 2000). Puri and Robison (2007) find that optimism and investment in stocks are positively correlated. In order to test whether this is also the case in our sample we test the following hypothesis:
Hypothesis 1: Optimistic individuals take more financial risks.
Puri and Robinson (2007), after differentiating extreme optimists from moderate ones, find that extreme ones take on higher risks than moderate ones. In order to check whether extreme optimists invest in riskier assets than moderate ones we test the following hypothesis:
Hypothesis 2: Extreme optimists invest in riskier assets than moderate optimists.
3.2 Sample and Data
In this study, the variable of interest is the individual participation in the stock market and the main dependent variable is optimism.
Nonetheless many other variables influence the independent variable; i.e. individual characteristics (Riley and Chow, 1992). Therefore, in order to draw a plausible conclusion about the influence of optimism on the stock market participation we need to create the following control variables: age, gender, education level, marital status, occupation, wealth and income.
In order to measure our variables of interest we use the DNB Household Survey (DHS). This survey includes a large dataset of economic and psychological features of financial behavior of 2,000 Dutch households approximately and gathers information with annual frequency since 1993 (Teppa and Vis, 2012). This survey contains five different questionnaires about employment and pensions, housing and mortgages, income and health, assets and debts, and economic and psychological concepts. For this study we use information gathered from 2003 until 2013 surveys, where we can find useful information for this study. This results in a total of 49,659 observations. Afterwards, we restrict the sample to households for which we have information on aggregate wealth, and we get 20,883 observations.
In table 1 we find the descriptive statistics for the control variables. In our restricted sample, from the total number of participants owning assets, 27.26% participate in equity market and the age of the participants goes from 16 to 122 and the mean age is 51. From those participants 54.59% are men and 45.41% are women, and 56.3% have a college education. Regarding household composition, 64.67% are the head of the household; 77.81% have children at home; 63.12% are married or living with a partner. With regards to employment, 49.43% are employed; 4.92% are self-employed (or own a business); 16.63% are unself-employed (including students); and 22.14% are retired.
3.3 Optimism measurement
As discussed earlier, in the literature we find two main ways to measure optimism. The first one consists on measuring optimism directly and the second one measures optimism indirectly. In this study we use both measures of optimism.
3.3.1 Direct Optimism Measurement
experiencing a good or bad future event relative to their peers (Weinstein, 1980).
In this study we first measure optimism directly using the DHS psychological and economic questionnaire where we can find questions similar to the ones asked in the LOT questionnaire. We measure it by analyzing the question that relates to people’s perceptions about their future financial situation: “How do you think the economic situation of your household will be in five years?” Respondents are allowed to choose between the following answers: “Much worse”, “Worse”; “(about) The same”; “Better”; “Much better”. Participants who believe their situation will improve are considered optimistic. We create a dummy variable that takes the value of 1 if the individual is considered optimistic and 0 otherwise.
Following Weinstein (1980), we analyze respondents’ perceptions of their own financial situation when compared to others. In the DHS participants are asked to rate the following statements “Compared to others in my environment, I am better off”; “If I compare myself to my friends, I think in general I am financially better”. Individuals answer using a scale from 1 to 7, where 1 means “totally disagree”, and 7 means “totally agree”. Participants who state they are better off than their peers (higher than 5) are considered optimistic. We use the same procedure as before and create a dummy for optimism.
3.3.2. Indirect Optimism Measurement
Following Puri and Robinson (2007) we measure optimism indirectly. In order to do so, we use the question regarding people’s perception of their life expectancy within the income and health questionnaire. The DHS ask the following question: “How likely is that you will attain the age of x?”; where x varies depending on the participant’s age. Participants under the age of 90 assign to each item a rating using a scale from 0 to 10, where 0 means “no chances at all” and 10 means “absolutely certain”.
For each person we collect his/her age and expected age based on the information underlying the “life tables”, and split it by gender. Then taking the people’s self-reported life-expectancy most closely connected to their expected age at death we create a dummy variable for optimism, where 1 means optimistic and 0 means non-optimistic.
3.3.3 Moderate vs Extreme Optimists
In this analysis we follow Puri and Robinson (2007) and distinguish between moderate and extreme optimists in order to check whether extreme ones are more likely to invest in riskier assets than moderate ones.
extreme optimistic participants the ones who think they have a very high chance that their financial situation will be better in the future or the ones who believe they are financially much better than other people from their environment.
We create a dummy variable for extreme optimism, where 1 means extreme optimistic and 0 means moderate optimistic.
