FINANCIAL LITERACY AND PENSION INVESTMENT DECISIONS

Financial Accountability & Management, 27(3), August 2011, 0267-4424 FINANCIAL LITERACY AND PENSION INVESTMENT DECISIONS NATALIE GALLERY, GERRY GALLERY, KERRY BROWN, CRAIG FURNEAUX AND CHRISANN PALM∗ INTRODUCTION Following worldwide trends, there has been a significant shift from defined benefit plans (DBPs) to defined contribution (DC) superannuation funds in Australia.1,2 Most DBPs in both the public and private sectors are closed to new members (Australian Prudential Regulation Authority (APRA) 2007), with total assets held in DBPs comprising only about 20% of the total held in all superannuation funds with more than four members (APRA, 2009). Closure of DBPs to new members and the introduction of compulsory superannuation have resulted in rapid growth of assets and membership in DC funds. Given that most members in those funds have a choice of how their superannuation savings are invested, the adequacy of the retirement benefits that individuals will ultimately receive is, in part, contingent on the decisions undertaken by fund members throughout their working lives. Most working Australians are faced with the decision of choosing their superannuation fund, and once they have selected a fund, they need to decide among the various investment options offered by the fund to determine where their superannuation savings should be invested.3 These decisions are ongoing, requiring members to periodically monitor and evaluate the performance of their chosen fund and investment option, and decide whether to switch to another fund and/or investment option. To achieve optimal outcomes in this complex decision-making environment requires decision-makers to have adequate levels of financial knowledge and skills. As financial literacy is key to informed ∗ Natalie Gallery, Gerry Gallery and Chrisann Palm are in the School of Accountancy, and Craig Furneaux is in the School of Management at the Queensland University of Technology; Kerry Brown is in the School of Tourism and Hospitality Management at Southern Cross University. The authors gratefully acknowledge funding provided by QSuper and input to this research project by Nicole Peterman, QSuper Financial Literacy Manager. They also thank Cameron Newton in the QUT School of Accountancy for his assistance with the development of the survey questionnaire, Frederic Fery and Kathryn Heiser for their IT support in developing and managing the survey website, Jake Shorter and Sukie Sawang for their research assistance on this project, and two anonymous referees, Irvine Lapsley (editor), and Kym Irving for helpful comments and suggestions on earlier drafts of this paper. Address for correspondence: Natalie Gallery, School of Accountancy, Queensland University of Technology, GPO Box 2434, Brisbane, Queensland 4001, Australia. e-mail: [email protected]  C 2011 Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA, MA 02148, USA. 286 FINANCIAL LITERACY AND PENSION INVESTMENT DECISIONS 287 retirement saving decisions, better financial education is necessary if individuals are to achieve their retirement income objectives (Arnone, 2005). However, emerging evidence suggests financial illiteracy is widespread and the ability of individuals to plan adequately for their retirement is limited. In a review of financial literacy surveys in twelve member countries, the OECD (2005) noted that financial literacy levels among general populations are low; particularly for certain groups of individuals, such as those who are less-educated, have lower incomes, and belong to minority groups. In Australia, poor financial literacy, combined with the shifting of responsibility for choosing a fund and investments from employers and trustees to individual members, has created a high risk environment for those members (PJCCFS, 2007). Poor financial literacy has negative consequences for both individuals and society, in that bad decisions detrimentally affect individuals’ long-term financial well-being, which, in turn, could mean a greater burden on taxpayers to support those individuals in retirement (FSA, 2006). While recognising the general need for education programs to address poor levels of financial literacy, the OECD identifies that to properly design and appropriately target those education programs, more research is needed to assess the education needs of particular groups (OECD, 2005). Empirical evidence of financial literacy levels is limited with most studies conducted in the US and UK. In Australia, the research is confined to broad-based consumer surveys, such as the 2005 ANZ Survey of Adult Financial Literacy in Australia, that examines consumers’ knowledge and understanding of basic financial matters. No known prior research has examined financial literacy in the context of more complex superannuation or pension investment decision-making. Motivated by this absence of empirical research on the adequacy of financial literacy levels in the context of superannuation decision-making, and the significant adverse economic and social consequences of poor financial decisionmaking in superannuation matters, this study explores the financial literacy of members of a large Australian public sector-based superannuation fund. Associations between explanatory factors relating to demographics, wealth, benefit type and sources of information, and three dimensions of financial literacy – general financial matters, general investment matters, and specific investment matters – are examined. The next section provides an overview of the Australian superannuation system and how investment decision-making has shifted to superannuation fund members, followed by a review of prior research on financial literacy, leading to the research questions addressed in this study. The research design is then described, the research findings are presented, and conclusions are presented in the final section. OVERVIEW OF SUPERANNUATION IN AUSTRALIA Australia is one of only a few countries that has legislated mandatory superannuation as part of its retirement incomes policy. Since 1992, employers  C 2011 Blackwell Publishing Ltd 288 GALLERY, GALLERY, BROWN, FURNEAUX AND PALM have been required to pay superannuation contributions on behalf of employees earning $450 or more per month under the Superannuation Guarantee (SG) legislation; the contribution rate started at 4% and increased to 9% by 2002. With the introduction of the mandatory system the superannuation industry grew rapidly and assets held in superannuation funds now total more than $1 trillion (APRA, 2009). In addition to expanding coverage of superannuation to almost all Australian workers, mandatory superannuation also gave rise to an increase in the number of defined contribution (DC) funds, and a decline in defined benefit (DB) funds.4 Following the worldwide trend of employers closing their DB funds (Ashcroft, 2009), almost all Australian DB funds (including public sector funds) are closed to new members. With superannuation assets in DC funds rapidly growing after the introduction of the SG system, there was increased pressure to offer members investment choice. As at June 2007, 80% of funds with over $100 million in assets offer their members investment choice (APRA, 2007). Thus, the majority of fund members are now responsible for deciding how their superannuation savings are to be invested. As part of broader reforms to the regulation of financial