What Drives Individuals Superannuation Investment Choices? Preliminary Evidence on Return Chasing - PDF

Please download to get full document.

View again

of 25
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information Report



Views: 5 | Pages: 25

Extension: PDF | Download: 0

Related documents
What Drives Individuals Superannuation Investment Choices? Preliminary Evidence on Return Chasing Marilyn Clark-Murphy School of Accounting, Finance & Economics Edith Cowan University Joondalup WA 6027
What Drives Individuals Superannuation Investment Choices? Preliminary Evidence on Return Chasing Marilyn Clark-Murphy School of Accounting, Finance & Economics Edith Cowan University Joondalup WA 6027 Tel: Key contact person and conference presenter Paul Gerrans School of Accounting, Finance & Economics Edith Cowan University Joondalup WA 6027 Tel: Craig Speelman School of Psychology Edith Cowan University Joondalup WA 6027 Tel: Keywords: Behavioural finance, retirement savings, investment choice. The authors gratefully acknowledge the support of the GESB, UniSuper, STA and HESTA in the conduct of this research and the work of our research assistant Jacqui Whale. Abstract Australian superannuation funds have increased the investment choice available to their members. Fund members can typically choose from a range of ready-made options or select their own asset allocations. Evidence suggests that individuals may be unduly influenced by historical returns in making their investment choices. Such a bias may produce a sub-optimal investment over the longer-term. This paper investigates the investment choices of members of four major not-for-profit superannuation funds in Australia. We find significant evidence that choices are driven by the historical return performance of the investment alternatives available. 1.0 Introduction More than 89 percent of the assets of defined contribution superannuation funds are in funds which offer some level of investment choice to their members (APRA, 2005). This paper seeks to explore one possible factor in individuals superannuation investment decisions: historical returns on the investment alternatives offered by the fund. The dataset employed covers the investment choices of members from four large not-for-profit superannuation funds. The choice offered to members of superannuation funds can include a choice from a menu of ready-made investment options, typically with a default option and lower and higher risk options. Choice can also extend to a do-it-yourself option where a member constructs their own selection from a menu of asset classes, for example cash, fixed interest securities and equity. Commentators both in academia and in the superannuation industry have expressed concern about whether employees have an appropriate level of financial education to enable them to make the relatively complex choices that are being placed before them (Delpachitra & Beal, 2002; Dunstan, 1999; Marshall, 1999). If inappropriate choices are made then, as has been demonstrated by the experience in the UK (Album, 1998), people s ultimate retirement incomes can be significantly reduced. This is a problem not just for the individual but also for policymakers who are seeking to reduce the reliance on state provided age pensions. Given the trend towards increased investment and fund choice for members of Australian superannuation funds, and the importance of superannuation performance as a determinant of standard of living in retirement, it is opportune to investigate the drivers of members investment choice. The present study finds evidence of return chasing as one significant driver. 1 The rest of this paper is organised as follows: the next section outlines the literature related to evidence of return chasing in investment decisions. The third section details the literature relating to member choice in Australia and overseas. The fourth section presents details of the dataset used in the present study. Section five sets out a preliminary investigation into return chasing as a possible influence on investment choice and discusses the results obtained. The final section summarises the work and identifies areas of future research. 2.0 Return Chasing Future returns from the range of financial investment asset classes involve uncertainty and are impossible to predict with precision. When making judgements under uncertainty it appears individuals can be influenced by factors not necessarily relevant to the final outcome. For example Tversky and Kahneman (1974) describe evidence that people exhibit poor ability when estimating the likelihood of particular events. Instead of using relevant factors which may be too numerous or complex for them to process, individuals rely on heuristics, or rules of thumb that help to reduce the complexity of the task into simpler judgements. While many such heuristics serve us well, Tversky and Kahneman (1974) demonstrate situations where these heuristics are shown to be unreliable. The use of heuristics in financial decisions has been widely documented in the behavioural finance literature (see, for