3.4 Risk Taking Measurement
We follow Puri and Robinson (2007) measure of risk taking and construct a similar variable for financial risk behavior using the DHS economic and psychological questionnaire. To measure individuals’ risk-taking in the past years this survey asks the participants to choose between the following five statements: “I have taken no risk at all”; “I have taken small risks every now and then”; “I have taken some risks”; “I have sometimes taken great risks”; “I have often taken great risks”. We then construct a risk-taking variable in order to check if more optimistic people have taken more risks. Following the literature findings, we expect optimism to have a positive influence in risk taking.
3.5 Participation in Financial Market Measurement
We define risky financial behavior as holding a positive amount of money in stocks and shares and construct one variable of portfolio allocation. Following Puri and Robinson (2007) we construct the following ratio, total equity to total financial assets, to check whether optimistic people are more eager to invest in risky assets. Total equity includes stocks and shares, and mutual funds. Following the hypotheses and the results found in the literature, we expect optimism to have a positive influence in the amount of equity invested.
3.6 Control variables
As mentioned before, several variables affect individual stock market participation and optimism; i.e. individual characteristics (Riley and Chow, 1992; Puri and Robinson, 2007; Barber and Odean, 2001). In order to avoid any misinterpretation we create the following control variables.
As age can influence investment decisions and optimism, we control for age differences and create an age variable (in log). Risk aversion decreases with age until it reaches a pick (Riley and Chow, 1992; Rooij et al., 2011). We then expect that people who are between 50 and 64 are more likely to participate in the financial market than younger or retired ones.
Marital status also affects individual participation in the financial market and the level of optimism (Barber and Odean, 2001; Riley and Chow, 1992). We create a dummy variable, where 1 means married and 0 single. Married people and/or people who have children tend to take less risk than single ones (Riley and Chow, 1992). Following these findings, we expect marital status to have a negative impact in equity holdings.
The education level and the financial literacy play a very important role when it comes to financial decisions (Rooij et al., 2011; Lusardi and Mitchell, 2014). As discussed above, those who have a lower education level or have little financial knowledge are less likely to invest in the stock market than those who have a high education or/and high financial literacy (Rooij et al., 2011; Lusardi and Mitchell, 2014). Therefore, we expect that people with higher education levels to have a higher participation in the stock market. We create a dummy variable for high education, where 1 means high level of education (including university and vocational colleges), and 0 means lower level of education.
Literature finds that participation in the stock market differs depending on the individual’s primary occupation. Individuals who own a business or are self-employed are more risk averse, and hence invest less in risky assets than other types of employees (Heaton and Lucas, 2000). We then expect self-employment and business ownership to have a negative influence on stock market participation. We create dummy variables for employees, self-employed and owners of a business, and retired.
Finally, as wealth can play an important role in financial decisions, we need to control for it (Riley and Chow, 1992). As discussed above, in general the higher the income the higher the amount of individual stocks held (Riley and Chow, 1992). We expect wealth to have a positive impact on individual participation in the stock market.
4. Model
Since we have information on multiple households at two or more point in time the simplest way to deal with this data would be to estimate the following pooled regression:
𝑒𝑞𝑢𝑖𝑡𝑦/𝑎𝑠𝑠𝑒𝑡𝑠!" = 𝛼 + 𝛽𝑥!"+ 𝑢!" (1)
where 𝑒𝑞𝑢𝑖𝑡𝑦/𝑎𝑠𝑠𝑒𝑡𝑠!" is the fraction of equity held to total assets, 𝑥!" are all the variables that have an influence on financial decisions.
However, this estimation does not include heterogeneity, which is present in data from entities or individuals that vary over time. In order to check whether we should use a simple pooled regression or fixed effects we run the following regression and some tests.
𝑒𝑞𝑢𝑖𝑡𝑦/𝑎𝑠𝑠𝑒𝑡𝑠!" = 𝛼!+ 𝜇!"𝑜𝑝𝑡𝑖𝑚𝑖𝑠𝑚 + 𝛽!"𝑐𝑜𝑛𝑡𝑟𝑜𝑙𝑠 !
!!!
+ 𝑢! (2)
where 𝑒𝑞𝑢𝑖𝑡𝑦/𝑎𝑠𝑠𝑒𝑡𝑠!" is the fraction of equity held to total assets, 𝑜𝑝𝑡𝑖𝑚𝑖𝑠𝑚 represents households who are optimistic, and controls are all the control variables that affect the amount of equity held as a proportion of total financial assets.