services and products, the Financial Services Reform (FSR) Act 2001 sought to increase consumer protection through an enhanced disclosure regime to make the understanding of financial products easier for investors, and to promote consumer confidence. Additional disclosures are a key part of those reforms. Accordingly, the Corporations Act 2001 imposes a requirement on superannuation funds to provide a Product Disclosure Statement (PDS) when members first join the fund. The objective of a PDS is to help members ‘compare and make informed choices’ about superannuation products (ASIC, 2007, RG168.7). While those legislated disclosure requirements are primarily aimed at new members, the PDS also serves the purpose of providing all DC members with information about the fund’s investment options on an ongoing basis that enables them to periodically review the investment options and decide whether to switch their superannuation to an alternative investment strategy. Adequate disclosure is just one of the prerequisites for informed decisionmaking. Making informed investment choices also requires individuals to have adequate financial literacy in terms of knowledge and understanding of various investment products and associated risks. This study explores the extent to which members of one large superannuation fund possess such financial capabilities. FINANCIAL LITERACY AND PRIOR RESEARCH Although there is no universally agreed definition of financial literacy, a definition commonly used is: ‘the ability to make informed judgements and take effective decisions regarding the use and management of money’ (Schagen and Lines, 2006, p. ii). Improving financial literacy of individuals to facilitate informed financial decision-making has become a global issue over recent years.  C 2011 Blackwell Publishing Ltd FINANCIAL LITERACY AND PENSION INVESTMENT DECISIONS 289 The OECD guidelines for member countries on core rights and protection of members in occupation pension plans include the right for members to be provided with: sufficient opportunity to acquire the financial skills or education and other assistance that they need in order to make appropriate investment decisions in their pension plans (OECD, 2003, p. 13). Recently, the OECD (2008) also set forth recommendations on financial education of pension fund members. A range of government-initiated and other programs addressing financial literacy issues have emerged in various countries. For example, the National Strategy for Financial Capability developed by the Financial Services Authority in the UK, programs of the Financial Literacy Education Commission in the US, and the Understanding Money program developed by the Financial Literacy Foundation in Australia. These programs are generally very broad in that they are targeted at improving basic financial knowledge and skills necessary for day-to-day financial decision-making such as managing household budgets and savings, and finances related to personal loans and home mortgages. Empirical research on financial literacy is largely confined to broad population surveys which assess individuals’ attitudes and behaviours in relation to general financial matters,5 and are generally based on the self-assessed responses of participants to questions aimed at measuring very basic financial literacy. Two recent studies in the Netherlands and the US test more advanced financial knowledge and skills. van Rooij et al. (2007) designed two modules included in the 2005 and 2006 DNB Household Surveys in the Netherlands to measure basic financial literacy, such as the effect of inflation and compounding interest, and more advanced financial knowledge, including understanding differences between stocks and bonds, how risk diversification works, and the relationship between bond prices and interest rates. Lusardi and Mitchell (2007) draw on the van Rooij et al. (2007) model to test basic financial literacy and what they term as ‘sophisticated financial literacy’. In both the van Rooij et al. (2007) and Lusardi and Mitchell (2007) studies, the ‘advanced’ and ‘sophisticated’ measures of financial literacy focus exclusively on knowledge and understanding of investment products and markets. As neither of those studies focuses on decision making in a pension context, it is difficult to generalise their findings to the setting of our study. In focusing on financial literacy relevant to investment decision-making in the context of superannuation funds, this study addresses the following two research questions: 1. How does the level of financial literacy of superannuation fund members vary across general financial and specific investment matters? 2. What demographic and other factors are associated with different financial literacy levels among superannuation fund members?  C 2011 Blackwell Publishing Ltd 290 GALLERY, GALLERY, BROWN, FURNEAUX AND PALM The identification of key aspects of financial literacy relevant to the superannuation context is an important step towards understanding context-specific components of financial literacy. In this regard, the findings of the study extend prior research beyond broad financial literacy relating to day-to-day household financial matters. The findings also have the potential to inform the development of financial literacy programs that are more specifically tailored to the superannuation domain. SAMPLE AND DATA Data for this study are drawn from a broader financial literacy survey of the members of QSuper (a large public sector-based superannuation fund) conducted by the researchers in late 2007. QSuper is the primary superannuation fund of individuals employed by public sector departments and organisations in the state of Queensland. At the time of the study, QSuper had over 490,000 members and around $25 billion in assets under management. Fund membership comprises a broad range of government employees including police, teachers, and health workers in occupations that extend from relatively low-skilled (e.g., gardeners, cleaners) to professionals and executives (e.g., doctors, department heads). While the fund has a defined benefit section, it is now closed to new members and all new members join the defined contribution (DC) section of the fund. At the time of the study, QSuper offered DC members a choice of eight investment options when joining the fund, and existing members could switch their investment option up to four times per year. The default option for members who do not make an investment choice is the ‘Balanced’ option. The online survey questionnaire captured information about respondents’ understanding of general and specific financial matters, investments and risk, sources of financial information and advice, retirement readiness, and superannuation knowledge. Additionally, demographic information relating to individuals’ education, work status, current dwelling, household situation and household income was collected. The survey was conducted via a dedicated website linked to a secure server at the researchers’ institution where all responses were recorded. An invitation to all members of the superannuation fund to participate in the survey was publicised on the fund’s homepage, with a link to the survey website. Emails were also sent by the fund administrator to the Human Resources sections of various departments requesting them to email their employees with an invitation to participate in the survey. Respondents were required to enter their superannuation fund client number to enable matching with other data on the fund’s database. However, all responses to the survey were anonymous as there was no access by researchers to data identifying respondents by name. A total of 2,370 responses were received of which 2,046 were matched to valid client numbers. Fourteen responses were eliminated due to missing data, leaving a final sample of 2,032 used in the analysis.  