Once
Bitten, Twice Shy
:
Do Macroeconomic Experiences Affect
Risk Tolerance
?
1Yichen Yao (Arietta) Student number: S2439158
MSc Finance
Faculty of Economics and Business University of Groningen
Supervisor: Dr. Auke Plantinga June 2014
A
BSTRACTUsing four Survey of Consumer Finances cross-sectional datasets representing the years 1989 through 2010, this study investigates whether personal experiences of stock market returns affect financial risk tolerance. This study finds that the effect of experienced stock market returns on financial risk tolerance in household heads of all age groups is not significant. However, when this study looks at different age groups, it finds that the experienced stock market returns have higher impact on age group of 20-30 and 40-50 than the age group of 30-40. The household heads in their 20s and 40s, who have experienced higher stock market returns have a greater willingness to take financial risk, are more likely to participate in the stock market, invest a higher fraction of their liquid assets in stocks if they participate. But the effect of experienced returns on the level of risk taking of household heads in their 30s is not significant.
Keywords:Behavioural finance, household finance, investment decision JEL Classification:D03, D14, D83, D84, E21, G11
1
1
1. Introduction
“Once bitten, twice shy” is a common phrase to describe a person who is much more careful to avoid repeating the same unpleasant experience he or she has already experienced. Similarly, investors also may have similar behaviour when investing in stocks. There is ample literature to support that a close link between individual experiences and personal behaviour. Friedman and Schwartz (1963) find that pessimism created by the Great Depression had a persistent effect on markets. Recently, Malmendier and Nagel (2011) also determine that individuals who had experienced lower stock market returns make households less likely to take financial risks, less likely to participate in the stock market, and, if they participate, invest a lower fraction of their liquid assets in stocks. Thus, how does individual experiences affect personal behaviour is an interesting topic to study.
From 2004 to early 2007, the financial markets exhibited low volatility. However, this situation was changed by the financial crisis of 2008: the financial market had become extremely volatile. Moreover, in the wake of the collapse of Lehman Brothers, the stock market reacted dramatically around the globe. In the US, equity wealth dropped 40% during just nine months from January to October 2008, from $20 trillion to $12 trillion. (Senbet and Gande, 2009) During this period of 2004-2007, a majority of investors had bad experiences that may influence their behaviour in the future, specifically their individual’s risk tolerance and stock market participation.
2
However, the financial crisis inspired more scholars to pay close attention to the field of behavioural finance. Behavioural finance combines disciplines of psychology and economics to explain and increase understanding of the reasoning patterns of investors, including the emotional processes involved and the degree to which they influence the decision-making process. Essentially, behavioural finance attempts to explain the what, why, and how of finance and investing, from a human perspective (Ricciardi and Simon, 2000). Currently, many economists study behavioural finance to explain many stock market anomalies (such as the January effect), speculative market bubbles (the recent retail Internet stock craze of 1999), and crashes (crash of 1929 and 1987).
Previous research done by Malmendier and Nagel (2011) examines differences in the influence of macro-economic shocks on long-term risk attitudes. Malmendier and Nagel (2011) find that those who had experienced a stock market boom throughout their life took more risk than those who had not. Also, those experiencing a persistent bull market were more likely to hold stocks and held a higher proportion of wealth in the form of stocks.
3
stock market indexes to describe the US stock market performance, and to find the periods of boom and recession: Dow Jones Industrial Average, National Association of Securities Dealers Automated Quotation (NASDAQ) Index and Standard & Poor's (S&P) 500 Index.
Based on the preceding conceptual framework, the following hypotheses are proposed: Hypothesis 1: Individuals who had only experienced a stock market boom have a higher financial risk tolerance.
Hypothesis 2: Personal experience in American stock market has a greater impact on age group of 20-30-years old and 40-50-years old than the age group of 30-40-years old.
Hypothesis 3: An individual who has a higher experienced stock return has higher financial risk tolerance.
The remainder of this research is organized as follows. Section 2 provides a review of literature on the effect of personal experiences on financial risk tolerance. Section 3 reports more detail on the underlying data and econometric methodologies. Section 4 presents the results. Section 5 concludes.
2. Literature review
A recent paper by Malmendier and Nagel (2011) shows that an individual’s risk taking is affected by the experience of financial markets that they have made over longer time spans. The main finding is that individuals, who have experienced low stock market returns throughout their lives, have lower willingness to take financial risk, are less likely to participate in the stock market. This evidence contradicts the assumption of standard neo-classical models that assume economic agents have stable risk preferences, and leads to a literature that studies the effect of the economic environment and personal experiences on the formation of preferences and economic behaviour.
4
there to last, but less drastic inflation experiences erode after 10 years (Ehrmann and Tzamourani, 2012).
Experience of recession also matters for future preferences. Having grown up during a recession causes individuals to be more likely to hold the belief that success in life depends more on luck than on effort, and therefore individuals are likely to have a more favourable attitude towards re-distributional policies (Alesina and Giuliano, 2011; Giuliano and Spilimbergo, 2009). Graham and Narasimhan (2004) also find that managers who experienced the Great Depression choose a more conservative capital structure with less leverage.
Also, a personal experience of financial market performance sharpens an individual’s behaviour. Kaustia and Knüepfer (2008) examine a positive link between investors’ experience on their investments in initial public offerings (IPO) on the stock market and their future IPO subscriptions. Moreover, Choi et al. (2009) indicate that personal experienced return plays an important role in making investors’ savings decisions. For instance, high personal experienced returns induce higher savings rates.
Experiences of rare events, such as the financial crisis, also have an impact on individual’s beliefs and behaviours. Friedman and Schwartz (1963) find that pessimism created by the Great Depression had persistent effects on markets. Malmendier and Tate (2005) also state that corporate managers who are “Depression babies2” avoid from external financing. Subsequently, Cogley and Sargent (2008) explain the equity premium with a model that assumes that the Great Depression created a long-lasting shift toward pessimistic beliefs. Furthermore, Necker and Ziegelmeyear (2013) identify that suffering wealth shock affects risk-taking via return expectations.