In order to check whether optimism influences risk-taking we run the following regression:
𝑟𝑖𝑠𝑘𝑡𝑎𝑘𝑖𝑛𝑔!" = 𝛼!+ 𝜇!"𝑜𝑝𝑡𝑖𝑚𝑖𝑠𝑚 + 𝛽!"𝑐𝑜𝑛𝑡𝑟𝑜𝑙𝑠 !
!!!
+ 𝑢! (3)
where 𝑟𝑖𝑠𝑘𝑡𝑎𝑘𝑖𝑛𝑔!" is the amount of risk taken in the past few years and controls are all the control variables that affect the amount of risk taken.
The coefficient of optimism is expected to be positive and significant in both regressions.
After running the redundant fixed effect test, we conclude that we should estimate our model using fixed effects. Nonetheless, the fixed effects model has some drawbacks as time invariant variables drop out of the model. Random effects model corrects for this, however this model assumes that individual effects are uncorrelated with the explanatory variables. In order to check which method we should use we perform the Hausman test and conclude that the fixed effects model gives a better estimation of this model for both cross-section and time effects.
We first estimate the model using cross-section fixed effects and then using time fixed effects.
In order to test the second hypothesis we need to check whether extreme optimistic individuals take higher financial risks than moderate ones. In order to do so, we include the dummy variable for extreme optimists and run the following regressions:
𝑒𝑞𝑢𝑖𝑡𝑦/𝑎𝑠𝑠𝑒𝑡𝑠!" = 𝛼!+ 𝜇!"𝑒𝑥𝑡𝑟𝑒𝑚𝑒 + 𝛽!"𝑐𝑜𝑛𝑡𝑟𝑜𝑙𝑠 ! !!! + 𝑢! (4) 𝑟𝑖𝑠𝑘𝑡𝑎𝑘𝑖𝑛𝑔!" = 𝛼!+ 𝜇!"𝑒𝑥𝑡𝑟𝑒𝑚𝑒 + 𝛽!"𝑐𝑜𝑛𝑡𝑟𝑜𝑙𝑠 ! !!! + 𝑢! (5)
where 𝑒𝑞𝑢𝑖𝑡𝑦/𝑎𝑠𝑠𝑒𝑡𝑠!" is the fraction of equity held to total assets, 𝑒𝑥𝑡𝑟𝑒𝑚𝑒 is the dummy variable that represents individuals who are extreme optimistic, and controls are all the control variables that affect the amount of risk taken; and 𝑟𝑖𝑠𝑘𝑡𝑎𝑘𝑖𝑛𝑔!" is the amount of risk taken in the past few years.
Robustness
After having performed all the regressions, we should carry out some robustness tests. In order to verify the influence of the independent variables on the dependent variable and the robustness of the results we first run a regression removing the optimism dummy variable, and afterwards we use different possible combinations of control variables. We run the following regressions:
𝑒𝑞𝑢𝑖𝑡𝑦/𝑎𝑠𝑠𝑒𝑡𝑠!" = 𝛼!+ 𝛽!"𝑐𝑜𝑛𝑡𝑟𝑜𝑙𝑠 ! !!! + 𝑢! (6) 𝑟𝑖𝑠𝑘 − 𝑡𝑎𝑘𝑖𝑛𝑔!" = 𝛼!+ 𝛽!"𝑐𝑜𝑛𝑡𝑟𝑜𝑙𝑠 ! !!! + 𝑢! (7)
Secondly, we run a regression that includes the optimism regressor only. 𝑒𝑞𝑢𝑖𝑡𝑦/𝑎𝑠𝑠𝑒𝑡𝑠!" = 𝛼!+ 𝜇!"𝑜𝑝𝑡𝑖𝑚𝑖𝑠𝑚 + 𝑢! (8)
𝑟𝑖𝑠𝑘𝑡𝑎𝑘𝑖𝑛𝑔!" = 𝛼!+ 𝜇!"𝑜𝑝𝑡𝑖𝑚𝑖𝑠𝑚 + 𝑢! (9)
5. Results
Table 1 reports the summary statistics of the participants characteristics from our restricted sample; i.e. participants from which we have information about their financial wealth. We create dummy variables for most of the demographics except for age and children.
The table gives the number of observations, means, standard deviations, minimum observations and maximum observations of the variables included in this study.