C 2011 Blackwell Publishing Ltd FINANCIAL LITERACY AND PENSION INVESTMENT DECISIONS 291 The demographic statistics show that just over half of the respondents are male, the average age is 46, over half are tertiary educated, their average household income is between A$80,000 and $90,000, and the average superannuation account balance is A$207,000. About 80% of respondents own their home (either mortgaged or mortgage-free), one-third live as a couple with no children, 41% are in a couple arrangement with children. A higher proportion of the respondents live in the State capital Brisbane (58%) than in regional areas of Queensland, and the vast majority work full-time (87%). By their nature, opportunity samples introduce bias. In comparison with the QSuper population, the sample comprises a higher proportion of members who are males, older, on higher salaries, and have higher superannuation account balances. For example, the QSuper population comprises 33% of members who are in the ‘Baby Boomers’ age category, compared with 61% of our sample in that age category. It is expected that these differences bias the sample conservatively toward members who are more financially literate and/or more interested in financial matters (for example, because they are closer to retirement). MEASURING FINANCIAL LITERACY Following prior research (van Rooij et al., 2007; and Lusardi and Mitchell, 2007), financial literacy was measured through objective tests of both basic and advanced financial knowledge and understanding. The nine questions on the survey instrument relating to general financial matters, such as understanding compounding interest, and general investment matters, such as understanding the importance of diversification, were based on questions from the ANZ Survey of Adult Financial Literacy in Australia (ANZ, 2005). A further set of 16 questions relating to the specific investment options offered by the participants’ superannuation fund were developed by the researchers.6 For each question a dummy variable was constructed, coded one (1) if the answer to the question was correct and zero (0) otherwise. In relation to the questions testing understanding of the relative risk associated with each of the fund’s eight investment options (Questions 10 to 17), respondents’ ratings were compared with the risk indicators presented in the fund’s Product Disclosure Statement (PDS). Those who rated the risk of the investment options consistently with the indication given in the PDS were scored correct (1) and zero (0) otherwise. This scoring system is presented in Table 1. The same scoring method applied to the questions on the level of expected long-term returns for each investment option. In addition, to assess the extent to which respondents understand the relationship between risk and return, a score was assigned to reflect the level of congruence between the responses to the risk question for each investment option and the responses to the corresponding question on returns for that product.  C 2011 Blackwell Publishing Ltd 292 GALLERY, GALLERY, BROWN, FURNEAUX AND PALM Table 1 Scoring System for Investment Product Risk Rating Investment Option a. Balanced b. Cash Plus c. Socially Responsible d. High Growth e. Cash f. Fixed Interest g. Australian Shares h. International Shares Fund’s PDS Risk Rating On the border between medium and high Medium, tending to low High, tending to medium High (mid) Low (mid) Medium (mid) High (mid) High (mid) Responses Scored as ‘Correct’ High risk; medium risk Medium risk; low risk High risk; medium risk High risk; very high risk Low risk; very low risk Medium risk; low risk High risk; very high risk High risk; very high risk RESULTS To avoid biases that could arise from simply summing the scores for survey question responses and to discern the underlying structure of the survey instrument, factor analysis was conducted. Initial principal component analysis of all 25 questions and the eight additional risk-return variables identified three interpretable factors: general financial matters literacy, general investment literacy and specific investment literacy. The resulting structure reflects the underlying levels of specificity and complexity inherent in the items.7 Analysis of Responses to Objective Questions Table 2 , Panel A shows that for the four questions (Q1-Q4) testing knowledge and understanding of general financial matters, the percentages of correct responses range from 51.4% to 91.4%. Panel B of Table 2 presents the cumulative numbers of correct responses for the four questions, showing that about threequarters of respondents answered either three or all four questions correctly (mean score = 3.04). While respondents generally have a good understanding of how interest rates operate and compounding effects of interest, only about half correctly identified that the best measure of their superannuation fund’s performance is returns net of all fees. About one-third indicated investment returns as the best indicator of fund performance (not shown in the table). The superannuation fund’s relatively low fees may explain this result; that is, members know the fund charges low fees and therefore consider only investment returns are relevant in evaluating fund performance. Alternatively, as members do not have the option of transferring their superannuation account to another fund while they are employed by the government, they may feel ‘locked in’ and therefore see no point in taking the fund fees into consideration when assessing fund performance. Notwithstanding the reason for the high proportion  C 2011 Blackwell Publishing Ltd 293 FINANCIAL LITERACY AND PENSION INVESTMENT DECISIONS Table 2 Frequencies of Responses for Basic General Financial Literacy Questions (N = 2,032) Panel A: Percentage of Correct, Incorrect, Don’t Know Responses Correct Incorrect Q1: Differences in interest rates Q2: Differences in interest payments and repayment frequency Q3: Understanding compound interest Q4: Understanding effect of fees on returns Panel B: Number of Correct Responses None 1 N Correct 30 1.5% 116 5.7% Don’t Know 88.5% 91.4% 4.2% 4.6% 7.3% 4.0% 72.6% 51.4% 22.5% 36.9% 4.9% 11.7% 2 3 All 4 Mean 349 17.2% 786 38.7% 751 37.0% 3.04 of respondents either answering the question incorrectly or indicating they do not know the answer, it is evident that a large number of respondents may not understand the effect of fees on overall fund performance. Table 3 presents the frequencies of responses to the five questions (Q5-Q9) on general investment matters with correct responses ranging from 74.5% to 93.5%. About 80% of respondents answered four or all five questions correctly (mean score = 4.16), indicating a relatively high level of financial literacy on general investment matters among this sample of superannuation fund members. The areas of weakness are in relation to basic understanding of the risk-return relationship of investments over the long-term and the importance of diversification. The next set of questions (Q10-Q25) tests more advanced levels of financial literacy in relation to the understanding of the relative risks and returns of the eight investment options available to members of the fund (shown in Figures 1 and 2). These are previously unexamined aspects of pension fund financial literacy. Respondents were asked to rate the relative risk of each option on the five-point scale ranging from very low risk to very high risk, and then to indicate the expected long-term returns for each option, ranging from very low to very high returns. The percentages of responses are graphically presented in Figures 1 and 2. It is interesting to note that in contrast to responses to the other two sets of more basic questions, ‘don’t know’ responses are over 12% for all 16 items, and over 20% ‘don’t know’ responses for several items. These differences suggest respondents are more uncertain in their knowledge of more advanced investment concepts, which is also reflected in the lower percentages of correct responses.  C 2011 Blackwell Publishing Ltd 294 GALLERY, GALLERY, BROWN, FURNEAUX AND PALM Table 3 Frequencies of Responses for General Investment Literacy Questions (N = 2,032) Panel A: Percentage of Correct, Incorrect, Don’t Know Responses Correct Incorrect Q5: Level of risk for investment with high return Q6: Effect of market fluctuations on investment value in long term Q7: Effect of market fluctuations on investment value in short term Q8: Importance of investment diversification Q9: Evaluating investment advertised as high return and no risk Panel B: Number of Correct Responses None 1 2 N Correct 27 1.3% 55 2.7% 91 4.5% Don’t Know 93.5% 2.4% 4.1% 74.5% 6.3% 19.3% 92.9% 1.3% 5.8% 76.0% 15.8% 8.2% 79.0% 15.5% 5.5% 3 4 All 5 Mean 242 11.9% 598 29.4% 1019 50.1% 4.16 The eight investment options are presented in order of risk in Figures 1 and 2, ranging from lower risk (Cash) to higher risk (High Growth).8 While larger proportions of respondents correctly rated the low risk option (Cash) and two of the high risk options (International Shares and High Growth), only half or fewer correctly rated the options between those two extremes. The Balanced option is the fund’s default investment option in which the majority of fund members’ assets are invested, yet only 44% correctly indicated the level of risk associated with that option. Surprisingly, of the 43% who provided an incorrect response for the ‘Balanced’ option, more than three-quarters rated it as low risk. Also surprising are the high proportions of either incorrect or ‘don’t know’ responses in relation to the predominantly cash-based products. For example, the 41% who provided an incorrect response for ‘Fixed Interest’ comprised 40% who indicated that the option is very low risk, whereas it is rated as medium risk in the fund’s Product Disclosure Statement (PDS). The ‘Cash’ option is rated low risk in the PDS but 21% of respondents rated it as medium, high, or very high risk. This fundamental misunderstanding of the nature of risk associated with various investment options within the superannuation fund, and specifically in relation to the ‘Balanced’ option in which the majority of members’ accounts are invested, gives rise to questions about whether large numbers of members have the requisite knowledge and skills to undertake even the most basic assessment of the relative risks of the investment options on offer. Figure 2 shows that respondents were generally more accurate in indicating the level of expected returns for each investment option, and in contrast to  C 2011 Blackwell Publishing Ltd FINANCIAL LITERACY AND PENSION INVESTMENT DECISIONS 295 Figure 1 Percentages of Correct, Incorrect and Don’t Know Responses on Rating the Risks Associated with Each of the Eight Investment Options Figure 2 Percentages of Correct, Incorrect and Don’t Know Responses on Rating the Expected Level of Long-Term Returns on the Eight Investment Options the results for rating the risk of the ‘Balanced’ option, just over three-quarters correctly rated the returns. The differences in accuracy between risk and returns of each investment option highlights that most respondents do not understand the basic risk-return relationship of investments. The modulus of differences shown in Figure 3 further highlights the disparity between the risk and return  C 2011 Blackwell Publishing Ltd 296 GALLERY, GALLERY, BROWN, FURNEAUX AND PALM Figure 3 Percentages of Responses Rating the Risk and Returns the Same for Each of the Eight Investment Options ratings with half or fewer respondents giving the same ratings for each product (e.g., if the risk was rated high, the expected returns for that product were also rated high). It could be argued that if risk is rated consistently higher than returns, then this indicates a conservative bias which is less alarming than if expected returns are rated higher than risk. However, the reverse is the case here. The analysis of the differences between the risk and return ratings (not reported in tables) indicates that for five of the options (Balanced, Cash Plus, Cash, Fixed Interest, and Australian Shares) the vast majority of differences are biased toward returns being rated higher than risk. In the case of the ‘Balanced’ option, 41% of respondents rated the returns higher than the risk, further highlighting inadequacies in the basic understating of the risk-return profile of the option in which most of the DC members’ assets are invested.9 Financial Literary Indices Factor scores for the three measures of financial literacy were obtained using the Anderson-Rubin method (Tabachnick and Fiddell, 2007). As indicated previously, simply summing the scores for responses to each question is likely to bias the results, given the overlap of concepts across questions. The three indices derived from the factor analysis are also indicators of the level of sophistication in financial literacy with the first two indices (general financial matters literacy and general investment literacy) regarded as more basic knowledge and skills,  C 2011 Blackwell Publishing Ltd FINANCIAL LITERACY AND PENSION INVESTMENT DECISIONS 297 and specific investment literacy representing higher levels of knowledge and understanding of financial concepts. Self-Assessed Financial Knowledge Respondents were asked the question ‘How do you rate your knowledge of financial matters?’ on a five-point scale ranging from very poor (1) to very good (5), and these self-ratings are compared with the scores derived from the three measures of financial literacy from the objective tests. On average, the respondents accurately self-assessed their general financial and general investment literacy. However, this distinction is less clear in relation to specific investment literacy. Segmenting respondents into quintiles based on their specific investment literacy scores, Figure 4 shows the distributions (in percentage terms) of respondents’ self-ratings across the quintiles. Just over half of those who self-rated their financial knowledge as poor or very poor also have specific investment literacy scores in the lowest two quintiles (very low and low). However, only 41% of those with good or very good self-ratings have literacy scores in the highest two quintiles (high and very high), with the other 59% falling into the medium or lower literacy score quintiles, indicating their objectively measured financial literacy is in fact moderate or poor. This Figure 4 Comparison of Respondents’ Self-Rated Financial Knowledge and Financial Literacy Scores on Specific Investment Matters  C 2011 Blackwell Publishing Ltd 298 GALLERY, GALLERY, BROWN, FURNEAUX AND PALM result may indicate over-confidence on more advanced financial matters, such as when evaluating different investment products.10 For example, when choosing an investment option or evaluating their choice, DC members who perceive they have higher levels of literacy but may not have the necessary understanding of the relative risks/returns of each investment option to make an informed judgement. Accordingly, a mismatch between self-assessed financial knowledge and actual understanding of more advanced investment matters potentially leads to overconfidence in investment decision-making that could result in undesirable long-term financial outcomes. Demographic and Contextual Factors Associated with Financial Literacy An individual’s financial literacy is influenced by a range of factors and therefore the context of the individual’s circumstances needs to be considered when evaluating financial literacy (Kempson et al., 2005). Contextual factors include the individual’s financial circumstances, life stage, information environment and information usage, and whether the individual consults intermediaries to assist with financial decision-making (Kempson et al., 2005). Demographic and contextual factors that have been found to be associated with financial literacy in prior research (see Agnew and Szykman, 2005; Worthington, 2006; van Rooij et al., 2007; and Lusardi and Mitchell, 2007) and included in our analsyis are: age (AGE), gender (GEND), education (EDUC), home ownership (DWELL), household situation, in terms of whether the person is living as a couple without children (HHNC) or with children (HHC), whether the individual lives in the city or a regional area (REG), and the individual’s working status (WORK). Individuals’ financial wealth and investment characteristics are captured by household income (HHINC), their superannuation account balance (SUPPACC), whether their main account is a defined contribution benefit (DCBEN), whether they chose the investment option (CHOICE), and whether they hold other investments in the form of cash products (OIcash), property (OIprop) or shares (OIshar) separate from their superannuation fund. A final set of variables capturing the sources of information and advice used by individuals in their financial decision-making are: whether the superannuation fund’s Product Disclosure Statements (PDS) is used, usage of other information available from the superannuation fund (INFOsf ), financial information available from other sources (INFOoth), whether the individual consults experts, in the form of accountant or financial planner (ADVex), or family, colleagues or friends (ADVnex) to assist with financial decision-making. Associations between these explanatory variables and the three objective measures of financial literacy (General Financial, General Investment and Specific Investment), and the subjective self-rated measure of financial knowledge, are tested using a multiple regression model, where it is assumed that the variables are a linear additive function with an error term.  C 2011 Blackwell Publishing Ltd FINANCIAL LITERACY AND PENSION INVESTMENT DECISIONS 299 Table 4 Descriptive Statistics (N = 2,032) AGE (years) EDUC (Below Year 12 = 1; Year 12 = 2; TAFE/Trade Cert = 3; Degree/Diploma = 4; Postgraduate = 5 HHINC ($000) SUPACC ($000) INFOsf (Number of fund information sources used) INFOoth (Number of other information sources used) Mean Median Std. Dev. Min. 45.7 3.5 48.0 4.0 10.8 1.3 17.0 1.0 73.0 5.0 80–90 207.1 4.2 80–90 133.7 4.0 214.4 1.7 <20 <0.1 0 >150 2,750.6 8.0 1.8 2.0 Coded GEND (Male = 1; Female = 0) DWELL (Own home (mortgaged or mortgage-free) = 1) HHNC (Couple with no children at home = 1) HHC (Couple with children at home = 1 REG (State capital city = 1 WORK (Working full-time = 1) DCBEN (Member’s main account is defined contribution benefit = 1 CHOICE (Member chose investment option for DC benefit account = 1) OIcash (Hold Investments in term deposit or cash management account = 1) OIprop (Hold Investments in property (owned/mortgaged) = 1) OIshar (Hold Investments in Australian and/or international shares = 1) PDS (Used fund’s Product Disclosure Statement before choosing investment option(s) = 1) ADVex (Consulted accountant or financial planner = 1) ADVnex (Consulted family members, colleagues or friends = 1) 1.3 0 Max. 4.0 1 0 54.5% 81.7% 45.5% 18.3% 33.6% 40.6% 57.8% 86.5% 45.1% 66.4% 59.4% 42.2% 13.5% 54.9% 45.9% 54.1% 42.3% 57.7% 31.0% 69.0% 45.8% 54.2% 25.7% 74.3% 72.6% 27.4% 75.8% 24.2% Table 4 presents descriptive statistics for the independent variables entering the regression model. Just under half (45%) of the respondents have DC as the main type of superannuation benefit, with defined benefit being the main type of benefit for the remainder. Another 15% of respondents whose main benefit is defined benefit also have a DC account;11 thus 60% in total have some or all superannuation benefits in a DC account (not shown in tables). Forty-six percent  C 2011 Blackwell Publishing Ltd 300 GALLERY, GALLERY, BROWN, FURNEAUX AND PALM of respondents (or three-quarters of those with a DC account) stated that they chose the investment option for their DC benefit, indicating a high level of engagement with superannuation among this sample of members. In terms of other types of investments held by respondents, 42% have term deposits or cash management accounts, 31% have direct property investments, and 46% invest in Australian or international shares. With an overall 60% of respondents holding direct investments in property and/or shares suggests respondents tend to be active investors. With respect to information sources, while 75% of members whose main account is a DC benefit stated that they chose their investment options, only about a quarter indicated they used the fund’s Product Disclosure Statements prior to selecting those options.12 In relation to other sources of information used to assist with financial decision-making relating to superannuation, on average, respondents accessed four or more of the resources provided by QSuper, such as newsletters, periodic fund reports, seminars and the telephone hotline, and two or more sources from outside the superannuation fund, such as financial newspapers, magazines and other paper and internet-based publications. This relatively extensive usage indicates the respondents are diligent in keeping themselves informed on financial matters relevant to their superannuation. In addition, 73% consult experts (accountant and/or financial planner) to assist with their financial decision-making. However, 76% also consult non-experts (family members, colleagues and/or friends), bringing into question the quality of that advice. Table 5 presents the results of the regression analysis for the three financial literacy indices and for the self-rated financial knowledge variable, with the independent variables grouped into the four categories of demographic factors, wealth factors, superannuation benefit type and sources of information/advice. Demographic factors: The demographic factors age (AGE), gender (GEND) and education (EDUC) coefficients are positive and significantly associated with all three measures of financial literacy, indicating older members, males, and those with higher levels of education