How Fact Sheets affect retirement income product knowledge,perceptions and choices

Abstract

Using an online experimental survey, we test how Fact Sheets presenting standardised information on retirement income products – Average Annual Income, an Income security Product Rating, Potential Income Shape, Access to Capital, and Death Benefits – influence product knowledge, perceptions and choices from a menu of annuity, phased withdrawal and hybrid products. Product choice is significantly influenced by annual income and income security, and the annuity and hybrid annuity products are most preferred in Fact Sheet versions where income security is most salient. We find no significant differences in knowledge of Fact Sheet information by version, with Access to Capital least understood in all versions.

JEL Classification:

D14, D83, D91

Keywords

Information disclosure product perceptions retirement income products

1. Introduction

Defined Contribution (DC) pension plans, in which retirees have flexibility in decumulation, are becoming widespread globally. In Australia, the setting for this study, workers have been obliged to belong to DC superannuation funds¹ since 1992, following the introduction of the superannuation guarantee which mandates employer contributions to individual accounts. As the system has matured, the balances at retirement in these accounts have grown, and superannuation fund assets in Australia are the fourth largest globally at 130% of gross domestic product (GDP) in mid-2020 (Organisation for Economic Co-operation and Development (OECD), 2021). Well-considered decisions around the decumulation of retirement savings are crucial to ensure the management of retirement risks and a comfortable life in retirement.

Australian retirees have considerable discretion in the drawdown of their retirement savings. At retirement, they can take one or a combination of a lump sum, an annuity or a phased withdrawal product. Most retirees take phased withdrawal products, known in Australia as account-based pensions, which provide a regular income (guided by tax-driven minimum drawdown rules) until the account balance is fully withdrawn (Australian Prudential Regulation Authority (APRA), 2021). While account-based pensions provide flexibility of withdrawal, they do not cover retirees for risks they may face in retirement relating to longevity, financial markets and/or inflation.

Retirement income products can be complex, particularly if designed to simultaneously address both income replacement and risk management. In Australia, there is low awareness and poor understanding of retirement income products and their features (Bateman et al., 2018), and low take-up by older Australians of retirement income products except account-based pensions (APRA, 2021). To partially address these issues, the Australian Treasury proposes a one-page Fact Sheet for retirement income products to provide standardised information on Average Annual Income, income security (represented as a Product Rating), Potential Income Shape, Access to Capital, and Death Benefits. The stated goal of the Fact Sheet is to improve comparability between products to enable consumers to select the retirement income product which is in their best interests (The Treasury, 2018b).

In this article, we test the effectiveness of the Fact Sheet as a tool to understand, compare and choose retirement income products.² We conduct an online survey incorporating choice tasks of retirement income products using variants of the proposed Fact Sheet with a representative sample of just over 1000 Australians close to the retirement age. The effectiveness of prescribed information to assist with retirement benefit decisions has been little examined in Australia or internationally.

The preference of Australian retirees for account-based pensions over products which offer insurance against retirement risks can be attributed to a combination of demand-side, supply-side, distribution and regulatory factors (Bateman and Piggott, 2011; Chomik et al., 2018; Iskhakov et al., 2015). On the demand-side, the so-called annuity puzzle is a well-known, global phenomenon, where rational factors, such as pre-existing annuitization and bequest motives, and behavioural factors, including framing and product complexity, have been found to explain subdued interest in voluntary annuities (Agnew et al., 2008; Bateman et al., 2017, 2018; Benartzi et al., 2012; Boyer et al., 2020; Brown et al., 2008, 2017, 2021; Lambregts and Schut, 2020). In Australia, the widespread eligibility to the means-tested, public lifetime annuity called Age Pension³ exacerbates these demand-side impediments (Iskhakov et al., 2015). Moreover, taking an account-based pension is the course of least action: it requires, at the very least, the conversion of an accumulation account to a decumulation account within the same superannuation fund, and is therefore likely to be associated with the stickiness of a default.

On the supply-side, the Australian market for retirement income products, particularly those providing longevity insurance such as life annuities, is small. Frictions in product distribution exist due to the absence of annuity products on many superannuation fund product menus and the lack of alignment with financial advisor incentives. In Australia, in 2021, there is only one active provider of lifetime annuities, down from over a dozen 20 years ago (Bateman and Piggott, 2011). Moreover, until recently, there were tax, Age Pension means test and regulatory barriers to the offer and take-up of a broader range of retirement income products (Department of Social Services, 2019).

It is in this context that the Financial System Inquiry, which reported in 2014, recommended that Australia’s superannuation system strengthened its focus on providing retirement incomes (Commonwealth of Australia, 2014: 90–91). The response from Australia’s policymakers has been to develop a ‘retirement income framework’ which includes the disclosure of standardised information about retirement income products on a one-page Fact Sheet (The Treasury, 2018b).⁴

Using choice tasks embedded in an online survey, we explore how the information items in the proposed Fact Sheet influence understanding, perceptions and choice of retirement income products from a menu which includes a lifetime annuity, a hybrid lifetime annuity/account-based pension, a hybrid account-based pension/deferred annuity and an account-based pension.

