![Sample our Economics, Finance,Business & Industry journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5931/TOC-Subject-Sample-2021-Economics-Finance-Business-Industry.jpg"})
Abstract
We conduct a clinical study of the investment behavior of 115 subjects. Using Norman's Big 5, Preference for Innovation and Risk-Taking Propensity (from Jackson's Personality Inventory), and Bem's sex-role inventory, we confirm the argument presented in Durand, Newby, and Sanghani [2008] that personality is related to investment choices and outcomes. We extend Durand et al. by demonstrating that investors’ reliance on two heuristics used to model market movements—the availability heuristic and the disposition effect—are associated with their personality traits.
ACKNOWLEDGMENTS
We thank seminar participants at the Asian FMA (Singapore, July 2010), the Academy of Behavioral Finance and Economics (Chicago, September, 2010) and Monash University for their helpful comments. We are grateful to Kalok Chan for his constructive comments. We are also grateful for Stephen Sheely's assistance in developing the trading platform used for our experiment. Leila Peggs and Michelle Siekierka acknowledge the financial support of the University of Western Australia Business School Honours Research Scholarship.
Notes
1. Homo economicus is the economically rational, utility maximizing, representative agent whose motivations govern the behavior of the market.
2. The personality metrics are discussed in detail on pp. 194–195 of Durand et al. [Citation2008].
3. See Durand et al. [Citation2008], p. 196.
4. The assignment was called “Archie's Challenge.” It involved students benchmarking their performance against the All-Ordinaries Index (proxying for the Australian market portfolio) and a benchmark portfolio of 10 randomly chosen stocks (Archie's stocks). Archie is the name of the course mascot, a toy polar bear. Blindfolded students randomly chose Archie's stocks. The blindfolded students were also spun around so that the ensuing dizziness would facilitate the random selection of stocks. We note that the use of darts is a more traditional way of choosing a random portfolio but this is probably not a wise thing to do in a crowded lecture theater.
5. Auditing was conducted by tutors marking the assignments. The tutors did not know the identities of the students participating in the study.
6. Although the students were investing in the period which has become known as the “Global Financial Crisis”, the study period was one in which investors enjoyed respite from the gloom: the benchmark index for Australian stocks, the All-Ordinaries, rose 9.7% during the trading period.
7. Subjects would have been required to study and use the CAPM in at least one, if not two, courses before taking Investment Analysis.
8. See footnote 4 for a short description of how the benchmark portfolio was chosen.
9. The sample is positively skewed in this metric (as is the proportion of the portfolio traded) due to the influence of a few investors who traded frequently. These were not excluded because they may well have been a product of the investors’ particular personality traits and, therefore, are valid observations.
10. αp is the intercept in the standard CAPM equation (Rp-Rf)t = αp+βp(Rm-Rf)t +ϵ. Rp is the return of the portfolio being analysed, Rf is the risk-free interest rate, Rm is the return of the market, t is a time subscript, βp is an estimated coefficient and ϵ is an i.i.d. error term.
11. Multifactor models of Australian returns are discussed in Halliwell, Heaney and Sawicki [Citation1999], Faff [Citation2001], Gaunt [Citation2004], Durack, Durand and Maller [Citation2004], Durand, Limkriangkrai and Smith [2006a], and Kassimatis [Citation2008]. Durand, Juricev and Smith [Citation2007] discuss the prominent size effect in the Australian market. The role of liquidity in Australian stock returns is discussed in Chan and Faff [2003] and Limkriangkrai, Durand and Watson [2008]. Momentum in Australian stock returns is discussed in Demir, Muthuswamy and Walter [2004], Durand, Limkriangkrai and Smith [2006b], and Brailsford and O’Brien [2008].
12. That is, the period from February 14, 2008, to March 14, 2008.
13. The Australian and the Financial Review are leading national newspapers. The West Australian and the Sunday Times are the major regional newspapers. The Australian, Financial Review and West Australian are published daily from Monday to Saturday; the Sunday Times is only published on Sundays.
14. The data were collected using Factiva, an electronic database which houses newspapers and other print items.
15. This is the case for all of the analyses presented in this paper. A stepwise methodology involves deleting the least significant personality variable in a regression until a model containing only coefficients with p-values less than 0.10 is obtained. The resulting model is considered the “best” one. The stepwise method was used in Durand et al.; they considered that it was required because of the size of the sample (18 investors). However, the stepwise method was found to be necessary in this study as well. This may be due to the fact that though the personality metrics are orthogonal, certain personality “packages” are more prevalent – subjects are to some extent like one another and thus similar personalities are captured by the same traits, with most personality metrics correlated (significant at the 5% level). This contention is supported by the analysis of the correlation (both parametric and non-parametric) between the personality variables of our subjects reported in Appendix C.
16. We also analyzed the book-to-market ratios and liquidity of portfolios chosen by our subjects but no variables were statistically significant. Therefore, following our decision not to report insignificant results, we do not present these results.
17. An interesting area of future research might involve examining the role of negative emotion in influencing the response of investors to new information. In this way, standard event studies might be augmented by examining the psychological traits of the investors responding to the new information.
18. Size is measured by the natural log of the market capitalization of the stocks included in the portfolio that is initially established, weighted by the proportion in which they are held.
19. The Herfindahl Index can range from 0 to 10,000. Higher numbers mean a higher concentration of particular stocks (and, hence, lower diversification).
20. We have gone back to trading forms completed by subjects to examine whether this is the case. Although many individuals stated they were pursuing a strategy of overall diversification, they tended to choose stocks based on a specific piece of information or a recommendation which they believed was reliable. This meant they tended to choose a small number of stocks which they believed they knew well. Individuals choosing stocks they know well is documented in the literature (de Bondt [1998]). For example, subject number 12 states their strategy is “… a mixture of diversification yet still placing an emphasis on the natural resources industry.” In choosing BHP as a stock in their portfolio, they state their reasons as follows: “BHP is the world's largest diversified resources company so I feel the size of the company's operations will grant some stability in the shares [sic] price. The amount of diversification in its operations which include mineral exploration, production and processing as well as hydrocarbon exploration, production and refining will also provide stability. Buying shares in this company contributes to my trading strategy of buying into the natural resources industry.”
21. Short momentum is the change in the value of the stocks in the portfolio in the month before the portfolio is formed, weighted by the proportion in which they are held in the portfolio.
22. Long momentum is the change in the value of the stocks in the portfolio in the year before the portfolio is formed, weighted by the proportion in which they are held in the portfolio.
23. Although Tobit has a superficial resemblance to ordinary least squares regression, the maximum likelihood estimation procedure used for estimation means that the value of R2 is not valid.
24. Total return is measured as the subject's wealth at the end of the trading period in excess of the starting $50,000 and is expressed as a percentage.
25. The “days owned” variable was included to investigate whether there was a relationship between the time the share was held and the type of stocks sold; this follows Grinblatt and Keloharju [2001].
26. We also conducted a test of the disposition effect using a logit regression of an indicator variable signifying whether the individual realised a net gain or a net loss in their first trade. No independent variable was found to be statistically significant and, therefore, we do not report the results in this paper.
27. One subject was removed as an outlier realised around $23,000 in this second trade. This is very large given the starting value of the portfolio was $50,000.