example, Shefrin, (2000)). Return chasing, i.e. allowing the investment decision to be unduly influenced by past performance, may be one such heuristic. Evidence of return chasing in retirement savings decisions has previously been identified by Cronqvist and Thaler (2004) who found that out of a large pool of funds available to Swedish fund members the one with the highest five-year trailing return attracted the largest market share. There are also signs of past performance as a factor in superannuation fund choice in 2 Australia; Clare (2006) finds that both those deciding to change from one fund to another and those deciding to remain with their existing fund identified past performance as a significant factor in their decision. Studies of mutual fund cash flows and investment have also found evidence of return chasing behaviour. In an Australian context Gerrans (2004) finds that the largest group in a sample of individual mutual fund investors rated historical performance highly as a factor influencing their decision to invest. In an earlier American study Capon, Fitzsimons and Prince (1996) also find that a fund s performance track record was the most important criteria for individual investors. Examining the cash flows of funds a range of studies have identified a positive link between past performance and cash inflows to the fund (see, for example Chevalier & Ellison (1997), Karceski (2002), Sapp (2004); Ippolito (1992); Sirri & Tufano, (1998)). The use of historical performance as a decision heuristic is not confined to individual (and arguably less expert) investors. Mei and Saunders (1997) find that past returns in the real estate market are a significant driver of property investment by American banks while Frazzini (2005) finds the stock choices of the mutual funds themselves exhibit return chasing behaviour. The use of return chasing as a heuristic to aid decision making when the investment area is unfamiliar is supported by studies into investment in overseas markets. Bohn and Tesar (1996) find that return chasing is a significant factor in U.S. equity investors decisions about investments in foreign markets. Choe, Kho and Stulz (2005) cite evidence of return chasing by foreign investors in the Korean stock market as a reason for their poor performance relative to local investors. 3 The present study seeks to add to this literature by exploring the investment decisions of individual members of four large not-for-profit Australian superannuation funds. There is an important difference between this population and the mutual fund investors considered by most previous studies. Australia s superannuation regime means that virtually all employees have mandated contributions made to a fund by their employers. Thus these investors are involuntary investors who may have no experience or interest in financial investment and yet are asked to make relatively complex investment decisions with significant implications for their income and lifestyle in retirement. By contrast those who invest in mutual funds have chosen to do so and are thus voluntary rather than involuntary investors. It may be reasonable to expect that involuntary, and possible inexperienced, investors are more likely to employ a return chasing heuristic. 3.0 Investment Choice The literature examining individual choice in a retirement savings framework is now large. A number of studies have examined the trend towards accumulation type DCF accounts, from the prescribed DBF accounts (Gustman and Steinmeier, 1992; Clare and Connor, 1999; Clark and Pitts, 1999; Dulebohn, Murray and Sun, 2000) Clark-Murphy and Gerrans (2001) and Gallery, Gallery et al. (2000) provide an analysis of the DBF/DCF choice in Australia. There is consistent Australian evidence that employees report feeling ill-informed and ill-equipped for the decisions presented to them relating to their superannuation decisions (Clare, 2002; Clark- Murphy & Gerrans, 2001; Plum Financial Services, 2001) though a recent survey suggests that this situation may be improving (Tuck, 2006). Several studies have suggest gender differences in risk aversion in general and in retirement investments in particular. The majority find women show greater risk aversion in the allocation 4 of funds to pension assets (Bernasek and Shwiff, 2001; Bajtelsmit, Bernasek and Jianakopolos, 1999; VanDerhei and Olsen, 2000,) and this is supported by Australian evidence (Gerrans and Clark-Murphy, 2004; Quinlivan, 1997). However Dwyer, Gilkeson and List (2002) find the level of risk aversion falls with increased financial education. Schubert, Brown, Gysler, & Brachinger (1999) find that women are not more risk-averse than men when financial decisions are put in context and more recently Brown, da Silva Rosa and McNaughton (2006) using an extensive Australian managed fund database suggest males are more risk averse. Agnew, Balduzzi and Sunden (2003), in a study of 7, (k) plans, find that men are more likely to make equity investments, that asset allocations tend to be extreme, with very high or very low allocations to equities, and very limited