We find that the understanding of the Fact Sheet information items is similar between participants who receive different Fact Sheets. Level of understanding is reasonable, except for the Access to Capital information item. Regarding product-specific knowledge, the annuity is the best understood product. Although the account-based pension is the dominant product in the Australian market, we find that it is the least understood product in our experimental survey. Participants who receive a Fact Sheet that only shows text have the best (objective) product-specific knowledge. However, perceived (subjective) understanding of the products does not differ between the alternative Fact Sheets.

The alternative Fact Sheet presentations influence perceptions (of risk and control) of the retirement income products, which in turn influence retirement income product choices. For example, participants who receive the Text-Only Fact Sheet perceive the annuity to be riskier and the account-based pension to offer more control than participants who received the Fact Sheet proposed by the Treasury. Participants’ risk perception follows the Product Rating information item, especially in Fact Sheet versions where the Product Rating is salient.

When using the Fact Sheet, participants chose the annuity and the annuity/account-based pension product most often: 33% and 34% of total choices, respectively. Contrary to the actual behaviour of Australian retirees (who predominantly choose account-based pensions), the account-based pension is selected least often (13% of total choices). Of the five information items proposed for inclusion in the Fact Sheet, the Product Rating (a 1–7 rating of protection against a fall in income due to inflation, market and longevity risk) and Average Annual Income are the most important drivers of decisions to choose retirement income products. In particular, the lifetime annuity and its hybrid variant are more likely to be chosen in Fact Sheet treatments where the Product Rating is more salient. We conclude that the Fact Sheets must be carefully designed to avoid unintentional influence on product choice.

The notion that standardised disclosure helps consumers make better choices is based on the assumption that consumers are ‘rational’ and maximise their utility with complete information (Janger and Block-Lieb, 2006; Loewenstein and O’Donoghue, 2006). This thinking underlies a global practice of developing information disclosure to facilitate decision-making in domains such as health, energy and financial behaviour (e.g. see Day and Brandt, 1974; Howlett et al., 2009; Ikonen et al., 2020; Markard and Holt, 2003).⁵ The broad conclusions of this literature are mixed with a large subset of papers finding that information disclosure is either ineffective or used in unexpected ways – a conclusion confirmed in a joint review by the Australian Securities and Investments Commission and the Dutch Authority for the Financial Markets (ASIC and AFM, 2019) titled ‘Disclosure: Why it Shouldn’t be The Default’.

However, prescribed information provision continues to be used by regulators, so it is important to continue to test alternative approaches to information disclosure formats. Furthermore, retirement income product disclosure is one of the forms of guidance (along with online calculators and other decision tools) to be provided to Australian superannuation fund members to assist with drawdown decisions under the Retirement Income Covenant (The Treasury, 2021). Apart from Australia, several other jurisdictions including Canada, the European Union and the United States are considering but have not yet introduced disclosure formats for retirement benefits (European Insurance and Occupational Pensions Authority (EIOPA), 2018; Office of the Superintendent of Financial Institutions (OSFI), 2016; US Securities and Exchange Commission (SEC), 2018). As far as we are aware, there has been little analysis of disclosure formats for retirement income products. An exception is Hiscox et al. (2017), a study by the Behavioural Economics Team of the Australian Government (BETA) conducted as a background study to the Retirement Income Disclosure Consultation Paper (The Treasury, 2018b). They analyse the impact of seven different information treatments on willingness to choose between an account-based pension and a hybrid product and find that information presented in text tables was clearest and most effective.

We contribute to the literature on the effectiveness of prescribed information provision for retirement income products in four ways. First, we test understanding of information items on the proposed Fact Sheet for retirement income products developed by the Australian Treasury. Second, we explore how subjective and objective understanding and perceptions of typical retirement income products are influenced by the information items on the proposed Fact Sheet.

Third, we analyse how stated choices between four illustrative retirement income products are influenced by product perceptions and the information items proposed for inclusion on the Fact Sheet. Finally, we test how personal characteristics such as financial competence and product knowledge, risk and time preferences, retirement planning, personality traits and demographics influence choices made using different versions of the Fact Sheet. As far as we are aware, there are no previous studies of disclosure formats for retirement benefits.

The remainder of this article is structured as follows. Section 2 outlines the design of the online choice tasks. Section 3 describes the sample and descriptive statistics, while section 4 presents results of regression models exploring the relationship between retirement income product perceptions and knowledge and product choices. Section 5 discusses our findings and concludes.

2. Survey design

We conducted an online survey with an embedded choice task of retirement income product decisions in Australia in October and November 2019, with 1007 participants drawn from the commercial web panel Pureprofile. We targeted participants aged between 55 and 66 years who had a superannuation account but had not yet started to decumulate. We specifically focused on those who were near retirement and would need to make a decumulation decision in the near future. Participants were contacted by email by the web panel provider and were paid a fixed amount for participation. They could also receive a bonus payment based on their score in a comprehension quiz which tested knowledge of the retirement income products and understanding of information items included in a retirement income product Fact Sheet.