In terms of socio-economic background, Dohmen et al. (2011) state that parental educational background affects willingness to take risks. That implies that parental education is also an important factor that influences young investors’ risk tolerance, which is why this
2
5
study controls the variable of education during the research. Subsequently, Guiso et al. (2004) show that in high-social-capital areas in Italy (measured by electoral turnout and blood donations), more households invest in stocks; for movers, social capital in the area of original birth remains relevant. Finally, Alesina and Fuchs-Schündeln (2007) demonstrate that persistent effects of communism on attitudes towards the role of the state in providing social services, insurance or redistribution through analyzing the data on German households.
It cannot be denied that individual’s beliefs and behaviour are closely related to individual experiences. But how long do these conditions persist is a good question. On this issue, the research done by Malmendier and Nagel (2011) gives a good answer. They use the 1960-2007 SCF to identify the impact of financial market experience on individual’s risk tolerance, and find that more recent experiences are relatively more important than more distant experiences, though their impact remains noticeable for some decades. Malmendier and Nagel’s findings also suggest that young individuals are particularly affected by more recent events.
6
3. Data and Methodology
3.1 Data
This study combines the 1989–2010 SCF datasets and historical data of US stock market returns. The standard SCF is obtained from the Board of Governors of the Federal Reserve System and available every three years. The survey not only provides various household background characteristics, but also includes information about households’ perceptions of their financial situations. This study applies a set of population weights to ensure the representativeness of the sample in the following calculations and analyses. To test the hypotheses, the study sample includes household heads in 20-50 years old age range in each of the 1989-2010 SCF datasets. The total sample size of this study is 11,341.
This study uses the Dow Jones Industrial Average, NASDAQ Index and S&P 500 to describe the US stock market performance. US stock markets had suffered two main booms and recessions over the past two decades (see Fig. 1). The inflation adjusted price trend of the Dow Jones Industrial Average, NASDAQ Index and S&P 500 all show the similar pattern. Fig. 1 illustrates that the equity wealth increased fast during the period of 1989-2000, and declined until the year of 2003. But in the early of 2004, the equity wealth started again to grow; this rise is faster and more dramatic than the previous boom, and it dropped again and also more sharply than the previous recession until the mid of 2009.
Fig. 1. The United States business cycles from 1989 to 2012.Notes: US stock market performance was
7
A stock market boom is defined as an annual return of above 20%, and a stock market recession is defined as an annual return of below -20% (see Appendix A). As a result, there are two obvious booms and recessions of stock market during the period from 1989 to 2010. The stock market had a high return during 1989-2001 and 2004-2007, while the return of stock market dropped dramatically during 2001-2003 and 2007-2010.
This study builds two measures related to the household’s risk tolerance as dependent variables. All the data items are presented as the average value, include stock holdings, liquid assets and equities. The first financial risk tolerance measure is a binary variable for stock market participation, available in four sample periods, namely 2001, 2004, 2007 and 2010. Stock market participation is defined as whether a household holds more than $0 worth of stocks. For determining stock market participation, a household head participates in the stock market if he or she holds any stocks directly or via the equity portion of mutual funds holdings.
The second measure of risk tolerance is the fraction of liquid assets invested in stocks. Liquid assets are defined as stock holdings and bonds, cash and short-term instruments (checking and savings accounts, money market mutual funds, certificates of deposit). Liquid assets and stock holdings both are available in four sample periods (2001, 2004, 2007 and 2010). To rule out technical error and ensure the accuracy of the information of questionnaire, each survey shows five results of interviews of every household. However, this study does not use individuals’ elicited willingness to take financial risk as a measurement of household heads’ financial risk taking, since self-reported risk tolerance of household is not presented in every survey.
8
This study uses the annual real returns of the S&P 500 stock market index from 1871 and 2013 to measure the stock market experiences of each household head. The annual real returns of the S&P 500 stock market index are collected from Yahoo Finance. The real stock returns are adjusted and deflated with the consumer prices index (CPI) inflation rate.