Table 1
Descriptive statistics of participants’ characteristics
Demographic Obs Mean St. Dev Max Min
Age 20,883 50.47 16.246 16 122 Men 20,883 0.546 0.498 0 1 Women 20,883 0.454 0.498 0 1 Head hh 20,883 0.647 0.478 0 1 College 20,850 0.563 0.496 0 1 Married 17,988 0.631 0.482 0 1 Employed 20,883 0.494 0.500 0 1 Self-employed 20,883 0.049 0.215 0 1 Retired 20,883 0.221 0.415 0 1 Unemployed 20,883 0.166 0.372 0 1 Children in hh 20,883 0.778 1.103 0 6
* The panel is unbalanced since some households do not participate in every survey, and some of them do not respond to some questions of interest, therefore we do not get the same number of observations for every variable.
The average age of participants owning assets is 50, and the majority are men (54.6%) and head of the household (64.7%); which is consistent with literature (Barber and Odean, 2001; Riley and Chow, 1992; Rooij et al., 2011). The majority of them has a college education (56.3%) and are married or living with a partner (63.1%) or have children in the household (77.8%). There are more employed people (49.4%) than self-employed (0.049%) or retired (22.1%) and 16.6% are unself-employed.
Table 2: descriptive statistics of optimism related to individuals' financial expectation (Optimism 1) and their perception of their financial situation compared to others (Optimism2).
Obs Mean St.Dev Min Max
Optimism 1 17,698 0.232 0.422 0 1
Optimism 2 17,046 0.222 0.416 0 1
Overall, less than a quarter of the sample is optimistic. With regards to their financial expectations, approximately 23.2% believe they will be financially better off in the next five years throughout the 11-years survey period.
Results are similar with regards to people’s perceptions of their situation related to others in their environment, where 22.2% think they are better off.
Table 3 reports the descriptive statistics of the indirect measure of optimism (life expectancy miscalibration). We consider optimistic individuals the ones who believe they will live longer than the life expectancy reported in the life tables corresponding to their age and sex. We use this measure to perform a sample validity check on our direct measure of optimism.
Table 3
Descriptive statistics of optimism related to individuals' self-reported life expectancy.
Obs Mean St.Dev Min Max
Optimism 17,890 0.416 0.493 0 1
With regards to their life expectancy, people are more optimistic than with regards to their financial situation. From 2003 to 2013, 41.6% of respondents are optimistic when using this method compared to 23% when using the direct method.
We run regressions by using equations 1 and 2 to test hypothesis 1 and hypothesis 2. We display the results in the following paragraphs.
Table 4
Optimism and Portfolio choice- Cross Section Fixed Effects
VARIABLES Equity /Assets Risk-Taking
Optimism 0.020*** 0.025** (0.004) (0.010) Age (logs) 0.076*** -0.144*** (0.011) (0.024) Wealth(logs) 0.007*** -0.008 (0.005) (0.007) Self-employed 0.044** 0.057* (0.015) (0.031) Married -0.004 0.011 (0.007) (0.013) Retired -0.001 -0.012 (0.008) (0.017) Unemployed -0.011* 0.016 (0.009) (0.020) Constant -0.261*** 1.246*** (0.056) (0.107) Observations 11,581 11,139 Number of id 3,248 3,153
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Overall, in Table 4 the estimated results are significant at a 95% confidence level, showing that the probability of investing in risky assets as a fraction of total assets and risk-taking increase when individuals are optimistic. Both regressions’ results show a positive relationship between risky financial behavior and optimism.
With regards to the first regression, we find that optimism is positively correlated with stock market participation, which is consistent with literature findings (Puri and Robinson, 2007; Ben-David et al., 2007). When individuals are optimistic, they hold a higher amount of equity relative to their total financial assets. The coefficient for optimism is positive (0.020) and is significant at 99% confidence level, which shows that optimism has a positive impact on investing in equity. This proves that optimistic investors are more likely to hold a higher fraction of equity among their total assets, which is in line with Puri and Robinson’s (2007) findings.
self-employed increases the investment in equity as a fraction of total financial assets, which is not consistent with literature (Riley and Chow, 1992). Other variables (i.e. other non-observable characteristics) that influence the participation in equity are reflected in the constant, which is negative and significant at a 95% level.