In the choice task, participants were asked to make 12 pairwise choices between four retirement income products using a retirement income product Fact Sheet informed by the Retirement Income Disclosure Consultation Paper (The Treasury, 2018b). The four retirement income products were as follows:

An inflation-indexed immediate life annuity (Annuity).

A hybrid product with 80% (of the purchase price) allocated to an inflation-indexed immediate life annuity and 20% to an account-based pension (ABP) invested in a balanced option (Annuity/ABP).⁶

A hybrid product with 77% (of the purchase price) allocated to an account-based pension invested in a balanced option and drawn down at the minimum rate and 23% to a deferred life annuity (DA) from age 85 years (ABP/DA).

An account-based pension invested in a balanced option and drawn down at the minimum rate (ABP).

To address possible bias due to real-world familiarity with the retirement income products, we avoided commercial product names. In the survey, we referred to the Annuity as ‘Product A’, and the Annuity/ABP, ABP/DA and ABP as ‘Product B’, ‘Product C’ and ‘Product D’, respectively. All products were fairly priced.⁷

The proposed Fact Sheet includes five information items considered relevant to retirement income product choice (The Treasury, 2018b) as follows:

Average Annual Income (net of fees and taxes) the product could provide for a purchase price of US$100,000 from retirement at age 67–97 years.

Product Rating (a measure of protection against downside income variation, expressed in terms of ‘income security’) calculated as downside variation of actual income from benchmark income (expected first year real income) for the life of the retiree due to the combination of market, inflation and longevity risks. It is presented on a scale from 1 (average income not secure) to 7 (average income secure).⁸ The numeric Product Rating is accompanied by a dot point summary of the underlying market, inflation and longevity risks.

Potential Income Shape of average annual income over the life of the product, illustrated in a line graph in terms of best (95th percentile), median and worst (5th percentile) outcomes.

Access to Capital showing the maximum amounts a consumer could withdraw if they decide to sell the product at specific ages, which vary depending on the type of product and the capital access schedule (Commonwealth of Australia, 2017).

Death and Reversionary Benefits.

Under the Product Rating metric, the Annuity offered in our study is rated 7 (average income secure), while the other retirement income products are rated less secure, at 5 (Annuity/ABP), 3 (APB/DA) and 2 (ABP).

In the choice tasks, the wording, order and presentation of the five information items are predominantly proposed by The Treasury in the Retirement Income Disclosure Consultation Paper (The Treasury, 2018b). The Fact Sheets were qualitatively tested with two focus groups in June 2019: one with people aged between 55 and 66 years, not yet retired, and the other with retirees aged 60 years and older. The focus group discussion provided valuable insights into areas of potential (mis)understanding by survey participants, and minor edits were made to the survey accordingly. In particular, to address concern about the possibility of product providers manipulating the information – such as the Product Rating – we added a footnote to the Fact Sheet to clarify that the calculations of the Product Rating, Access to Capital, and Death Benefit follow government regulations. We did this to enhance participant trust of the information on the Fact Sheets.

The experiment design also included one within-subject treatment to test the effect of exclusion of the information item ‘Potential Income Shape’, as we hypothesised that this may somewhat replicate the information already provided in the Average Annual Income and Product Rating. As such, given four retirement income products, participants completed six pairwise product choices with and six pairwise product choices without the Potential Income Shape information item. The order of the within-subject treatment was randomised.

We tested four variants of the Fact Sheet (four treatments) in a between-subjects design, with survey participants randomly assigned to one of the four treatments as follows:⁹

Treasury-Graph includes the information items as proposed in The Treasury (2018b) with two information items – Access to Capital and Death Benefit – presented in a single graph.

Treasury-Table includes the information items as proposed in The Treasury (2018b), with Access to Capital and Death Benefit presented in a single table.

Text-Only has the same information items as the Treasury-Graph and Treasury-Table but presents the information items in text only.

The No-Rating treatment uses the same design as the Treasury-Graph treatment but excludes the numeric Product Rating.¹⁰

Examples of the Fact Sheets we tested can be found in Appendix C (Online Appendix).

After completion of the 12 product choices, participants were asked a series of questions about trust in and perceptions of the Fact Sheets and products, and to test knowledge. First, they were asked about their perception of the Fact Sheets in terms of transparency of the product provider, trust in the product provider and control over the product. The next set of questions canvassed their subjective understanding of the four retirement income products and their perceptions of the riskiness of, and degree of control offered by, the products. They then answered two sets of questions to elicit their understanding of the Fact Sheet information items and their objective knowledge of the retirement income products offered. This included questions about which products provide a regular income, allow flexible drawdowns, include a death benefit after age 85 years, provide inflation-indexed payments, provide income influenced by changes in asset returns and provide unlimited access to capital. We ask all participants except those in the No-Rating treatment group whether a higher Product Rating means that the retirement income product provides less risky income. Finally participants completed standard questions on financial literacy (Lusardi and Mitchell, 2011); numeracy (Lipkus et al., 2001); graph literacy (Galesic and Garcia-Retamero, 2011); trust in relation to superannuation funds, the financial services industry and government (Agnew et al., 2012; Koh et al., 2021); risk preferences (Dohmen et al., 2011) and time preferences (Vischer et al., 2013); psychological traits (Gosling et al., 2003); bequest motives; and standard demographics including wealth, income, age, gender and marital status. We included questions on trust in product providers, financial institutions and government to explore whether previous findings on the association between trust and positive retirement decisions also applied in our study.¹¹ Selected screen shots from the survey can be found in Appendix C (Online Appendix).