3.2 Methodology
This study investigates the influence of past experiences on the risk attitude of households. Following Malmendier and Nagel (2011), this study synthesizes the experienced returns of a household using a weighted average of these returns conditional on a weighting parameter λ. The weighting scheme is flexible enough to allow households to give either higher or lower weights to more recently experienced returns. Specifically, for each household head i in year t, the experienced return is constructed as follows:
𝑨𝒊𝒕(𝝀) = ∑𝒂𝒈𝒆𝒌=𝟏𝒊𝒕−𝟏𝝎𝒊𝒕(𝒌, 𝝀)𝑹𝒕−𝒌 , (1)
where 𝝎𝒊𝒕 (𝒌, 𝝀) = (𝒂𝒈𝒆𝒊𝒕−𝒌) 𝝀 ∑𝒂𝒈𝒆𝒊𝒕−𝟏𝒌=𝟏 (𝒂𝒈𝒆𝒊𝒕−𝒌)𝝀
, (2)
where Rt-k is the stock market return in year t-k, t is the reference period of the survey; k is the
birth year of the household. The weights 𝜔it (k, λ) depend on the age of the household head
and a weighting parameter λ at time t (ageit). A weighting parameter λ determines the shape of
the weighting function and the steepness of the slope. Malmendier and Nagel (2011) plot possible weights 𝜔it (k, λ) for the example of a 50-year-old household head as a function of
9
The following regression model indicates how this study estimates the household head’s sensitivity to experienced returns:
𝒚𝒊𝒕= 𝜶𝒕+ 𝜷𝑨𝒊𝒕( 𝝀) + 𝜹𝒙𝒊𝒕+ 𝜺𝒊𝒕, (3)
where yit is the measure of household head’s risk tolerance, the variable indicating whether or not a household head participates in the stock markets, or the fraction of liquid assets invested in stocks. xit denotes a vector of control variables, εit is residual and Ait (λ) denotes experienced returns. Depending on Malmendier and Nagel (2011), the value of weighting parameter λ is supposed λ = 1.5 to apply for all tests in this study. Setting λ = 1.5 in all tests is easy to compare test results and analyze data. To capture the experience effects of a single period of boom or recession of the stock market on a household head’s risk taking, this study chooses the household heads who are above 20-year-old and below than 55-year-old. The following four models for four periods are based on empirical model (3): 2001: 𝒚𝟏= 𝜶𝟏+ 𝜷𝟏𝑨𝒊(𝝀) + 𝜹𝒙𝒊+ 𝜺𝒊, (4)
2004: 𝒚𝟐 = 𝜶𝟐+ 𝜷𝟐𝑨𝒊(𝝀) + 𝜹𝒙𝒊+ 𝜺𝒊, (5)
2007: 𝒚𝟑 = 𝜶𝟑+ 𝜷𝟑𝑨𝒊(𝝀) + 𝜹𝒙𝒊+ 𝜺𝒊, (6)
2010: 𝒚𝟒 = 𝜶𝟒+ 𝜷𝟒𝑨𝒊(𝝀) + 𝜹𝒙𝒊+ 𝜺𝒊. (7)
Ordinary linear regression requires the dependent variable to be continuous. When assessing the impact of experiences on the fraction of liquid assets invested in stocks of households, this study uses ordinary least squares method. This study employs a binary logit model when looking at the changes of stock market participation caused by experience returns. The binary logit model is advantageous because it is appropriate when the response takes one of only two possible values representing the presence or absence of an attribute of interest.
3.3 Descriptive Statistics
10 Table 1
Descriptive statistics. Notes: Observations are weighted by SCF sample weights. In terms of age filtered, 1=20-30-year-old, 2=30-40-year-old and 3=40-50-year-old. In terms of education level, 2=high school diploma, 3=some college, 4=college degree. Income, liquid assets and stock holding all are measured by US dollar ($).
10th pct Median 90th pct Mean Std.dev. #Obs.
Panel A: All Households
Age Filtered 1 2 3 2.332 0.854 11,341
Education Level 2 3 4 3.000 1.058 11,341
Income 17,281 68,106 657,597 616,758 4,031,779 11,341
Liquid Assets 11 4,901 163,718 189,264 1,702,739 11,341
Stock holding 0 0 157,152 697,244 7,815,278 11,341
Fraction of liquid assets invested in stocks 0 0 0.762 0.157 0.303 10,442 Stock market participation 0 0 0.797 0.190 0.325 7,056 Experienced real stock return (λ=1.5) 0.075 0.093 0.124 0.101 0.018 11,341
Panel B: Households in 2001 Age Filtered 1 2 3 2.853 0.815 2,728 Education Level 2 3 4 3.090 1.051 2,728 Income 20,660 90,703 1,259,764 733,790 3,208,072 2,728 Liquid Assets 184 9,682 272,374 283,042 2,123,665 2,728 Stock holding 0 0 857,754 1,587,838 13,391,727 2,728
Fraction of liquid assets invested in stocks 0 0 0.207 0.239 0.359 2,570 Stock market participation 0 0.014 0.906 0.246 0.344 1,905 Experienced real stock return (λ=1.5) 0.074 0.097 0.120 0.097 0.017 2,728
Panel C: Households in 2004 Age Filtered 1 2 3 2.198 0.783 2,541 Education Level 2 3 4 3.024 1.059 2,541 Income 17,726 70,905 633,415 592,752 3,522,998 2,541 Liquid Assets 1 4,810 161,051 164,495 1,046,820 2,541 Stock holding 0 0 156,488 519,009 5,588,032 2,541
Fraction of liquid assets invested in stocks 0 0 0.756 0.165 0.305 2,317
Stock market participation 0 0 0.806 0.203 0.325 1,572
Experienced real stock return (λ=1.5) 0.092 0.092 0.119 0.097 0.011 2,541
Panel D: Households in 2007 Age Filtered 1 2 3 2.179 0.794 2,408 Education Level 2 3 4 3.015 1.059 2,408 Income 18,100 71,216 726,169 787,871 4,752,148 2,408 Liquid Assets 10 5,260 167,399 221,645 2,497,671 2,408 Stock holding 0 0 83,815 608,682 6,483,243 2,408
Fraction of liquid assets invested in stocks 0 0 0.695 0.142 0.288 2,218
Stock market participation 0 0 0.752 0.167 0.304 1,560
11
10th pct Median 90th pct Mean Std.dev. #Obs.
Panel E: Households in 2010 Age Filtered 1 2 3 2.137 0.812 3,364 Education Level 2 3 4 2.946 1.059 3,364 Income 15,248 55,705 271,408 433,815 4,376,325 3,364 Liquid Assets 1 2,739 86,630 115,340 863,680 3,664 Stock holding 0 0 20,000 215,971 2,335,314 3,664
Fraction of liquid assets invested in stocks 0 0 0.485 0.098 0.245 3,337
Stock market participation 0 0 0.629 0.143 0.312 2,019
Experienced real stock return (λ=1.5) 0.075 0.087 0.124 0.088 0.012 3,664 Comparing Panel A and B, not only stock market participants tend to be wealthier than
the average household, with a median holding of $283,042 in liquid assets rather than $189,264 in the full sample, but also the amount of stock holding (the mean of stock holding is $1,587,838) is higher than that’s in full sample. Actually, there is a similar situation that comparing Panel A and D. In contrast, Panel C and E both show the different results, for example, income level, liquid assets and stock holding on average all tend to a lower level. Panel A also shows that a wide wealth disparity in America. The income in Panel A has 10th and 90th percentiles of $17,281 and $657,597.