Further, we also find that optimistic individual are more likely to take higher risks. The coefficient is positive (0.025) and significant at a 95% confidence level, which means that optimistic people have taken more risks in the past. However, among demographic variables, age is significant and negatively correlated with risk taking, which is surprising since we find that age is positively correlated with the amount of equity held. The life-cycle could be the reason for this. For example, some people took fewer risks in the past few years maybe because they were still in early stages of the life-cycle or because they retired (Lusardi and Mitchell, 2014; Agarwal et al., 2009). The rest of the variables do not vary that much. Nonetheless, here the constant is positive and significant at a 99% confidence level.
Table 5 presents the results of the estimates after using time fixed effects after correcting for heteroskedasticity. We use the same variables as above.
Table 5
Optimism and portfolio choice – Time Fixed Effects
Variables Equity/Assets Risk-Taking
Optimism 0.015*** -0.011 (0.004) (0.011) Age (logs) 0.082*** -0.049** (0.010) (0.021) Wealth (log) 0.008*** -0.003 (0.002) (0.006) Self-employed 0.045*** 0.082* (0.026) (0.027) Married -0.001 -0.017 (0.005) (0.011) Retired -0.006 0.020 (0.007) (0.015) Unemployed -0.011 0.017 (0.007) (0.018) Constant -0.293*** 0.927* (0.043) (0.095) Observations 11,581 11,139 Number of id 3,248 3,153 R-squared 0.011 0.173
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Overall, when we use time fixed effects we find significant results. However, we do not find the same results for both regressions.
equity held as a fraction of total assets. For the control variables we find the same results as we find when we use cross-section fixed effects.
In the second regression we find opposite results. Here, optimism is not significant; hence it does not have an influence on risk-taking when using time-fixed effects. Wealth is also negative and insignificant, which is in not in line with the discussed literature (Riley and Chow, 1992). The other control variables have a similar influence as before.
Table 6 presents the results after running cross-sectional fixed effects for equations 4 and 5, and corrected for heteroskedasticity. The second and forth column present the results for the extreme optimistic households, and the third and fifth column present the results for the moderate ones. Extreme optimistic individuals are the ones who think they are much better off than others in their environment. In the survey participants who rated with a 7 are considered extreme optimistic and the ones who rated with a 5 or 6 are considered moderate optimistic.
Table 6
Extreme and Moderate Optimism- Cross Section Fixed Effects
Extreme Moderate Extreme Moderate VARIABLES Equity/Assets Equity/Assets Risk-taking Risk-taking
Optimism 0.028* 0.016*** 0.089*** 0.017 (0.016) (0.004) (0.028) (0.011) Age (logs) 0.069*** 0.069*** -0.148*** -0.128*** (0.011) (0.010) (0.024) (0.024) Wealth (log) 0.007 0.008* -0.007 -0.007 (0.005) (0.004) (0.007) (0.007) Self-employed 0.340*** 0.330*** 0.058* 0.056* (0.030) (0.029) (0.030) (0.031) Married -0.004 -0.003 0.011 0.013 (0.007) (0.007) (0.013) (0.013) Retired -0.001 0.000 -0.012 -0.017 (0.008) (0.008) (0.017) (0.017) Unemployed -0.011 -0.008 0.016 0.012 (0.009) (0.008) (0.020) (0.020) Constant -0.251*** -0.259*** 1.255*** 1.180*** (0.055) (0.049) (0.107) (0.105) Observations 11,581 12,589 11,139 11,319 Number of id 3,248 3,510 3,153 3,185
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
who find that extreme optimists take higher risks than moderate ones.
With regards to the first regression we find that both moderate and extreme optimists hold a high amount of risky assets as a fraction of total financial wealth. Both coefficients are significant and statistically significant. The coefficient for extreme optimists is slightly higher (0.028) than the one for moderate optimists (0.016).
Robustness tests
In this section we provide the results after carrying robustness checks.
Table 7 presents the results after running cross-section fixed effects without including optimism as an explanatory variable.
Table 7
Robustness test removing optimism regressor. Cross-section fixed effects.
Variables Equity/Assets Risk-taking
Age (log) 0.073*** -0.148*** (0.009) (0.024) Wealth (log) 0.007*** -0.007 (0.003) (0.007) Self-employed 0.044*** 0.059* (0.015) (0.031) Married -0.006 0.011 (0.006) (0.013) Retired -0.005 -0.012 (0.008) (0.017) Unemployed -0.012 0.016 (0.007) (0.020) Constant -0.272*** 1.256*** (0.041) (0.107) Observations 11,581 11,139 Number of id 3,248 3,153
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Table 8 presents the results after running cross-section fixed effects without including any control variable and just including the optimism dummy variable (direct measure). Table 8
Robustness test removing all control variables. Cross-section FE
VARIABLES Equity/Assets Risk-taking
Optimism 0.055*** 0.024*** (0.018) (0.009) Constant 0.053*** 0.608*** (0.018) (0.004) Observations 17,698 16,511 R-squared 0.004 0.000
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
When we run a regression with optimism as a unique regressor we also find significant and positive results as the ones we find when we regress adding the control variables. We then conclude that the coefficients are robust.