3. Descriptive statistics

This section first describes our sample, including demographics, actual retirement plan characteristics and financial competence. We then report summary statistics for the objective understanding of the Fact Sheet information items, objective and subjective understanding as well as perceptions of the retirement income products offered in the choice task and retirement income product choices both in aggregate and by Fact Sheet treatment.

3.1. Sample

Our sample is broadly representative of the Australian population in our targeted age group (55–66 years), as illustrated in Appendix A (Online Appendix). Our sample has a similar gender mix but is younger, slightly less likely to be partnered and more educated than the Australian population in the same age range. As we require that survey participants be members of a super fund in the accumulation phase, we are not surprised that our sample is more likely to be employed and receive a higher income than Australians in that age group on average.

As illustrated in Table 1, the average age of the full sample is 60 years old and 42% are male; 63% of participants are either married or in a long-term relationship. On average, participants have two children. A total of 81% have completed tertiary education, and 83% are either full- or part-time employed. The average household wealth is US$992,300.¹² 93% report being a member of a large superannuation fund and 9% of a self-managed superannuation fund.¹³ Just over 50% of participants reported that they belong to a DC scheme, 16% reported being a member of a DB scheme and one third of participants did not know the type of super scheme to which they belonged. On average, participants answered 2.48 out of 3 questions correctly on financial literacy, 1.67 out of 3 questions correctly on numeracy, 3.55 out of 6 questions on superannuation and 3.24 out of 4 questions on graph literacy. Responses to retirement planning questions were mixed. About 54% of participants had worked out how much money they needed for retirement; 21% reported that they use a financial adviser, while 42% currently without an adviser plan to use one in the future. A total of 84% of participants stated that they intended to leave a bequest. Randomisation checks find that the treatment groups’ demographics do not differ statistically significantly for any demographic variable listed in Table 1 except for marital status (F statistic = 2.83, p = 0.0372).

Table 1.

Descriptive statistics.

	M	SD
Demographics
Age	59.65	3.26
Gender (% female)	58%
Marital status (% married/partnered)	63%
No. of children	1.85	1.48
Education (% tertiary)	81.13%
Employed (%, full or part time)	83.41%
Household wealth (US$ ‘000)	992.3	6.739
Subjective life expectancy	84.62	8.73
Type of superannuation fund (%)
Large superannuation fund	93%
Self-managed superannuation fund	9%
Do not know	2%
Benefit type (%)
Defined contribution	51%
Defined benefit	16%
Do not know	33%
Financial competence (number correct)
Financial literacy (out of 3)	2.48	0.78
Financial literacy – Don’t Knows (out of 3)	0.3	0.6
Numeracy (out of 3)	1.67	1.06
Superannuation knowledge (out of 6)	3.55	1.07
Graph literacy (out of 4)	3.24	1.02
Planning for retirement (%)
Worked out how much need for retirement	54%
Financial advice (%)
Has adviser now	21%
Plans to have adviser in future	42%
Preferences (Likert-type scale: 1–10)
Patience	6.49	2.333
General risk preference	4.14	2.52
Financial risk preference	3.70	2.64
Bequest (% yes)	84%

A complete set of summary statistics can be found in Appendix B (Online Appendix).

3.2. Understanding of Fact Sheet information items

Responses to questions testing understanding of the Fact Sheet information items were modest. Two thirds of participants understood that Average Annual Income does not refer to income in the first year of payment. About 68% of participants in the Treasury-Graph, Treasury-Table and Text-Only treatment groups understood that the Product Rating is a measure of income security, and 58% of all participants answered correctly that it measures three types of risks; 15% of participants answered ‘Don’t Know’ to the question about the risk types, and 7% to the question about the Product Rating and income security. There was slightly less confusion with the Death Benefit question (79% answered correctly, 13% answered ‘Don’t Know’). Participants were most confused with the information item Access to Capital. Only one third of participants correctly answered the question, while 11% answered that they ‘Don’t Know’. It appears that many participants think Access to Capital gives them information on how much money is always available to them as a lump sum, rather than how much initial capital is returned when they sell the product. Of participants in the Treasury-Table, Treasury-Graph and Text-Only treatment groups, 3.88% answered none of the information item questions correctly, 96.12% had at least one correct answer, 86.25% at least two correct answers, 67.51% at least three, 39.36% at least four and only 12.2% of participants answered all five questions correctly. We found no statistical differences in information item knowledge between the three treatments. Of participants allocated to the No-Rating treatment, 5.71% answered none of the information item questions correctly, 94.29% answered at least one correctly, 83.81% at least two, 55.24% at least three and 19.05% answered all four information item questions correctly. When we take the four questions all participants answered, knowledge about the information items does not differ statistically significantly between treatment groups (Table 2, Panel A, rows 1–2).