movement in allocations. Chernev (2004) discusses evidence of extremeness aversion in choice and a tendency to go for the compromise option. This suggestion may support the experience of most Australian superannuation funds that the majority of fund members remain in the default option. By contrast Benartzi and Thaler (2001) find that employees, whether male or female, are likely to adopt a naïve diversification strategy in employer-sponsored superannuation, dividing their funds equally between each of the investment strategies offered, although they also identify a tendency to choose a middle option (Benartzi & Thaler, 2002). Evidence of naïve diversification has also been found in Sweden (Hedesstrom, Svedsater, & Garling, 2004) in a study which supports extremeness aversion. Thus it has been suggested that plan design, the alternatives offered and the way funds can be divided, may all significantly influence the choices made (Chernev, 2004). However the relationship between naïve diversification and plan design has been questioned by Huberman & Jiang (2006). 5 Literature from other fields of consumption suggests that too many options may not facilitate good or satisfying choices and this is now being applied to superannuation decisions (Sethi- Iyengar, Huberman, & Jiang, 2004). In this context it has also been suggested (Papke, 2004) that the presence of investment choice increases the proportion of funds members hold in equity and the likelihood that members will make voluntary contributions. Taken as a whole the existing literature suggests that a wide range of factors may influence individuals investment decision making. Many of these factors are behavioural in nature and go beyond the inputs employed in modern portfolio theory as part of the rational decision making framework. 4.0 Overview of Funds and Data Four superannuation funds have allowed access to their membership data to enable examination of member investment strategy choices. The Health Employees Superannuation Trust Australia (HESTA), the Superannuation Trust of Australia (STA), the Government Employees Superannuation Board (GESB) and UniSuper have combined assets of $39 billion and 1.8 million members. HESTA, STA and UniSuper are industry funds whereas GESB is a publicsector fund. HESTA s 508,665 members 1 are predominantly from health and community services; it was one of the first industry funds to offer choice to its members in STA started as the fund for the manufacturing sector though it now has members in a variety of industries including automotive, entertainment and transport and total members of 515, Choice was introduced to STA members in July The majority of GESB members receive automatic membership into the West State Super Scheme when they join the Western Australia public sector. West State Super has 232,677 members and first offered choice in April HESTA STA UniSuper is the industry fund for employees in Australian universities. UniSuper was formed from the merger of the Superannuation Scheme for Australian Universities (SSAU) and the Tertiary Education Superannuation Scheme (TESS) in SSAU was established as a defined benefit fund (DBF) whereas TESS was formed as a defined contribution fund (DCF). The TESS fund is now known as the Award Plus Plan (APP). A majority of tertiary sector employees were members of both. SSAU members were given the opportunity in 1998 to move from the DBF to a DCF. This newly created DCF is now known as the Investment Choice Plan (ICP). Therefore members would have either one DCF account, a DBF and DCF account, or two DCF accounts. The analysis in this paper focuses on the members who opted into the DCF and the members who also had the TESS DCF account. UniSuper subsequently simplified this system though this is outside the period of data sampled. A full description of the history of SSAU and TESS is provided in the Appendix. UniSuper had 360,382 individual members with 447,946 accounts at June 2005, 174,204 members had one DCF account, 26,542 had two DCF accounts and 49,375 had a DCF and DBF account. UniSuper s assets totalled $15 billion at 30 June Overview of Investment Choice Available The level of investment choice varies between each of the four funds who have made their member database available. HESTA, STA, and GESB allow members the choice of a selection of readymade options, which have a specified investment strategy, or a do-it-yourself (DIY) option where members choose their own investment strategy. However, members in UniSuper 3 UniSuper 2005/6 annual report 7 can only choose from a selection of readymade options. 