Table 9 presents robustness test following Puri and Robinson (2007) and using the indirect measure of optimism; i.e. life miscalibration; to see whether we find the same results when using a different optimism measure.
Table 9
Robustness test using indirect measure of optimism. Cross-section FE
Variables Equity/Assets Risk-taking
Optimism 0.018** 0.051** (0.010) (0.026) Age (log) 0.196 -1.546*** (0.123) (0.324) Wealth (log) 0.026*** 0.001 (0.005) (0.017) Self-employed -0.005 -0.089 (0.075) (0.151) Married 0.000 0.001 (0.016) (0.032) Retired -0.027 0.036 (0.024) (0.053) Unemployed -0.040 0.024 (0.030) (0.068) Constant -0.971* 7.051*** (0.510) (1.378) Observations 2,159 2,094 Number of id 698 683
When we run the regressions using the indirect optimism measurement the results do not change much. The coefficient for optimism is still positive and significant for both regressions at a 95% confidence level. However, in the first regression the coefficient of age and self-employment are not significant anymore.
6. Conclusion
Recent studies in finance have shown that management and individuals’ financial decisions are related to factors that are not captured by the classical finance theory (Malmendier and Tate, 2005; Byrne and Utkus, 2013; Nofsinger, 2005; Ackert and Daves, 2010; Kahneman, 2011). The purpose of this paper was to analyze whether Dutch households are optimistic and whether this influences their financial decision-making. In order to draw a plausible conclusion about the influence of optimism on households’ financial decisions we included some control variables that might influence individual participation in the stock market; i.e. individual characteristics (Riley and Chow, 1992; Puri and Robinson, 2007). We measured the level of individual optimism and the amount of financial risk taken using the DHS, and fond that optimism is an important determinant of investing in risky assets. The coefficient of optimism is positive and significant at a 95% confidence level. Individuals who are optimistic invest more in equity and have taken higher risks in the past years than those who are not optimistic. This finding remains the same in both panel data analysis. However, when using time fixed effects model optimism does not have any influence on the risk taken in the past few years. We further follow Puri and Robinson (2007) analysis and provide some results that support the hypothesis that extreme optimists take higher financial risks than moderate ones. These results are in line with Puri and Robison (2007) findings that extreme optimists take on higher financial risks.
After carrying out some robustness analysis, we dismiss the possibility that the relation between optimism and investment in risky assets is driven by other variables outstanding in the literature like individual characteristics. When running a regression with optimism as a unique regressor we find the same significant and positive results as when we run the regression including the control variables. We also find similar results when running the regression including just the control variables and not the optimism dummy variable. We also show that this relationship does not disappear when we run the regression using life miscalibration as a measure of optimism.
References
Ackert, L.C., Deaves, R., 2010. Behavioral Finance. Cengage Learning.
Agarwal, S., Driscoll, J. C., Gabaix, X., Laibson, D. I., 2009. The age of reason: financial decisions over the life-cycle with implications for regulation. Brookings Papers on Economic Activity 40, 51–117.
Antoni, M.H., Goodkin, K., 1988. Host moderator variables in the promotion of cervical neoplasia. Journal of Psychosomatic Research 32, 327-338.
Ariely,D., 2009. The curious paradox of optimism bias. Bloomberg Business. Available at: http://www.bloomberg.com/bw/magazine/content/09_34/b4144048821798.htm Baker, M., Ruback, R., Wurgler, J., 2007. Behavioral corporate finance: a survey. Handbook of Corporate Finance: Empirical Corporate Finance. New York.
Baker, H.K., Riccardi, V., 2014. How biases affect investor behavior. The European Financial Review, 7-10.
Balasuriya, J., Muradoglu, Y.G., Ayton, P., 2010. Optimism and portfolio choice. Working Paper. University of London.
Barber, B.M., Odean, T., 2000. Trading is hazardous to your wealth: the common stock investment performance of individual investors. Journal of Finance 55(2), 773-885. Barber,B.M., Odean,T., 2001. Boys will be boys: gender, overconfidence, and common stock investment. Quarterly Journal of Economics 116, 261-292.