Table 2.

Knowledge and perceptions, by Fact Sheet treatment.

	Treasury-Graph	Treasury-Table	Text-Only	No-Rating	F statistic	Full sample
Panel A: Objective knowledge of
Information items	2.268 (1.08)	2.379 (1.15)	2.354 (1.11)	2.524 (1.09)	1.47	2.35 (1.11)
Information items – Don’t Knows	0.54 (1.01)	0.45 (0.94)	0.48 (0.93)	0.29 (0.78)	1.99	0.47 (0.94)
Annuity	4.16 (1.37)	3.78 (1.29)	4.54 (1.47)	3.97 (1.30)	16.13***	4.14 (1.40)
Annuity/ABP	4.02 (1.68)	3.91 (1.50)	4.02 (1.55)	3.94 (1.72)	0.35	3.98 (1.59)
ABP/DA	3.67 (1.34)	3.423 (1.25)	3.957 (1.22)	3.476 (1.33)	9.66***	3.66 (1.29)
ABP	3.936 (1.58)	3.648 (1.46)	4.013 (1.70)	3.924 (1.54)	3.01*	3.87 (1.59)
Panel B: Subjective understanding of
Annuity	4.84 (1.34)	4.91 (1.29)	4.76 (1.24)	4.74 (1.24)	0.80	4.83 (1.28)
Annuity/ABP	4.80 (1.32)	4.85 (1.25)	4.73 (1.23)	4.70 (1.17)	0.61	4.78 (1.26)
ABP/DA	4.68 (1.32)	4.76 (1.24)	4.69 (1.23)	4.65 (1.16)	0.34	4.71 (1.25)
ABP	4.656 (1.35)	4.733 (1.26)	4.710 (1.26)	4.58 (1.26)	0.46	4.69 (1.29)
Panel C: Perceived riskiness of
Annuity	3.39 (1.51)	3.26 (1.38)	3.61 (1.74)	3.64 (1.30)	3.26*	3.44 (1.53)
Annuity/ABP	3.70 (1.28)	3.53 (1.15)	3.82 (1.35)	3.85 (1.09)	3.33*	3.70 (1.25)
ABP/DA	4.32 (1.10)	4.12 (1.05)	4.45 (1.03)	4.16 (0.93)	5.38**	4.29 (1.06)
ABP	4.76 (1.24)	4.55 (1.25)	4.66 (1.32)	4.59 (1.05)	1.54	4.65 (1.25)
Panel D: Perceived control with
Annuity	3.97 (1.49)	4.14 (1.39)	3.78 (1.40)	4.17 (1.21)	4.03**	3.98 (1.41)
Annuity/ABP	4.00 (1.38)	4.07 (1.24)	3.95 (1.26)	4.04 (1.17)	0.45	4.01 (1.28)
ABP/DA	3.67 (1.28)	3.82 (1.18)	3.74 (1.25)	4.01 (1.14)	2.15	3.77 (1.23)
ABP	3.58 (1.42)	3.69 (1.30)	3.90 (1.34)	3.70 (1.26)	3.11*	3.72 (1.35)
Observations	299	298	305	105		1007

This table shows the means and standard deviations (in parentheses) by treatment and in aggregate. ABP: account-based pension; DA: deferred annuity. The maximum score for information item knowledge is four, and for the objective product-specific knowledge, six. The perceptions were rated on a Likert-type scale ranging from 1 to 7.

***

p < 0.001, ^**p < 0.01, ^*p < 0.05.

3.3. Retirement income product knowledge

We also tested the participants on their knowledge of the four retirement income products (product-specific knowledge) after they had read the Fact Sheets and completed the product choice tasks. The mean (median) score was 15.66 (15) correct answers out of a possible 24, with a minimum of six correct answers. Table 2 (Panel A, rows 3–6) reports the number of correct answers for each product by Fact Sheet treatment. On average, participants answered 4.14 of six questions correctly for the Annuity, 3.98 for the Annuity/ABP, 3.66 for the ABP/DA, and 3.87 for the ABP. We ran ANOVA’s to test for differences in product knowledge between the Fact Sheet treatment groups. The average number of correct answers differ between Fact Sheets for the Annuity and the ABP/DA (both at the 0.1% significance level). Participants in the Text-Only treatment group outperform the other three Fact Sheet treatment groups (statistically significant at the 1% significance level, pairwise comparisons with Bonferroni-adjusted standard errors). Participants in the Treasury-Table treatment group underperform participants in the other groups (statistically significant at the 1% significance level except for the comparison with the No-Rating treatment group). Furthermore, even though the ABP is the retirement income product of choice outside the experiment, knowledge of the ABP by survey participants was no better than the three other retirement income products.