are the most financially literate. In relation to the self-rated measure of financial knowledge (last column of Table 5), only GEND is significant, indicating that although better-educated and older members do not perceive their knowledge of financial matters any differently from younger, less well-educated members, this is not reflected in their actual financial literacy, based on the objective measures. A possible implication is that younger and less well-educated members may be more over-confident of their financial capabilities relative to older and better-educated members. The fact that these cohorts of members feel more confident than their actual financial knowledge suggests a need for a targeted financial education programs to address this knowledge deficiency. Wealth factors: Results for wealth factors show that higher levels of financial literacy are associated with wealthier households, as indicated by the positive  C 2011 Blackwell Publishing Ltd FINANCIAL LITERACY AND PENSION INVESTMENT DECISIONS 301 Table 5 Regression Results (N = 2,032) Independent Variables (1) General Financial Literacy Coeff. (2) General Investment Literacy t-value ∗∗ −7.798 Demographic Factors AGE 0.007 GEND 0.190 EDUC 0.062 HHNC −0.024 HHC −0.004 REG −0.035 WORK −0.015 Wealth Factors DWELL HHINC SUPACC OIcash OIprop OIshar Coeff. (3) Specific Investment Literacy t-value −1.996 ∗∗ −8.565 2.673∗∗ 4.075∗∗ 3.801∗∗ −0.376 −0.058 −0.816 −0.227 0.010 0.202 0.088 −0.005 0.047 0.018 −0.037 0.288 0.022 0.049 −0.007 0.063 0.052 4.710∗∗ 2.951∗∗ 2.492∗ −0.155 1.365 1.150 Benefit Type/Choice DCBEN 0.045 CHOICE 0.210 Coeff. (4) Self-rated Financial Knowledge t-value Coeff. t-value −0.915 ∗∗ −3.758 1.801 9.087∗∗ 4.172∗∗ 4.439∗∗ 5.517∗∗ −0.081 0.783 0.430 −0.586 0.007 0.186 0.058 −0.054 0.091 0.042 0.040 2.792∗∗ 3.906∗∗ 3.463∗∗ −0.851 1.434 0.960 0.603 0.000 0.134 −0.005 0.079 −0.046 0.058 −0.118 0.091 3.459∗∗ −0.345 1.517 −0.894 1.625 −2.197∗ 0.101 0.025 0.021 0.041 −0.038 0.139 1.686∧ 3.438∗∗ 1.073 0.982 −0.833 3.123∗∗ 0.123 0.016 −0.056 0.054 0.027 0.130 1.974∗ 2.097∗ −2.763∗∗ 1.220 0.577 2.793∗∗ 0.062 0.024 0.016 0.063 0.081 0.226 1.223 3.797∗∗ 0.990 1.755∗ 2.098∗ 5.958∗∗ 0.861 3.713∗∗ 0.056 0.275 1.080 4.964∗∗ 0.004 0.170 0.070 2.944∗∗ 0.054 0.285 1.238 6.049∗∗ Information/Advice Sources PDS 0.007 INFOsf −0.022 INFOoth 0.083 ADVex 0.058 ADVnex 0.030 0.109 −1.523 4.367∗∗ 1.152 0.610 −0.064 0.033 0.061 0.177 −0.002 −1.067 2.354∗ 3.274∗∗ 3.592∗∗ −0.045 −0.132 0.039 0.057 0.223 0.062 −2.101∗ 2.709∗∗ 2.948∗∗ 4.311∗∗ 1.209 0.086 0.054 0.196 0.106 −0.070 1.685∧ 4.583∗∗ 12.377∗∗ 2.534∗ −1.69∧ Adj R2 F-stat 0.136 17.020∗∗ (Constant) −1.857 0.179 23.160∗∗ 0.104 12.800∗∗ 0.321 48.900∗∗ Note: ∧ , ∗ , ∗ ∗ significant at the 0.1, 0.05, and 0.01 levels (two-tailed test). associations with owning a home (DWELL), and higher income households (HHINC)13 for general financial literacy. These wealth factors and the additional indicator of investing in shares (OIshar) are also positively associated with the two investment literacy measures.14 It is however, surprising that the amount of a member’s superannuation account balance (SUPACC), as another indicator of wealth, is positively associated with financial literacy only in relation to general financial matters. For the two investment literacy measures, no association is found between account balance and general investment literacy, and the association with specific investment literacy is negative. These inconsistent findings suggest that the size of the superannuation account balance is not a reliable indicator of members’ financial literacy. Moreover, the negative coefficient for superannuation account balance (SUPACC) implies  C 2011 Blackwell Publishing Ltd 302 GALLERY, GALLERY, BROWN, FURNEAUX AND PALM that members with more wealth at risk have a poorer understanding of more sophisticated investment matters. In contrast to the mixed findings for the objective financial literacy measures, the significant coefficients on the wealth indicators – household income, superannuation account balance and other investments (OIcash, OIprop, OIshar) – are consistently positive for self-rated financial literacy. While those who hold other investments in the form of cash (OIcash) or property (OIprop) self-assess their financial capabilities higher than others who do not hold such investments, these differences are not evident for any of the objective measures of financial literacy. This disparity between those members’ selfassessed and actual financial capabilities suggests they may be over-confident in their investment-related decision-making and may, as a consequence, make sub-optimal investment decisions. Benefit type/investment choice: As discussed earlier, the defined benefit section of QSuper is closed to new members and therefore all new employees join the DC plan. At the time of this study, the main account (highest value) for 45% of the members is a DC account and an additional 15% of members have a DC account in addition to their main defined benefit account. Given that those members whose main account is a DC plan bear the investment risk associated with their superannuation, they could be expected to have greater financial literacy relative to members whose main account is defined benefit. Surprisingly, this is not evident in the results; DCBEN is not significant for any of the literacy indices. However, the significant positive coefficient for CHOICE indicates that DC plan members actively choosing investment options are better equipped to make informed choices. Sources of information/advice: The results for the final set of variables in Table 5 reveal divergent findings about members’ sources of information and advice. The fund’s Product Disclosure Statement (PDS) is a key document providing up-to-date information about each of the fund’s investment options, including details of asset allocations, risks and expected returns. The univariate analysis highlighted that of the members who actively chose their investment option, only a little more than half used the PDS prior to making that choice. The significantly negative coefficient for PDS indicates that those with lower levels of specific investment literacy are likely to use the PDS when exercising investment choice. While it is encouraging that those members recognise the need to refer to such information, their lower levels of understanding of the relative risk and returns associated with the fund’s investment options brings into question whether those members are making informed choices. Also, the tendency of members with higher levels of specific investment literacy not to use the PDS suggests those members may be overlooking important information when making their investment choices. On the other hand, the relatively low level of PDS usage by members generally, and by members with higher specific investment literacy, may simply reflect the criticism that PDS documents are  C 2011 Blackwell Publishing Ltd FINANCIAL LITERACY AND PENSION INVESTMENT DECISIONS 303 generally too long, complex and difficult for the average superannuation fund member to understand (see note 12). While policy-makers are aiming to make PDSs more readable, whether the content is understood (in whatever form) is dependent on users’ having a prerequisite understanding of fundamental investment concepts. In contrast to the finding for PDSs, members who use more of the information sources offered by the superannuation fund (INFOsf ) have higher levels of investment literacy, and those who use more external information sources (INFOoth) have higher scores for all measures of financial literacy. Members who use more of both the internal superannuation fund and external information sources also perceive they possess higher levels of financial capability. Also, those with higher levels of financial literacy on both general and specific investment matters are more, rather than less likely to consult experts, such as accountants or financial planners (ADVex). In comparison, those who consult non-experts (ADVnex), such as friends, colleagues and family members, do not have significantly different levels of financial literacy from other members, but they do appear to rate themselves lower in financial knowledge relative to other members. Analysis of explanatory factors: Further analysis shows that across the four regression models, the four categories of predictor variables have differing explanatory power. Based on semi-partial correlation analysis (not shown in tables), for the General Financial Literacy regression (1), about one third of the explanatory power (adjusted R2 = 13.6%) is attributable to unique variability of the significant independent variables (with the remainder attributable to shared variability), with demographic and wealth factors the dominant sources, each contributing about 12% unique variability. However, in the General Investment Literacy model (2) (adjusted R2 = 17.9%), only the demographic factors are dominant, contributing 15% unique variability compared to around 6% for the other three groupings. For Specific Investment Literacy (3) (adjusted R2 = 10.4%), the demographic factors and information/advice sources dominate, contributing 15% and 17% unique variability, respectively, with wealth factors contributing 10%. In contrast to results for the three objective measures of financial literacy, information/advice sources overwhelmingly dominates in the Self-rated Financial Knowledge regression (4) (adjusted R2 = 32.1%), contributing 19% unique variability, with relatively small amounts added by the other factors (demographic = 2%; wealth = 6%; investing choice = 4%). Thus the semi-partial correlation analysis highlights that variations in respondents’ perceptions of their financial capabilities are predominantly explained by the information sources they use, whereas variations in their actual financial literacy is predominantly explained by other factors. This finding suggests that for some individuals, using more information sources may result in false confidence in their financial decision-making abilities.  C 2011 Blackwell Publishing Ltd 304 GALLERY, GALLERY, BROWN, FURNEAUX AND PALM CONCLUSION As in many other countries, retirement savings in Australian superannuation funds have shifted from being predominantly defined benefit to defined contribution (DC) benefits, with almost all Australian defined benefit schemes now closed to new members. The decline in defined benefits and the introduction of mandatory superannuation in the early 1990s led to massive growth in the number of members in DC funds and the assets held in those funds. With the growth in coverage and value of superannuation, government policies and superannuation fund initiatives have increasingly encouraged members to take a more active role in the management of their retirement savings. The expansion of investment choices within superannuation funds and the expectation that members choose how their superannuation assets are invested raises the question of whether individuals have the financial capabilities to make informed choices. Given potential adverse consequences of poor choices, it is imperative to understand the extent of financial capability among superannuation fund members in order to inform the development of education programs that address specific needs and deficiencies in financial literacy levels. This study takes initial steps towards providing empirical evidence about the financial literacy of superannuation fund members, with specific focus on decision-making in the superannuation fund context relating to investment choices. While there have been other financial literacy surveys conducted in Australia (see ANZ, 2005; and FLF, 2007) and elsewhere, this is the first study to test more sophisticated financial literacy in the context of superannuation/pension investment decision-making. Based on responses from 2,032 members of QSuper, a large public sector superannuation fund, to an on-line survey questionnaire, the findings show that although the majority of respondents scored highly on general financial and investment matters, they did not perform as well on the more advanced financial literacy measure relating to specific investment products, with large proportions of either incorrect and ‘don’t know’ responses. The analysis also highlights differences among demographic groupings, with women, younger members, and those with lower levels of education generally scoring lower on all three financial literacy measures. Wealth is identified as a discriminating factor in relation to investment matters, with household income and holding other investments in shares found to be positively associated with general and specific investment literacy. Similarly, members who chose the superannuation investment option, use a greater number of different sources of information, and consult experts to aid them in financial decision-making, tend to have higher levels of investment literacy. Comparisons of subjective tests of financial literacy and objective tests indicate that individuals appear to accurately assess their own literacy in general financial and investment matters. However, in relation to specific investment literacy, only those with lower levels of financial literacy on the objective tests correctly self-rated their financial literacy as poor. Those who self-rated their financial literacy as high tended to achieve low scores on the objective tests for  C 2011 Blackwell Publishing Ltd FINANCIAL LITERACY AND PENSION INVESTMENT DECISIONS 305 advanced investment literacy, indicating there may be over-confidence of skills on more advanced investment matters. Overall, the findings show that respondents in this study generally have good basic financial literacy, but significant deficiencies are identified in relation to more advanced investment matters. The analysis highlights demographic groupings of members who are more likely to have lower levels of financial literacy and therefore at greatest risk of making sub-optimal superannuation investment decisions. These findings emphasise the need for development of education programs to address shortcomings in superannuation fund members’ investment knowledge and skills to ensure they are equipped with appropriate levels of financial literacy to make informed investment decisions. A limitation of this study is that it examines the financial literacy of only one of the large superannuation funds in Australia, and is also limited to public sector employees. Nevertheless, it identifies that, on average, this sample of members generally lack the financial understanding necessary for investment decisionmaking. Lack of financial skills and apparent over-confidence among some of those members in choosing investment options potentially leads to undesirable and unexpected financial outcomes for those individuals. The long-term financial well-being