Barberis, N., Shleifer, A., Vishny, R., 1998. A model of investor sentiment. Journal of Financial Economics 49, 307-343.
Ben-David, I., Graham, J.R., Harvey, C.R., 2007. Managerial overconfidence and corporate policies. Working Paper, University of Chicago and Duke University.
Beyer, S., Bowden,E.M., 1997. Gender differences in self-perceptions: convergent evidence from three measures of accuracy and bias. Personality and Social Psychology Bulletin 23, 157-172.
Biais, B., Hilton, D., Mazurier, K., Pouget, S., 2005. Judgmental overconfidence, self- monitoring, and trading performance in an experimental financial market. Review of Economic studies 72(25), 287-322.
Brunnermeier, M., Parker,J., 2005. Optimal expectations. American Economic Review 95, 1092-1118.
Buehler,R., Griffin, D., Ross,M., 1994. Exploring the “planning fallacy”: why people underestimate their task completion times. Journal of Personality and Social Psychology 67, 366-381.
Byrnes, J.P., Miller, D.C., Schafer, W.D., 1999. Gender differences in risk-taking: a meta-analysis. Psychological Bulletin 125(3), 367-383.
Byrne, A., Utkus, S.P., 2013. Understanding how the mind can help or hinder investment success. Vanguard Asset Management, United Kingdom.
Campbell,J.,Calvet,L.E.,Sodini,P.,2009. Measuring the financial sophistication of households. American Economic Association, 99(2),393-98.
Carr, A., 2011. Positive Psychology: the science of happiness and human strengths. Routledge, London and New York.
Chen,I.J., Lin,S.H., 2012. Will managerial optimism affect the investment efficiency of a firm? Procedia Economics and Finance 2, 73-80.
Chesterman,E., Cohen,F., Adler,N., 1990. Trait optimism as a predictor of pregnancy outcomes. First International Congress on Behavioral Medicine, Uppsala, Sweden. Cohn,R.A., Lewellen,W.G.,Lease,R.C., Sclarbaum,G.G.,1975. Individual investor risk aversion and investment portfolio composition. Journal of Finance 30, 605-620.
Friedman,L., Weinberg, A., Webb,J., Cooper,H., Bruce,S., 1995. Skin cancer prevention and early detection intentions and behaviour. American Journal of Preventive Medicine 11, 59-65.
Gervais, S., Heaton, J.B., Odean, T. ,2010. Overconfidence, compensation contracts, and capital budgeting. Journal of Finance 3, 83-128.
Gillham, J.E., 2000. The science of optimism and hope: research essays in honor of Martin E. P. Seligman. Templeton Foundation Press, Philadelphia.
Heaton,J.B., 2002. Managerial optimism and corporate finance. Financial Management 31, 33-45.
Heaton,J.B.,Lucas,D., 2000. Portfolio choice and asset prices: the importance of entrepreneurial risk. The Journal of Finance 55, 1163-1198.
Helweg-Larsen,M., Shepperd,J.A., 2001. Do moderators of the optimistic bias affect personal or target risk estimates? A review of the literature. Personality and Social Psychology Review (5), 74-95.
Kahneman,D., 2011. Bias, blindness and how we truly think (part 1). Bloomberg Business.
Kenny,C., 2015. How optimism strengthens economies. Bloomberg Business. Available
at:
http://www.bloomberg.com/bw/articles/2015-01-08/how-optimism-strengthens-economies.
Keynes, J.M., 1936. The General Theory of Employment, Interest and Money. Cambridge University Press.
Lusardi, A., Mitchell, O.S., 2014. The economic importance of financial literacy: theory and evidence. Journal of Economic Literature 52(1), 5-44.
Malmendier, U., Tate, G., 2005. CEO overconfidence and corporate investment. Journal of Finance 60, 2661-2700.
Mauboussin, M.J., 2009. Think twice: harnessing the power of counterintuition. Harvard Business School Press, 4-6.
McKenna,F.P., 1993. It won’t happen to me: unrealistic optimism or illusion of control? British Journal of Psychology 84, 39-50.
Miller, E., 1977. Risk, uncertainty, and divergence of opinion. Journal of Finance 32, 1151–1168.
Nofsinger,J.R, 2005. Social mood and financial economics. Journal of Behavioural Finance 6, 144-160.