4 The three funds current offerings are summarised in Table 1. Each fund s options have evolved since first introduced. Insert Table 1 HESTA and STA members have the most extensive range of investment options which includes individual asset classes and the readymade options. Members are free to combine any mixture in a DIY option. GESB members can only construct a DIY investment strategy from a selection of asset classes. However GESB members can still spread their investment strategy across different readymade options. During the period under analysis, UniSuper members had the most restricted investment choice. Members could choose one investment option and not spread their investment allocation across the investment options Fund Investment Choice Data From the introduction of choice in July 1995 until December 2004, 44,393 HESTA members made 48,874 investment changes. This includes only changes applied to future contributions as members were able to choose a different strategy for their existing balance, the changes made to existing balances will be considered in future work. In the period between July 1997, when STA introduced choice, and December ,969 members made 27,488 changes. Between July 1997 and December 2002 these changes applied to a member s existing balance and future contributions. Since January 2002 the changes applied only to future contributions. A total of 17,609 GESB members made 19,688 changes between July 2001 and June These changes applied to both the existing balance and future contributions. Between July 1998 and June 2004, 22,170 UniSuper members who moved from the DCF to the DBF made 28,386 changes to their 4 UniSuper has since introduced a DIY choice 5 As discussed in the Appendix the ability to choose more than one investment option was introduced for members in July ICP. A total of 13,583 UniSuper members of the APP made 15,159 changes. The total of 43,545 changes applied to both the existing balance and future contributions. 5.0 Member Investment Choice and Historical Performance The historical performance of investment alternatives is a possible external driver of member choice. To investigate this, the historical performance of the investment options offered by each fund are calculated for trailing twelve month, six month and monthly periods using the funds published asset class monthly returns. In Table 2 the first three rows in each panel present historical performance comparisons for the old and new investment options using monthly credited returns for all decisions made after the first month of investment choice, six-monthly returns for all decisions made after the first six-months, and twelve-monthly returns after the first year. The fourth row of each panel uses published data from the fund s annual report. This was included principally to include those members making their first choice in the first year following the introduction of choice. Given the time period elapsed these changes would otherwise not be able to be analysed for a 12-month historical period. Insert Table 2 When the historical performance of old and new choices is compared, it suggests that once member investment choice is underway, the new choices selected are likely to have performed significantly better than the old option in the trailing month, six-month, and twelve-month period. Of the 17 data points 70% show a significant positive difference between trailing returns on the old and new options. These results establish that there is evidence of return chasing behaviour in overall choice patterns. It is appropriate to explore whether this behaviour is consistent across the first and subsequent choices made by members. Table 3 to Table 6 analyse, for each fund, the historical 9 returns of the investment option a member is moving from and to, according to the number of the choice the member is making. The first panel summarises results for members making their first investment choice, the second panel summarises results for members making their second investment choice, and so on, up to choice number four. (Choices beyond the fourth are not examined as the number of choices made by members is too small). Once again results indicate significant, although slightly less consistent, return chasing behaviour. Across the 64 data points 48 (75%) show a significant positive difference between trailing returns on the old and new options. Only 10 (15%) show a significant negative difference while a further 6 (10%) show no significant difference. Insert Table 3,4,5,6 The first choice made by members of HESTA provides a unique group of results with all observations showing a significant negative difference between trailing returns on the old and new options. This is reflected in the results reported in Table 2 where data for HESTA shows less positive differences than that for the other funds. Similar first choice results are not observed for STA and UniSuper although negative differences are present for the one month and annual return observations in the first choice of GESB members. It may be that material sent to members early in their membership of HESTA has some overriding influence on the nature of their first choice, after which they revert to return chasing behaviour. It is possible that certain events, such as the publication of the fund s annual report, may be a driver of members decisions to change their investment option. To investigate this investment changes are mapped across the year for each fund. The results appear in Figures 1 to 4. There do appear to be occasional spikes for some fun
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks