320 likes | 644 Views
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE. Barry Strydom, Avish Gokul, & Andrew Christison School of Economics & Finance University of Kwazulu-Natal JANUARY 2009 Presented at the 2009 SAFA Conference, Cape Town. FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE.
E N D
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE Barry Strydom, Avish Gokul, & Andrew Christison School of Economics & Finance University of Kwazulu-Natal JANUARY 2009 Presented at the 2009 SAFA Conference, Cape Town
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE • Modern Portfolio Theory An investor’s optimal portfolio represents the combination of available investment opportunities that will maximise his/her utility function for risk and return (Bodie, Kane & Marcus, 2007: 181; Yook & Everett, 2003: 1)
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE • An investor’s relative risk aversion is thus critical in determining optimal portfolio allocations (Viceria, 1999: 59; Campbell & Viceria, 2002: 11; Kapteyn & Teppa, 2002: 35; Henderson, 2005: 9) • “Thus, in terms of investment advice, identification of an investor’s risk tolerance determines the asset allocation decision … In order to satisfy the ‘know your client rule’, financial advisors should have some conception of an investor’s psychological comfort and tolerance towards investments with uncertain outcomes.” (Subedar, McCrae & Gerace, 2006: 6,9)
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE • Risk Tolerance • Inverse of Risk Aversion Walls & Dyer, 1996: 1007; Barsky, Juster, Kimball & Shapiro, 1997: 542; Gron & Winton, 2001: 593; Yao, Gutter & Hanna; 2005: 52
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE • Demographic Determinants • Gender Grable & Joo(1999); Hanna, Gutter & Fan (1998) – not statistically significant Barsky, et al (1970) ; Halek and Eisenhauer, (2001); Bajtelsmit, Bernasek and Jianakoplos (1999); Miller and Hoffmann, (1995); Cohn, Lewellen, Leasea and Schlarbaum, (1975); Powell and Ansic, (1997) and Baker and Haslem, (1974). • Race Yao, Gutter & Hanna (2005); Barsky et al (1997) ; Sung & Hanna (1996) • Religion Halek & Eisenhauer (2001); Barsky et al (1997) • Wealth Yook and Everett (2003); Baltelsmit et al, (1999) • Age Riley & Chow (1992); Wang & Hanna (1997) • Marital status Hanna & Lindamood (2005); Halek & Eisenhauer (2001); Sung & Hanna (1996)
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE Hypotheses 1. There is no difference between the results of the UKZN study and the the Hanna & Lindamood study 2. There is no difference in the results of the two questions measuring risk aversion 3.a There is no difference in the results for males between the two questions measuring risk aversion 3.b There is no difference in the results for females between the two questions measuring risk aversion 4. There is no difference between the risk tolerance of males and females 5. There is no difference in risk tolerance across race 6. There is no difference in risk tolerance across religion 7. Risk tolerance is correlated to income
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE • Measuring Financial Risk Tolerance • Actual Behaviour • Investment Type Questions • Subjective Questions • Combination
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE • Methodology Barsky, Juster, Kimball & Shapiro (1997) “Suppose that you are the only income earner in the family, and you have a good job guaranteed to give you your current (family) income every year for life. You are given the opportunity to take a new and equally good job, with a 50-50 chance it will double your (family) income and a 50-50 chance it will cut your (family) income by a third. Would you take the new job?”
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE Barsky et al : respondents might value current job for reasons other than income. Further concern raised by Hanna et al (2001): • Hypothetical income decreases are permanent • Developed alternative set of questions based on pension choices Hanna & Lindamood (2004) • Concern that hypothetical income choices too complex for many respondents to fully understand • Therefore developed a graphical version (please refer to handout)
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE SAMPLE The H&L study comprised of 152 observations obtained from students at Ohio State University 74% male, mean age of 23 This study surveyed 3rd & 4th year Accounting and Finance students on the Pmb campus of UKZN. In total 84 responses were obtained 54% male, mean age of 23
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE FINDINGS: Hypothesis 1 The results of the UKZN study correspond to the Hanna & Lindamood study Hanna & Hanna Lindamood et al UKZN Extremely Low Risk Aversion (A< 1.0) 2 1 7 Very Low Risk Aversion (1.0 ≤ A < 2.0) 16 5 6 Low Risk Aversion (2.0 ≤ A < 3.8) 36 22 32 Moderate Risk Aversion (3.8 ≤ A < 7.5) 37 44 25 High Risk Aversion (7.5 ≤ A < 9.3) 5 10 10 Very High Risk Aversion (9.3 ≤ A ≤ 14.5) 2 7 13 Extremely High Risk Aversion (A > 14.5) 2 11 7
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE Hanna & Lindamood Mean Risk Aversion = 4.4 Hanna et al Mean Risk Aversion = 6.6 UKZN study Mean Risk Aversion = 6
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE The Mean level of risk aversion for the UKZN respondents is significantly higher than the corresponding Hanna & Lindamood sample. P-value 0.001027813 Chi-square 22.3922182
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE FINDINGS: Hypothesis 2 There is no difference in the results of the two questions measuring risk tolerance
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE FINDINGS: Hypothesis 3 3.a There is no difference in the results for males between the two questions measuring risk aversion
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE FINDINGS: Hypothesis 3 3.b There is no difference in the results for females between the two questions measuring risk tolerance
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE Significant correlation for males but surprisingly low. No correlation for females. Powell & Ansic (1997: 609) – differences in risk preference could be explained by the way tasks were framed Schubert, Brown, Gysler & Brachinger (1999: 384) “the comparative risk propensity of male and female subjects in financial choices strongly depends on the decision frame” The implication is that framing is important in measuring risk tolerance and that the effect is greater for women than men.
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE FINDINGS: Hypothesis 4 There is no difference between the risk tolerance of males and females
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE 8: Mann-Whitney on gender (FRT 2) Test Statistics(a) Table 9: Mann-Whitney on gender (FRT 1) Test Statistics(a) a Grouping Variable: Gender of respondent
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE Kruskal-Wallis test Table 10: Crosstabulation of racial groups against FRT 2 Ranks FINDINGS: Hypothesis 5 There is no difference in risk aversion across race The Chi-Square measure = 9,748 (3 degrees of freedom) which is significant at the 1% level.
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE Table 11: Median analysis of racial groups Frequencies FINDINGS: Hypothesis 5 There is no difference in risk aversion across race
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE Table 12.a: Mann-Whitney comparing Whites and Indians Test Statistics(a) a: Grouping Variable: Racial group to which respondent belongs Table 12.b: Mann-Whitney comparing Whites and Blacks Test Statistics(a) a: Grouping Variable: Racial group to which respondent belongs
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE Table 13: Crosstabulation of religious groups against FRT 2 Ranks FINDINGS: Hypothesis 6 There is no difference in risk aversion across religion Table 14: Kruskal-Wallis test against religion Test Statistics(a,b) a Kruskal Wallis Test b Grouping Variable: Religious group to which respondent belongs
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE Table 15: Median analysis of religious groups Frequencies FINDINGS: Hypothesis 6 There is no difference in risk aversion across religion Table 16: Mann-Whitney comparing Christians and Hindus Test Statistics(a) a Grouping Variable: Religious group to which respondent belongs
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE Table 17: Crosstabulation of FRT 2 against household income Financial Risk Tolerance 2 against Category of household Income FINDINGS: Hypothesis 7 Risk aversion is correlated to income
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE Table 18: Correlation of FRT 2 and household income Correlations FINDINGS: Hypothesis 7 Risk aversion is correlated to income
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE LIMITATIONS OF THE STUDY • Relatively small sample size • Income data • Educational homogeneity • Lack of other demographic variety e.g. age, marital status
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE AREAS FOR FURTHER RESEARCH • Explore ‘framing’ aspects between questions, especially amongst females. • Expand survey to compare different educational backgrounds • Survey that captures a representative sample of ages • Survey that captures a representative sample of income • Greater biographical data to explore links with other risk-taking behaviour
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE CONCLUSION A significant difference exists between the results of this study and the Hanna & Lindamood study and while the results of the subjective risk tolerance question are positively correlated with the SCF Investment question, the level of correlation is low. This indicates that more analysis would be required to substantiate (or otherwise) the usefulness of this approach for measuring individuals’ risk aversion in a South African context.
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE CONCLUSION Gender, race, religion and income were all seen to be significantly related to risk aversion. This suggests that there is value in tailoring specific investment options on the basis of these demographic variables. Further research is needed, however, to more fully understand the interaction between these variables.
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE CONCLUSION While the results are generally in line with what one would expect, a significant difference in risk aversion particularly among the female respondents, was observed between the different risk aversion questions. This raises serious questions about the impact of framing on the measure of risk aversion. The implication that females are more sensitive to framing than males is particularly intriguing and worthy of further research.
FINANCIAL RISK TOLERANCE: A SOUTH AFRICAN PERSPECTIVE THANK YOU