Panel A shows 19.0% of households participate on average in the stock market during 1989-2010. These rates represent the US population (not the SCF sample). Every SCF shows different numbers of households that are interviewed (as the time increases, the amount of household also increase). To ensure the selected sample is representative, this study adds the analysis of 10th and 90th percentiles of observations into the descriptive statistics.
12
start to participate in the stock market following an upward tendency, they could get high experienced stock market return, vice versa. Malmendier and Nagel (2011) also find that the mean stock market participant is 0.342 over their sample period. But in this study, the mean stock market participant is 0.190 in full sample (total observation is 11,341) with four sample periods, which is lower than the long-term period. That implies that recent financial events have higher impact on household’ behaviour, more and more investors are less willingness to invest money in stock market.
5. Results
5.1 Fraction of Liquid Assets Invested in Stocks
Table 2 shows the results of effect of experienced stock returns on the fraction of liquid assets that household heads invested in stocks. Following Malmendier and Nagel (2011), λ=1.5 is set in the following linear regression model:
𝒚𝒊𝒕= 𝜶𝒕+ 𝜷𝑨𝒊𝒕( 𝝀) + 𝜹𝒙𝒊𝒕+ 𝜺𝒊𝒕, (8) where yit is the fraction of liquid assets invested in stocks. This study estimates the model with
least squares. Table 2
Fraction of liquid assets invested in stocks.Notes: t-statistics are reported in brackets, ***, **, and * stand for p < 0.001, p < 0.01, and p < 0.1, respectively.
2001 2004 2007 2010
Experienced return coefficient β 0.026 0.353 0.656 0.224
(0.059) (0.566) (1.954) (0.883)
Weighting parameter λ Fixed Fixed Fixed Fixed
1.5 1.5 1.5 1.5
Income controls Yes Yes Yes Yes
Liquid assets controls Yes Yes Yes Yes
Household characteristics Yes Yes Yes Yes
13
Table 2 shows that with the liquid assets controls, the experienced stock market returns do not have a statistically significant effect on the percentage invested in stocks in all age groups. Although the effect of experienced stock return on changing household’s financial risk tolerance is not significant, this test is still meaningful, since the p-values of tests of each sample period are all statistically significant. Additionally, the coefficient of the impact of experienced stock return in 2007 is estimated at 0.656, which has the strongest impact on the household portfolio choice in sample period, but it is lower than what was found by Malmendier and Nagel (2011) (1.734) from 1960 to 2007. Malmendier and Nagel (2011) use the data include both of the young and the old, and the sample period is longer than in this study, which focuses on young household heads.
With income controls, liquid assets controls and household characteristics controls, age also shows a statistically significant and positive effect on the percentage invested in stocks. However, the four years’ coefficients are similar in this study, which are around 0.04, and the p-values are all lower than 0.001. That implies age is an important factor of affecting the fraction of liquid assets invested in stocks, and it is worth to study. Similarly, the control variable of education level also plays an essential part in this study. Dohmen et al (2011) find parental educational background affects willingness to take risks. This study also finds the education level affects household heads’ financial risk tolerance. But knowledge of investments cannot be simply equated with an individual’s level of education. For instance, a bachelor’s degree holder in finance that has invested for 20 years should know a lot more about investing than a Ph.D. degree holder in another field who only started investing a few months ago. Those with modest knowledge about and very little experiences with investments may have a very different perception of financial risks as compared with those who are familiar with investing and have been investing for a long time.
14
combination of a person’s wealth and his or her financial knowledge. At last, he suggests that financial behaviours may be driven more by a person’s ability to take on financial risks rather their willingness to take on financial risk. The old is more likely to have stronger financial behaviours than the young, such as own a home, a car3 or have good credit card behaviours. In those circumstances, this study minimizes the influence of age and control the education level of household heads.
Table 2 also estimates that the control variables of gender and race both have a significant effect on invested in stock, but there are negative. Although the test results of this study present that the relationship between the financial risk tolerance and marital status is not apparent, these two variables are still correlated. There are thousands of research studies and hundreds of books on studying the relationship between the financial risk tolerance and demographic characteristics. Yao et al (2004) find that characteristics such as gender, marital status and race that have a significant negative relationship with risk tolerance. Yao et al (2005) further identify that Blacks and Hispanics are less likely to be willing to take some financial risk but are more likely to be willing to take substantial financial risk than Whites.
To capture the different effects of experienced stock market returns on personal risk taking in different age groups of the young, Table 3 illustrates the different influence degree of age groups on the percentage invested in stock in sample periods.
Table 3
Fraction of liquid assets invested in stock with different age groups. Note: ***, **, and * stand for p < 0.001, p < 0.01, and p < 0.1, respectively. 2001 2004 2007 2010 Age group (20-30) Coefficient β -0.099* -0.061*** -0.063*** -0.058*** Age group (30-40) Coefficient β -0.038* -0.005 0.022 -0.019 Age group (40-50) Coefficient β 0.09* 0.071*** 0.055** 0.063***
Income controls Yes Yes Yes Yes
Liquid assets controls Yes Yes Yes Yes
Household characteristics Yes Yes Yes Yes
3
15
This study adds the dummies in Eq. (8) to distinguish the difference in financial risk tolerance of household heads of various ages. Table 3 shows that all coefficients of the impact of experienced returns in age group of 20-30 in four sample years are negative, but the p-values all show significant results. Since the household heads are above 20 years old in the data sample, since twenties are matter in personal life (Arnett, 2000). Most people get the first job in their twenties, which means they start to save income, get a credit score and have the responsibility to face their own mistakes. Also, they have sound cognitive ability to make decisions. Meg Jay4 says, “People in twenties are more sensitive to surprise and criticism, are more likely to take feedback personally, and magnify events to enormous proportions in their minds.” That implies that for people in their twenties start to invest in stocks, suffering the boom or recession of stock return could lead to a significant influence on their investing behaviour.
Each generation experiences a unique demographic, political, and socioeconomic environment during their formative years. Differing experiences shared by a generation may contribute to dissimilar attitudes toward financial risks between those in different generations (Yao et al, 2011). According to Meg Jay’s TED5 Talk, “80% of life's most defining moments happen by age 35.” For example, individuals in 30s who have a decent job with stable income, make decisions by themselves. As stated, the middle to late 30s are often characterized by settling down, who spend much more time on advancing their careers and gaining stability in their personal lives, such as marriage and child-rearing. The experienced return may be one of the effects on the investment decision for the 30-year-olds. Thus, experienced return does not have a significant influence on household head’s risk taking in age group of 30-40.
Table 3 also illustrates that the household heads in age group (40-50) are affected significantly by the experienced returns, and the p-values in four sample periods are all significant. People in forties not only gather sufficient experience on their careers and lives
4
Dr. Meg Jay is an assistant clinical professor at the University of Virginia and maintains a private practice in Charlottesville, Virginia.
5
16
but also have stronger positive financial behaviours to invest. Generally, those people are more likely to take risk and invest more money in the stock market, but in the meanwhile, when those people suffer dramatic stock return booms or recessions, they are affected significantly.
5.2 Stock Market Participation
For testing the effect of experienced return on stock market participation, this study applies for logit model:
𝑷(𝒚𝒊𝒕= 𝟏|𝒙𝒊𝒕, 𝑨𝒊𝒕(𝝀)) = 𝜶 + 𝜷𝑨𝒊𝒕(𝝀) + 𝜹𝒙𝒊𝒕+ 𝜺𝒊𝒕. (9)
The independent variable can be defined as:
𝒚𝒊𝒕 { 𝑷 = 𝟏, 𝑻𝒉𝒆 𝒉𝒐𝒖𝒔𝒆𝒉𝒐𝒍𝒅 𝒉𝒆𝒂𝒅 𝒉𝒐𝒍𝒅𝒔 𝒎𝒐𝒓𝒆 𝒕𝒉𝒂𝒏 $𝟎 𝒊𝒏 𝒔𝒕𝒐𝒄𝒌 𝒎𝒂𝒓𝒌𝒆𝒕 𝑷 = 𝟎, 𝑶𝒕𝒉𝒆𝒓𝒘𝒊𝒔𝒆 ,
the binary indicator yit equals 1 if a household head i has any stock holdings at time t. The key
independent variable is the effect of experienced returns, 𝐴𝑖𝑡(𝜆), on the probability of stock
market participation. The vector 𝑥𝑖𝑡 includes the same income controls and household
characteristics as in the logit model. Table 4 shows the estimated influence of experienced stock returns on the stock market participation.
Table 4
Stock market participation. Note: z-statistics are reported in brackets, ***, **, and * stand for p < 0.001, p < 0.01, and p < 0.1, respectively.
2001 2004 2007 2010
Experienced return coefficient β -1.240 -2.134 3.769 3.257 (-0.369) (-0.399) (1.092) (0.921)
Weighting parameter λ Fixed Fixed Fixed Fixed
1.5 1.5 1.5 1.5
Income controls Yes Yes Yes Yes
Liquid assets controls Yes Yes Yes Yes
Household characteristics Yes Yes Yes Yes
17
Table 4 shows that the experienced returns do not have a positive and significant effect on stock market participation while controlling for liquid assets, income assets and household characteristics. But the test results on the effect of experienced returns on household heads’ stock market participation are similar with the test results are shown in Table 2. Also, the control variables like age and education level both have a significant effect on household’s stock market participation, whereas there is a negative and highly significant effect on stock market participation, including the variable of gender and race. However, the results of p=0 and p=1 indicate that the stock market participation of household heads in the full sample is decreasing. After the boom years, such as the period of 1989-2001, the household heads had experiences of high returns, and they were willing to take high risks to invest in the stock market. Subsequently, the willingness of household heads participate in the stock market is declining, since the US stock markets have suffered the subprime mortgage crisis during the period of 2001-2003. Table 4 represents the stock market participation rate of household heads in full sample has decreased from 52.7% in 2001 to 44.6% in 2004.
Although, the US stock market started to recover in early of 2004, and afterwards it experienced a small peak, such that this economic environment still could not attract more household heads to invest in the stock market. To some extent, the experienced returns have an impact on the households invest in the stock market. Table 4 also indicates that the number of household heads did not hold any stock increased from 55.4% in 2004 to 61.3% in 2007. The majority of household heads in the age group of 40-50 who suffered the boom of stock return during the period of 1989-2001 and the recession from 2001 to 2003, so that the changes of those people’s risk tolerance more or less due to the experienced returns.
18
memories of hyperinflation tend to stay in people’s minds and affect attitudes in a much more persistent manner. That explains the decrease of household heads’ participation in the stock market. The data of SCF 2010 also report the similar phenomenon that the percentages of household heads that hold any stock are dropped from 38.7% in 2007 to 31.9% in 2010.
Appendix B shows the experienced stock return of 0.121, which group of household heads has the highest participation rate (5.8%). The second highest participation rate is 5.5%, which are household heads who have experienced stock return of 0.120. After checking the working table of experienced return of each age of household head in same periods (see Appendix C), the results show that the household heads in age group of 30-40-years old and 40-50-years old are willing to take risks, especially after the boom years.
However, after recession years, the situation has been changed 55.6% of households participated in stock market in 2004, and these household heads’ ages are between 40 and 50-years old, and the stock market participation rates of the rest of other two age groups were quite different, at 12.6% for 20-30 year-olds and 31.8% for 30-40-year-olds, respectively. There are a few different possibilities as to the cause. The household heads in age group of 20-30 need much more time to familiarize themselves with the investment climate and accumulate financial wealth, while people in 30-40 may pay more attention on their personal lives and career. The similar condition exists in the rest of sample periods (2007 and 2010). Appendix B also shows that the stock market participation rates decreased dramatically in each age group from 2001 to 2004, even household in the age group of 20-30 rarely held any stock in 2010.
6. Conclusions
19
The results of this study do not imply the same level of significance the experienced stock market returns has on a household head’s financial risk tolerance in all age groups. However, when this study looks at different age groups, the results show different. The experienced stock market returns have a greater impact on 20-30-year-olds and 40-50-year- olds than the 30-40-year-olds. The household heads in their 40s who have higher experienced stock market returns have higher risk tolerance. The household heads in their 40s, who have stronger financial ability to invest in stock markets, are more likely to participate in stock markets, have a higher willingness to invest in stock market and invest a higher fraction of their liquid assets in stocks if they participate. The household heads in their 20s, who may have just started to know what finance is and try to invest in stock markets, may be significantly affected by their experienced stock market returns. However, the results do not show the significant effect of experienced stock market returns on the financial risk tolerance of household heads in their 30s, since that age group may pay more attention on their personal lives and careers.
This study also provides evidence to support that household characteristics such as gender and race that have a significantly negative relationship with risk tolerance. Other characteristics that have a significant positive relationship with risk tolerance, such as the effect of education level, may also have a significant relationship with risk tolerance.
20
References
Alesina, A., Giuliano, P., 2011. Preferences for redistribution. In: Bisin, A., Benhabib, J. (Ed.), Handbook of Social Economics. North Holland, Amsterdam, pp. 93-132.
Alesin, A., Fuchs-Schündeln, N., 2007. Goodbye Lenin (or not?): The effect of communism on people. American Economic Review 97, 1507-1528.
Ameriks, J., Zeldes, S., 2004. How do household portfolio shares vary with age?. Working paper. Columbia University.
Arnett, J., 2000. Emerging adulthood: A theory of development from the late teens through the twenties. American psychologist 55, 469.
Barber, B., Odean, T., 2008. All that glitters: The effect of attention and news on the buying behaviour of individual and institutional investors. Review of Financial Studies 21, 785-818.
Barro, R., 2006. Rare disasters and asset markets in the twentieth century. The Quarterly Journal of Economics 121, 823-866.
Bekaert, G., Hoerova, M., 2013. The VIX, the variance premium and stock market volatility. National Bureau of Economic Research.
Blanchflower, D., 2007. Is unemployment more costly than inflation?. National Bureau of Economic Research.
Brunnermeier, M., Nagel, S., 2006. Do wealth fluctuations generate time-varying risk aversion? Micro-evidence on individuals' asset allocation. National Bureau of Economic Research.
Choi, J., Laibson, D., Madrian, B., Metrick, A., 2009. Reinforcement learning and savings behaviour. The Journal of finance 64, 2515-2534.
Cogley, T., Sargent, T., 2008. The market price of risk and the equity premium: A legacy of the Great Depression?. Journal of Monetary Economics, 55, 454-476.
Dohmen, T., Falk, A., Huffman, D., Sunde, U., Schupp, J., Wagner, G., 2011. Individual risk attitudes: Measurement, determinants, and behavioural consequences. Journal of the European Economic Association 9, 522-550.
Døskeland, T., Hvide, H., 2011. Do individual investors have asymmetric information based on work experience?. The Journal of Finance 66, 1011-1041.
21
Shum, P., Faig, M., 2006. What explains household stock holdings?. Journal of Banking and Finance 30, 2579-2597.
Friedman, M., Schwartz, A., 2008. A monetary history of the United States, 1867-1960. Princeton University Press.
Giuliano, P., Spilimbergo, A., 2009. Growing up in a recession: Beliefs and the macroeconomy. National Bureau of Economic Research.
Grable, J., 2000. Financial risk tolerance and additional factors that affect risk taking in everyday money matters. Journal of Business and Psychology 14, 625-630.
Grable, J., 1997. Investor risk tolerance: Testing the efficacy of demographics as differentiating and classifying factors. Virginia Polytechnic Institute and State University.
Graham, J., Harvey, C., Puri, M., 2013. Managerial attitudes and corporate actions. Journal of Financial Economics 109, 103-121.
Graham, J., Hazarika, S., Narasimhan, K., 2011. Financial distress in the Great Depression. Financial Management 40, 821-844.
Graham, J., Narasimhan, K., 2004. Corporate survival and managerial experiences during the Great Depression. Manuscript, Duke University.
Guiso, L., Sapienza, P., Zingales, L., 2000. The role of social capital in financial development. National Bureau of Economic Research.
Hertwig, R., Barron, G., Weber, E., Erev, I., 2004. Decisions from experience and the effect of rare events in risky choice. Psychological Science 15, 534-539.
Jay, M., 2012. The Defining Decade: Why Your Twenties Matter--And How to Make the Most of Them Now. Grand Central Publishing, New York.
Kaustia, M., Knüpfer, S., 2008. Do investors overweight personal experience? Evidence from IPO subscriptions. Journal of Finance 63, 2679-2702.
Lerner, J., Small, D., Loewenstein, G., 2004. Heart strings and purse strings carryover effects of emotions on economic decisions. Psychological Science 15, 337-341.
Lombardelli, C., Saleheen, J., 2003. Public expectations of UK inflation. Bank of England Quarterly Bulletin 43, 281-290.
22
Malmendier, U., Nagel, S., 2009. Learning from inflation experiences. Unpublished manuscript. University of California, Berkeley.
Malmendier, U., Tate, G., Yan, J., 2011. Overconfidence and early-life experiences: The effect of managerial traits on corporate financial policies. Journal of Finance 66, 1687-1733.
Malmendier, U., Nagel, S., 2011. Depression babies: Do macroeconomic experiences affect risk-taking? Quarterly Journal of Economics 126, 373-416.
Malkiel, B., 1996. A Random Walk Down Wall Street. Norton. New York.
Necker, S., Ziegelmeyer, M., 2013. Risk taking after the financial crisis. Mimeo, University of Freiburg.
Nisbett, R., Ross, L., 1980. Human inference: Strategies and shortcomings of social judgment. Prentice Hall, Englewood Cliffs, New Jersey.
Ricciardi, V., Simon, H., 2000. What is behavioural finance? Business Education and Technology Journal 2, 1-9.
Sahm, C., 2007. Stability of risk preference. Working paper. Division of Research and Statistics and Monetary Affairs, Federal Reserve Board.
Senbet, L., Gande, A., 2009. Financial Crisis and Stock Markets: Issues, Impact, and Policies. The Dubai Economic Council.
Sung, J., Hanna, S., 1996. Factors related to risk tolerance. Financial Counselling and Planning 7, 11-20.
Vissing-Jorgensen, A., 2004. Perspectives on Behavioural Finance: Does" Irrationality" Disappear with Wealth? Evidence from Expectations and Actions. In: NBER Macroeconomics Annual 2003, Volume 18, pp. 139-208.
Wagner, J., 2014. An empirical analysis linking a person’s financial risk tolerance and financial literacy to financial behaviours.
Yao, R., Sharpe, D., Wang, F., 2011. Decomposing the age effect on risk tolerance. The Journal of Socio-Economics 40, 879-887.
Yao, R., Gutter, M., Hanna, S., 2005. The financial risk tolerance of Blacks, Hispanics and Whites. Financial Counselling and Planning 16, 51-62.
23
A
PPENDIXA
THE ANNUAL REAL RETURNS OF THE S&P 500 FROM1871 TO 2013
Year S&P 500 CPI Inf. ADJ S&P 500 Year S&P 500 CPI Inf. ADJ S&P 500
24
A
PPENDIXA
(CONTINUED)
Year S&P 500 CPI Inf. ADJ S&P 500 Year S&P 500 CPI Inf. ADJ S&P 500
1937 -32.11% 1.89% -34.00% 1899 3.66% 14.43% -10.77% 1936 32.55% 1.89% 30.66% 1898 29.32% 1.90% 27.42% 1935 54.93% 4.49% 50.44% 1897 20.37% 0.00% 20.37% 1934 -8.01% 1.37% -9.38% 1896 3.25% -1.54% 4.79% 1933 56.79% 1.19% 55.60% 1895 5.01% 3.09% 1.92% 1932 -5.81% -10.79% 4.98% 1894 3.63% -6.28% 9.91% 1931 -44.20% -5.73% -38.47% 1893 -18.79% -6.96% -11.83% 1930 -22.72% -5.28% -17.44% 1892 6.14% 1.40% 4.74% 1929 -9.46% 0.53% -9.99% 1891 18.88% -6.34% 25.22% 1928 47.57% -1.73% 49.30% 1890 -6.16% 1.19% -7.35% 1927 37.10% -3.17% 40.27% 1889 7.09% -6.86% 13.95% 1926 11.51% -1.26% 12.77% 1888 3.34% 0.00% 3.34% 1925 25.83% 4.22% 21.61% 1887 -0.64% 5.99% -6.63% 1924 27.10% 0.00% 27.10% 1886 11.98% -5.74% 17.72% 1923 5.45% 2.44% 3.01% 1885 30.06% -1.59% 31.65% 1922 29.07% -3.05% 32.12% 1884 -12.32% -9.51% -2.81% 1921 10.15% -13.37% 23.52% 1883 -5.49% -8.22% 2.73% 1920 -13.95% 2.22% -16.17% 1882 3.61% -2.07% 5.68% 1919 19.67% 15.20% 4.47% 1881 0.27% 6.88% -6.61% 1918 18.21% 20.06% -1.85% 1880 26.63% -2.67% 29.30% 1917 -18.62% 12.47% -31.09% 1879 49.37% 23.10% 26.27% 1916 8.12% 12.12% -4.00% 1878 16.29% -18.44% 34.73% 1915 31.20% 2.55% 28.65% 1877 -1.06% -13.54% 12.48% 1914 -5.39% 0.94% -6.33% 1876 -14.15% -0.79% -13.36% 1913 -4.73% 2.86% -7.59% 1875 5.44% -5.80% 11.24% 1912 7.18% 7.73% -0.55% 1874 4.72% -6.37% 11.09% 1911 3.52% -2.30% 5.82% 1873 -2.49% -5.59% 3.10% 1910 -3.39% -8.40% 5.01% 1872 11.16% 1.72% 9.44% 1909 16.15% 11.61% 4.54% 1871 15.64% 1.82% 13.82% 1908 39.47% 3.10% 36.37% 1907 -24.21% -0.86% -23.35% 1906 0.64% 4.52% -3.88% 1905 21.29% 0.00% 21.29% 1904 32.16% 6.22% 25.94% 1903 -17.09% -5.12% -11.97% 1902 8.28% 7.55% 0.73% 1901 19.45% 5.97% 13.48% 1900 20.84% -4.77% 25.61%
25
A
PPENDIXB
STOCK MARKET PARTICIPATION WITH DIFFERENT AGE GROUPS
Notes: The overall results of households participate in stock markets with three age groups from 2001 to 2010, excluding the date of households who do not participate in stock markets, and applying for population weights.