Odean, T., 1998. Are investors reluctant to realize their losses? Journal of Finance 53, 1775-1798.
Odean, T., 1999. Do investors trade too much? American Economic Review 89, 1279-1298.
Olsen, R.A.,Cox,C.M., 2001. The influence of gender on the perception and response to investment risk: the case of professional investors. The Journal of Psychology and Financial Markets 2, 29-36.
Plantinga, A., 2012. Reader Portfolio Management. University of Groningen.
Puri, M.,Robinson,D.T, 2007. Optimism and economic choice. Journal of Financial Economics 86, 71-99.
Riley, W.B., Chow, K.V., 1992. Asset allocation and individual risk aversion. Financial Analysts Journal 48 (6), 32-37.
Rooij, M., Lusardi, A., Alessie, R., 2011. Financial literacy and stock market participation. Journal of Financial Economics 101(2), 449-472.
Scheier,M., Matthews, K.A., Magovern, J., Lefebvre,R., Abbot,R., 1989. Dispositional optimism and recovery from coronary artery bypass surgery: the beneficial effects on physical and psychological well-being. Personality and Social Psychology 57, 1024-1040.
Scheier,M., Carver, C.S., Bridges, M.W., 1994. Distinguishing optimism from neuroticism: a re-evaluation of the life orientation test. Journal of Personality and Social Psychology 67, 1063-1078.
Schulz,R., Bookwala,J., Knapp,J., Scheier,M., Williamsom,G., 1996. Pessimism, age and cancer mortality. Psychology and Ageing 11, 304-309.
Seligman,M., 2000. Optimism, pessimism, and mortality. Mayo Clinic Proceedings 75, 133-134.
Sharot,T., 2011. The Optimism Bias. Pantheon Books, New York.
Shefrin, H., 2005. A Behavioral Approach to Asset Pricing Theory. Elsevier North-Holland, Amsterdam.
Shefrin, H., 2010. How psychological pitfalls generated the global financial crisis. In Voices of Wisdom: Understanding the Global Financial Crisis, Research Foundation of CFA Institute.
Stockman, F., 2009. The hazards of optimism. The Boston Globe, San Antonio.
https://www.bostonglobe.com/opinion/2013/02/05/could-optimism-have-caused-financial-meltdown/7bgthgu8aiMkFpyJSniBnM/story.html
Tennen,H., Affleck, G., 1987. The costs and benefits of optimistic explanations and dispositional optimism. Journal of Personality 55, 377-392.
Teppa, F., Vis, C., 2012. The center panel and DNB household survey: methodological aspects. DNB Occasional Studies 10(4).
Weinstein,N., 1980. Unrealistic optimism about future life events. Journal of Personality and Social Psychology 39, 806-820.
Weinstein, N., Klein, W.M., 1996. Unrealistic optimism: Present and future. Journal of Social and Clinical Psychology 15, 1–8.
Zang,Y., Fishbach,A., Dhar, R.W., 2007. When thinking beats doing: the role of optimistic expectations in goal-based choice. Journal of Consumer Research 34, 567-578.
APPENDIX
Appendix A: Table A1 reports the mean life expectancy for each group of age and gender gathered from http://www.pensioenpad.nl.
Table A1: reports the average age Dutch population lives depending on their age and gender.
Age Men Women
16 80.14 83.70 55 81.73 85 65 83.3 86.30 70 84.48 87.20 75 86 88.32 80 88.03 89.81 85 90.71 91.89 90 93.63 94.66 .
Appendix B: Correlation matrix of all the variables included in the regression. Table B: Correlation matrix between variables in the regression
Appendix C: Table C1 provides the estimated results after running a pooled OLS regression. Equity/Assets is the amount of total equity to total financial assets. Risk-taking is the amount of risk taken in the past few years. We use the measure of direct optimism in this case. The rest of the variables are control variables that influence the dependent variable and can vary over time.
VARIABLES Equity/Assets Risk-taking
Optimism 0.039*** 0.018* (0.005) (0.011) Age (log) 0.069*** -0.088*** (0.009) (0.019) Wealth (log) 0.018*** -0.011* (0.003) (0.006) Self-employed 0.392*** 0.047* (0.012) (0.025) Married -0.006 -0.005 (0.005) (0.010) Retired 0.006 -0.010 (0.006) (0.013) Unemployed 0.002 0.019 (0.008) (0.017) Constant -0.371*** 1.075*** (0.042) (0.089) Observations 11,581 11,139 R-squared 0.105 0.006