of those individuals critically hinges on the superannuation choices they make throughout their working lives. If individuals make wrong choices resulting in inadequate superannuation savings to fund their retirement, they are likely to fall back on the safety net of the government-provided age pension and thus increase the tax burden on future generations. Finally, the overarching question is: whose responsibility is it to educate superannuation fund members to facilitate informed investment decisionmaking? As part of the recent inquiry into superannuation, the Parliamentary Joint Committee on Corporations and Financial Services (PJCCFS) reported that while: Government initiatives can stimulate people’s interest [in superannuation] and provide generic material on the system and interpretive information and advice, . . . funds and advisers usually have a more direct input into educating consumers on their own superannuation arrangements (PJCCFS, 2007, p.176). With the responsibility placed squarely with superannuation funds, it is critical that their member education programs are well-targeted and address specific shortcomings in members’ financial literacy. Further and ongoing empirical research into the financial literacy of superannuation fund members is clearly needed to inform education programs and, in turn, provide input to financial literacy and superannuation/pension policy-making in Australia and other jurisdictions. NOTES 1 The term ‘superannuation’ is used in Australia to refer to employment-related retirement benefits, whereas the term ‘pension’ is more commonly used in other countries.  C 2011 Blackwell Publishing Ltd 306 GALLERY, GALLERY, BROWN, FURNEAUX AND PALM 2 In defined contribution superannuation plans (also referred to as accumulation plans), a member’s benefit comprises contributions to the plan, plus earnings on those contributions, less tax and expenses. In defined benefit plans, a member’s benefit is determined by a formula which is typically a multiple of the member’s final average salary just prior to retirement, taking into account years of fund membership and the member’s age. 3 Choice of fund legislation came into effect from 1 July, 2005, requiring many employers to offer their employees a choice of superannuation fund; this requirement does not apply where certain awards and workplace agreements are in place. Most superannuation funds that have an accumulation component offer members investment choice. 4 In 1983, 82% of superannuation fund members were in defined benefit funds, but by 2006, 97% of members were in funds providing either only accumulation benefits or a mix of accumulation and defined benefits (APRA, 2007). 5 See for example FSA (2006) in the UK, and ANZ (2005) and FLF (2007) in Australia. 6 The detailed set of questions used to test financial literacy in this study is available from the authors on request. 7 Details of factor loadings are available from the authors on request. 8 To avoid potential bias that might be caused by presenting the investment options by relative levels of risk/expected returns, the order of investment options presented in the questionnaire is the same as the order in which those options are presented in QSuper’s Product Disclosure Statement. 9 To assess only those members with an accumulation account, members whose main account is defined benefit were removed from the sample; the results are virtually identical. 10 This result does not change when defined benefit members are removed from the sample. 11 Accumulation accounts held by those defined benefit members would generally comprise additional voluntary superannuation contributions made by the members. 12 Since their introduction, PDSs have been criticised for being too long, complex and difficult to understand (see for example PJCCFS, 2007), which may explain why so few respondents used them. 13 The personal income of members was included in tests as an alternative measure of income; the results are qualitatively the same. 14 The finding that members with share investments outside their superannuation are more financially literate is consistent with Banks and Oldfield’s (2007, p.147) ‘reverse causality’ argument. That is, rather than financial literacy leading to the propensity to invest, the act of investment increases financial literacy as individuals seek to increase their financial literacy in order to understand the investments they hold. REFERENCES Agnew, J. and L. Szykman (2005), ‘Asset Allocation and Information Overload: The Influence of Information Display, Asset Choice, and Investor Experience’, The Journal of Behavioral Finance, Vol. 6, No. 2, pp. 57–70. ANZ (2005), ANZ Survey of Adult Financial Literacy in Australia (Available at: http://www.anz.com/ aus/aboutanz/Community/Programs/pdf/ANZ_Survey_2005.pdf). Arnone, W.J. (2005), ‘Educating Pension Plan Participants’, in R.L. Clark and O.S. Mitchell (eds.), Reinventing the Retirement Paradigm (Oxford University Press, Oxford), pp. 163–72. Ashcroft, J. (2009) ‘Defined-Contribution (DC) Arrangements in Anglo-Saxon Countries’, OECD Working Papers on Insurance and Private Pensions No. 35. Australian Prudential Regulation Authority (APRA) (2007), Insight: Celebrating 10 Years of Superannuation Data Collection 1996–2006 (Sydney). ——— (2009), Annual Superannuation Bulletin – June 2008 (Sydney). Australian Securities & Investments Commission (ASIC) (2007), Regulatory Guide 168 Disclosure: Product Disclosure Statements (and other disclosure obligations). Banks, J. and Z. Oldfield (2007), ‘Understanding Pensions: Cognitive Function, Numerical Ability and Retirement Saving’, Fiscal Studies, Vol. 28, No. 2, pp. 143–70. Financial Literacy Foundation (FLF) (2007), Financial Literacy: Australians Understanding Money (Barton). Financial Services Authority (FSA) (2006), Financial Capability in the UK: Establishing a Baseline (London).  C 2011 Blackwell Publishing Ltd FINANCIAL LITERACY AND PENSION INVESTMENT DECISIONS 307 Kempson, E. et al. (2005), ‘Measuring Financial Capability: An Exploratory Study’, Consumer Research No. 37 (Financial Services Authority, London). Lusardi, A. and O. Mitchell (2007), ‘Financial Literacy and Retirement Preparedness: Evidence and Implications for Financial Education’, Business Economics, Vol. 42, No. 1, pp. 35–44. Organisation for Economic Co-operation and Development (OECD) (2003), OECD Guidelines for the Protection of Rights of Members and Beneficiaries in Occupational Pension Plans (Available at: http://www.oecd.org/dataoecd/16/33/34018295.pdf). ——— (2005), Improving Financial Literacy: Analysis of Issues and Policies (OECD Publishing, Paris). ——— (2008), Recommendation on Good Practices for Financial Education Relating to Private Pensions (Available at: http://www.oecd.org/dataoecd/4/21/40537843.pdf). Parliamentary Joint Committee on Corporations and Financial Services (PJCCFS) (2007) The Structure and Operation of the Superannuation Industry (Canberra). Schagen, S. and A. Lines (1996), Financial Literacy in Adult Life (National Foundation for Educational Research, UK). Tabachnick, B.G. and L.S. Fiddell (2007), Using Multivariate Statistics (5th ed., Pearson Education, Boston). van Rooij, M. et al. (2007), ‘Financial Literacy and Stock Market Participation’, NBER Working Paper No. W13565. Worthington, A.C. (2006), ‘Predicting Financial Literacy in Australia’, Financial Services Review, Vol. 15, No. 1, pp. 59–79.  C 2011 Blackwell Publishing Ltd