Nonlinear linkages between financial risk tolerance and demographic characteristics

Abstract

We explore the nonlinear linkage between financial risk tolerance and demographic characteristics. Our tests support the nonlinear role of age, income and number of dependents.

Acknowledgements

The authors would like to thank FinaMetrica Limited (formerly ProQuest Limited) for providing access to their database and Geoff Davey, CEO for information on the operation of their risk profiling system. The support provided by an ARC Linkage grant (LP0560992) is also gratefully acknowledged.

Notes

¹ Females typically show a lower preference for risk than males – for example, Lewellen et al. (1978); Bajtelsmit and Bernasek (1996); Powell and Ansic (1997); Grable (2000); Grable and Joo (2000); Halek and Eisenhaeur (2001).

²Roszkowski et al. (1993) suggest that single people have higher financial risk tolerance than married individuals because they have less responsibilities than married people, particularly with respect to dependents. However, a number of studies have failed to identify any significant relationship between marital status and financial risk tolerance (McInish, 1982; Masters, 1989; Haliassos and Bertaut, 1995).

³ Higher attained levels of education have been found to be positively related to higher financial risk tolerance – see, for example, Haliassis and Bertaut (1995); Sung and Hanna (1996).

⁴ The impact of income and wealth tend to support a positive relationship with levels of risk tolerance – see, for example, Friedman (1974); Cohn et al. (1975); Riley and Chow (1992); Schooley and Worden (1996); Shaw (1996); Grable and Lytton (1999).

⁵ The FinaMetrica Personal Financial Profiling system is a proprietary, computer-based risk tolerance measurement tool. It has been available commercially to the Australian financial planning industry since 1998 and was introduced in the US in 2002. See www.FinaMetrica.com.au for further information about the FinaMetrica system.

⁶ A value of 1 indicates the respondent did not complete high school; a value of 2 that they did complete high school; a value of 3 that they have a trade or diploma education; and a value of 4 that they have a university or higher qualification.

⁷ A value of 1 indicates an individual income under $30 000; a value of 2, an income between $30 000 and $50 000; a value of 3, an income between $50 000 and $100 000; a value of 4, an income between $100 000 and $200 000; and a value of 5, an income over $200 000.

⁸ A value of 1 indicates a combined income under $30 000; a value of 2, a combined income between $30 000 and $50 000; a value of 3, a combined income between $50 000 and $100 000; a value of 4, a combined income between $100 000 and $200 000; and a value of 5, a combined income over $200 000.

⁹ A value of 1 indicates net assets under $50 000; a value of 2, net assets between $50 000 and $150 000; a value of 3, net assets between $150 000 and $500 000; a value of 4, net assets between $500 000 and $1 000 000; and a value of 5, net assets over $1 000 000.

¹⁰ The variable CINC is interacted with DMARR, since it is only validly defined for ‘married’ respondents.

¹¹ The nonlinear effect in NASS is dropped due to insignificant results.