2001 2004 2007 2010
AGE AIT % AGE AIT % AGE AIT % AGE AIT %
2 0.121 5.8% 2 0.092 31.0% 3 0.123 11.8% 3 0.125 11.2% 3 0.120 5.5% 3 0.092 26.5% 3 0.110 8.6% 3 0.118 7.5% 2 0.080 5.0% 1 0.092 12.4% 3 0.124 6.3% 2 0.094 7.0% 3 0.117 4.5% 3 0.123 6.8% 3 0.121 6.0% 3 0.119 6.7% 3 0.119 4.4% 3 0.119 6.7% 3 0.116 5.8% 3 0.115 6.4% 2 0.074 4.3% 3 0.108 5.3% 3 0.126 5.8% 3 0.121 5.0% 3 0.113 4.3% 3 0.116 5.1% 2 0.083 5.5% 3 0.126 5.0% 2 0.075 4.2% 3 0.122 4.7% 3 0.118 5.5% 3 0.111 4.7% 2 0.089 4.2% 2 0.071 0.3% 2 0.074 5.0% 2 0.074 4.5% 1 0.071 4.0% 2 0.087 0.3% 2 0.095 4.5% 3 0.129 4.5% 1 0.101 4.0% 3 0.094 0.3% 2 0.080 4.3% 2 0.120 4.0% 1 0.081 3.8% 2 0.075 0.1% 3 0.099 4.0% 3 0.124 3.9% 1 0.074 3.7% 3 0.080 0.1% 2 0.091 3.5% 2 0.092 3.7% 2 0.082 3.6% 1 0.103 0.1% 2 0.075 3.0% 2 0.096 3.3% 2 0.093 3.6% 2 0.103 0.1% 2 0.088 2.8% 2 0.081 2.5% 3 0.107 3.5% 3 0.093 2.8% 1 0.070 2.3% 2 0.097 3.3% 2 0.106 2.5% 1 0.074 2.2% 2 0.113 3.3% 1 0.090 2.0% 1 0.083 2.2% 3 0.116 3.2% 1 0.104 1.8% 1 0.114 2.2% 2 0.091 3.1% 1 0.108 1.8% 1 0.089 1.6% 3 0.115 3.0% 1 0.071 1.7% 1 0.148 1.6% 1 0.086 2.6% 1 0.102 1.7% 2 0.084 1.4% 1 0.087 2.6% 1 0.062 1.3% 1 0.091 1.4% 3 0.111 2.3% 1 0.140 1.0% 1 0.105 1.4% 1 0.096 2.0% 1 0.125 0.8% 1 0.107 1.4% 1 0.099 1.8% 1 0.064 0.5% 2 0.108 1.2% 1 0.104 1.8% 1 0.093 0.8% 1 0.128 1.7% 1 0.132 0.8% 1 0.114 1.2%
Total 100.0% Total 100.0% Total 100.0% Total 100.0%
1 (20-30) 292 29.1% 1 (20-30) 88 12.6% 1 (20-30) 76 12.6% 1 (20-30) 114 17.7% 2 (30-40) 404 40.3% 2 (30-40) 223 31.8% 2 (30-40) 187 31.0% 2 (30-40) 178 27.6% 3 (40-50) 307 30.6% 3 (40-50) 390 55.6% 3 (40-50) 341 56.4% 3 (40-50) 352 54.7%
Total 1003 100.0% Total 701 100.0% Total 604 100.0% Total 644 100.0%
P=1 52.7% P=1 44.6% P=1 38.7% P=1 41.3%
26
A
PPENDIXC
WORKING TABLE OF EXPERIENCED RETURNS OF EACH AGE OF HOUSEHOLD HEAD
2001 2004 2007 2010 2001 2004 2007 2010
Age Weights Experienced Return Age Weights Experienced Return
27
A
PPENDIXC
(CONTINUED)
2001 2004 2007 2010
Age Weights Experienced Return
28 148.162 0.126 0.062 0.125 0.114 27 140.296 0.147 0.063 0.090 0.105 26 132.575 0.114 0.106 0.108 0.148 25 125.000 0.085 0.135 0.062 0.132 24 117.576 0.080 0.160 0.064 0.093 23 110.304 0.065 0.123 0.113 0.113 22 103.189 0.042 0.091 0.146 0.063 21 96.234 0.043 0.086 0.176 0.064 20 89.443 0.008 0.070 0.135 0.121 19 82.819 -0.022 0.043 0.100 0.160 18 76.368 -0.003 0.045 0.096 0.197 17 70.093 -0.010 0.004 0.078 0.153 16 64.000 -0.008 -0.033 0.047 0.113 15 58.095 0.043 -0.011 0.050 0.110 14 52.383 0.058 -0.021 0.001 0.091 13 46.872 0.013 -0.022 -0.045 0.056 12 41.569 0.029 0.040 -0.020 0.062 11 36.483 0.054 0.058 -0.035 0.001 10 31.623 0.069 0.000 -0.039 -0.058 9 27.000 0.072 0.018 0.038 -0.029 8 22.627 0.079 0.048 0.064 -0.053 7 18.520 0.091 0.067 -0.018 -0.065 6 14.697 -0.027 0.071 0.001 0.041 5 11.180 0.081 0.082 0.041 0.083 4 8.000 0.160 0.106 0.072 -0.050 3 5.196 0.113 -0.123 0.083 -0.039 2 2.828 -0.087 0.027 0.129 0.011 1 1.000 0.007 0.213 0.364 0.058 0 0.000 0.000 0.000 0.000 0.000
Notes: Following the approach by Malmendier and Nagel (2011) to get the experienced stock market return of each household head. Setting the value of parameter λ = 1.5 in the following formulas:
𝑨𝒊𝒕(𝝀) = ∑𝒂𝒈𝒆𝒌=𝟏𝒊𝒕−𝟏𝝎𝒊𝒕(𝒌, 𝝀)𝑹𝒕−𝒌 , where 𝝎𝒊𝒕 (𝒌, 𝝀) = (𝒂𝒈𝒆𝒊𝒕−𝒌)𝝀
∑𝒂𝒈𝒆𝒊𝒕−𝟏𝒌=𝟏 (𝒂𝒈𝒆𝒊𝒕−𝒌)𝝀
,
where Rt-k denotes the stock market return in year t-k, t is the reference period of the survey; k is the birth year of the household.