3.4. Retirement income product perceptions

To explore the role of perceptions about the retirement income products, we asked participants about their perceived understanding of each product, how risky they perceive each product to be and how much control they perceive they would have with each product (White et al., 2019). The perceptions of the products by Fact Sheet treatment are summarised in Table 2, Panels B, C and D. On average, participants perceived understanding of the Annuity, and the Annuity/ABP was highest (4.83 and 4.78, respectively), followed by the ABP/DA (4.71) and the ABP (4.69). The differences between Fact Sheet treatment groups are not statistically significant (Table 2, Panel B).

If participants understood the information on the Fact Sheet, we would expect perceptions of product riskiness to be negatively correlated with the Product Rating (i.e. the measure of income security). This is generally the case, as the perceived riskiness of the four products is lowest for the Annuity (rated 7), followed by the Annuity/ABP (rated 5), the ABP/DA (rated 3) and then the ABP (rated 2). The differences between treatment groups are statistically significant for the ABP/DA (99% significance level) and the Annuity and the Annuity/ABP (95% significance level). Specifically, participants in the Text-Only treatment group are more likely to rate the ABP/DA riskier than those in the Treasury-Graph group (0.32, p = 0.001) (Table 2, Panel C).

We would also expect that participants understood that they would have more control with the ABP than with the other products, as the Access to Capital is always greater. This is not generally the case. In fact, the ABP is rated as providing the least control in all treatment groups except Text-Only where it ranks second, after the Annuity/ABP. Differences between perceptions of control are statistically significant for the Annuity (p < 0.01) and the ABP (p < 0.05). Participants in the Text-Only group perceive less control with the Annuity than those in the Treasury-Graph group (−0.36, p = 0.01) (Table 2, Panel D).

3.5. Retirement income product choices

We now turn to retirement income product choices after participants have viewed the Fact Sheets. We find that in aggregate across all pairwise choices for four Fact Sheet treatment groups, participants were most likely to choose the Annuity and the Annuity/ABP, which they did 33% and 34% of the time, respectively (see Figure 1). The ABP/DA was chosen 20% of the time, and in contrast to the actual behaviour of Australian retirees (who predominantly choose ABPs) (APRA, 2021), the ABP was least preferred by survey participants, chosen 13% of the time. To put the current behaviour of Australian retirees in context, we note that typically the ABP is the only retirement income product on offer on superannuation fund product menus, the two hybrid products included in our study (Annuity/ABP, ABP/DA) are not currently offered and lifetime annuities have a very low profile in Australia and are subject to low awareness and understanding (Bateman et al., 2018).¹⁴ The low demand for lifetime annuities in Australia is in line with the international experience (APRA, 2021; Bateman and Piggott, 2011; Chomik et al., 2018). Table 3 shows the preferences for retirement income products by Fact Sheet treatment. Columns 2–4 present the percentage of times a product is chosen in all pairwise choice settings. Columns 5–7 report how many times a retirement income product is chosen when it was available in the pair. The second and fifth columns report the percentages when Potential Income Shape information item is not shown (labelled ‘4 info. items’), and the third and sixth columns report the percentages when it is shown (labelled ‘5 info. items’).

Table 3.

Preferences for retirement income products.

	Treasury-Table treatment							Treasury-Graph Treatment
	% chosen in all choice settings			% chosen in choice settings where the product was available				% chosen in all choice settings			% chosen in choice settings where the product was available
	4 info. items	5 info. items	Total	4 info. items	5 info. items	Total		4 info. items	5 info. items	Total	4 info. items	5 info. items	Total
Annuity	33.89	34.39	34.14	67.78	68.78	68.28	Annuity	37.64	37.53	37.58	75.28	75.06	75.17
Annuity/ABP	34.23	36.23	35.23	68.45	72.46	70.46	Annuity/ABP	33.05	35.23	34.14	66.11	70.47	68.29
ABP/DA	18.06	18.23	18.14	36.12	36.45	36.29	ABP/DA	20.92	20.97	20.95	41.83	41.95	41.89
ABP	13.82	11.15	12.49	27.65	22.30	24.97	ABP	8.36	6.26	7.33	16.78	12.53	14.65
N	1794	1794	3588	897	897	1794	N	1788	1788	3576	894	894	1788
	Text-Only treatment							Treasury-Graph treatment
	% chosen in all choice settings			% chosen in choice settings where the product was available				% chosen in all choice settings			% chosen in choice settings where the product was available
	4 info. items	5 info. items	Total	4 info. items	5 info. items	Total		4 info. items	5 info. items	Total	4 info. items	5 info. items	Total
Annuity	25.63	26.83	26.23	51.26	53.66	52.46	Annuity	33.33	33.02	33.17	66.67	66.03	66.35
Annuity/ABP	32.51	33.11	32.81	65.03	66.23	65.63	Annuity/ABP	28.57	33.65	31.11	57.14	67.30	62.22
ABP/DA	18.58	20.66	19.62	37.16	41.31	39.23	ABP/DA	27.62	25.40	26.51	55.24	50.79	53.02
ABP	23.28	19.40	21.34	46.56	38.80	42.68	ABP	10.48	7.94	9.21	20.95	15.87	18.41
N	1830	1830	3660	915	915	1830	N	630	630	1260	315	315	630

ABP: account-based pension; DA: deferred annuity. Bold percentages indicate the most preferred product within participants in the specific treatment group.

Figure 1.

Choice probabilities between the four retirement income products.

In the Treasury-Table and Text-Only treatment groups, the Annuity/ABP is most preferred, followed very closely by the Annuity. In the Treasury-Graph treatment, the Annuity is most preferred, followed closely by the Annuity/ABP. In the No-Rating treatment group, the Annuity is also most preferred overall as well as in the within-subject treatment with four information items. The ABP is the least preferred in all treatments except the Text-Only treatment. In the Text-Only treatment, the ABP is chosen twice to nearly three times as much than in the other three treatments. We also find that the ABP is less preferred when the Potential Income Shape information item is shown, compared with the treatment where it is not. The Annuity/ABP is more preferred when Potential Income Shape is shown. These summary statistics suggest that the ABP is preferred where the Product Rating is presented as text rather than a bold image and therefore less salient (Text-Only treatment), where the information on Death Benefits and Access to Capital is less detailed (Treasury-Table treatment) and where Potential Income Shape is excluded, as the minimum drawdown settings inhibit a smooth payment pattern (as illustrated in Appendix C (Online Appendix): Retirement income product Fact Sheets).

4. Regression analysis: product knowledge, perceptions and product choice

In this section, we further explore the relationship between retirement income product knowledge, perceptions and product choice in a series of multivariate regressions. First, we report regression models analysing how retirement income product knowledge and perceptions vary by Fact Sheet treatment. We then report regression models analysing the associations between understanding of the Fact Sheet information items, product-specific knowledge and product perceptions and retirement income product choice for each Fact Sheet treatment.

4.1. Retirement income product knowledge and product Fact Sheets

First, we examine the association between the alternative product Fact Sheets and objective knowledge of the four retirement income products – Annuity, Annuity/ABP, ABP/DA and ABP. In Table 4, Panel A reports the coefficients of ordinary least squares (OLS) regressions of the Fact Sheet treatments on the objective product knowledge for each retirement income product, with Treasury-Table as the baseline treatment group. We add information item knowledge, number of Don’t Knows for the information item knowledge questions, graph literacy, superannuation knowledge, financial literacy, number of Don’t Knows for financial literacy questions, numeracy, age, gender, being married or in a long-term relationship, being potentially eligible for a full or part Age Pension and being a member of a Defined Contribution pension plan as controls.

Table 4.

Panel A. OLS regressions of product-specific knowledge and perceptions.

	Panel A: Product-specific knowledge of				Panel B: Perceived understanding of
	Annuity	Annuity/ABP	ABP/DA	ABP	Annuity	Annuity/ABP	ABP/DA	ABP
Treasury-Graph	−0.40*** (0.10)	−0.20 (0.12)	−0.26* (0.10)	−0.38*** (0.11)	0.06 (0.10)	0.03 (0.10)	0.06 (0.10)	0.07 (0.10)
Text-Only	0.32** (0.11)	−0.09 (0.12)	0.25* (0.10)	−0.02 (0.12)	−0.12 (0.09)	−0.07 (0.09)	0.01 (0.10)	0.06 (0.10)
No-Rating	−0.18 (0.14)	−0.20 (0.18)	−0.22 (0.15)	−0.14 (0.16)	−0.14 (0.13)	−0.13 (0.12)	−0.07 (0.12)	−0.10 (0.13)
Information item knowledge	0.11* (0.05)	0.21*** (0.05)	0.08 (0.05)	0.27*** (0.05)	−0.05 (0.04)	−0.07 (0.04)	−0.05 (0.04)	−0.10* (0.05)
Information item knowledge – Don’t Knows	0.01 (0.06)	−0.14* (0.06)	−0.03 (0.05)	−0.15* (0.06)	−0.42*** (0.05)	−0.39*** (0.05)	−0.39*** (0.05)	−0.37*** (0.05)
Graph literacy	0.05 (0.05)	0.18*** (0.05)	0.00 (0.04)	0.15** (0.05)	0.04 (0.04)	0.02 (0.04)	0.03 (0.04)	0.02 (0.04)
Superannuation knowledge score	0.11* (0.04)	0.14** (0.05)	0.10* (0.04)	0.14** (0.05)	0.01 (0.04)	−0.01 (0.04)	−0.01 (0.04)	0.02 (0.04)
Financial literacy score	−0.04 (0.08)	0.31*** (0.09)	0.05 (0.08)	0.25** (0.09)	0.00 (0.09)	−0.04 (0.08)	0.01 (0.08)	0.08 (0.09)
Financial literacy – Don’t Knows	0.03 (0.11)	0.23 (0.12)	0.08 (0.10)	0.31** (0.11)	−0.18 (0.11)	−0.17 (0.11)	−0.18 (0.11)	−0.09 (0.11)
Numeracy score	0.31*** (0.05)	0.14** (0.05)	0.22*** (0.04)	0.21*** (0.05)	0.09* (0.04)	0.08* (0.04)	0.08* (0.04)	0.04 (0.04)
Age	−0.01 (0.01)	−0.03 (0.01)	−0.00 (0.01)	−0.03* (0.01)	−0.00 (0.01)	0.01 (0.01)	0.01 (0.01)	0.01 (0.01)
Male	−0.16 (0.08)	−0.22* (0.09)	−0.13 (0.08)	−0.35*** (0.09)	0.05 (0.08)	0.05 (0.07)	0.10 (0.08)	0.13 (0.08)
Married/Long-term relationship	0.18* (0.09)	−0.07 (0.10)	−0.02 (0.08)	0.01 (0.09)	−0.00 (0.08)	0.08 (0.08)	0.00 (0.08)	0.10 (0.08)
Full/Part Age Pension	−0.05 (0.10)	−0.00 (0.11)	−0.06 (0.10)	0.01 (0.12)	−0.14 (0.09)	−0.13 (0.09)	−0.19* (0.09)	−0.22* (0.09)
Defined Contribution	−0.03 (0.08)	−0.03 (0.09)	−0.09 (0.08)	0.07 (0.09)	0.22** (0.08)	0.24** (0.07)	0.20** (0.07)	0.20* (0.08)
Product-specific Knowledge					0.09** (0.03)	0.14*** (0.02)	0.02 (0.03)	0.09*** (0.03)
Constant	3.75*** (0.81)	3.25*** (0.88)	3.01*** (0.78)	3.17*** (0.90)	4.76*** (0.75)	4.04*** (0.75)	4.42*** (0.75)	3.54*** (0.77)
Observations	24,168	24,168	24,168	24,168	24,168	24,168	24,168	24,168
Adjusted R²	14.75%	17.16%	9.50%	20.86%	17.57%	19.09%	15.03%	15.56%
	Panel C: Perceived risk of				Panel D: Perceived control of
	Annuity	Annuity/ABP	ABP/DA	ABP	Annuity	Annuity/ABP	ABP/DA	ABP
Treasury-Graph	−0.15 (0.11)	−0.16 (0.09)	−0.18* (0.09)	−0.17 (0.10)	0.17 (0.12)	0.08 (0.11)	0.14 (0.10)	0.11 (0.11)
Text-Only	0.34** (0.12)	0.17 (0.10)	0.12 (0.09)	−0.13 (0.10)	−0.18 (0.12)	−0.04 (0.11)	0.12 (0.10)	0.38*** (0.11)
No-Rating	0.28 (0.15)	0.18 (0.13)	−0.15 (0.11)	−0.19 (0.12)	0.17 (0.14)	0.03 (0.14)	0.35** (0.13)	0.17 (0.14)
Information item knowledge	−0.10 (0.06)	−0.08 (0.04)	−0.05 (0.04)	0.03 (0.05)	−0.11* (0.05)	−0.12** (0.04)	−0.13** (0.04)	−0.22*** (0.05)
Information item knowledge – Don’t Knows	0.04 (0.06)	−0.02 (0.05)	−0.10* (0.04)	−0.10* (0.05)	−0.29*** (0.06)	−0.28*** (0.05)	−0.22*** (0.05)	−0.25*** (0.06)
Graph literacy	−0.09 (0.05)	−0.08 (0.04)	−0.06 (0.04)	−0.04 (0.04)	−0.03 (0.05)	−0.04 (0.05)	−0.07 (0.04)	−0.02 (0.05)
Superannuation knowledge score	−0.09 (0.05)	−0.05 (0.04)	−0.01 (0.03)	−0.04 (0.04)	−0.05 (0.04)	−0.04 (0.04)	−0.04 (0.04)	−0.01 (0.04)
Financial literacy score	−0.13 (0.09)	−0.00 (0.08)	−0.02 (0.07)	−0.02 (0.08)	−0.29*** (0.09)	−0.26** (0.09)	−0.22** (0.08)	−0.12 (0.08)
Financial literacy – Don’t Knows	−0.16 (0.12)	−0.07 (0.11)	−0.08 (0.09)	−0.12 (0.10)	−0.51*** (0.11)	−0.38*** (0.10)	−0.43*** (0.10)	−0.26* (0.10)
Numeracy score	−0.34*** (0.05)	−0.23*** (0.04)	−0.01 (0.04)	0.09 (0.05)	0.03 (0.05)	0.05 (0.05)	−0.00 (0.04)	−0.05 (0.05)
Age	−0.00 (0.01)	−0.02 (0.01)	−0.02 (0.01)	−0.01 (0.01)	−0.00 (0.01)	0.01 (0.01)	0.01 (0.01)	0.02 (0.01)
Male	0.21* (0.09)	0.18* (0.08)	0.04 (0.07)	0.01 (0.08)	0.06 (0.09)	0.06 (0.08)	0.22** (0.08)	0.30*** (0.09)
Married/Long-term relationship	−0.04 (0.09)	−0.03 (0.08)	0.02 (0.07)	−0.07 (0.08)	−0.13 (0.09)	−0.04 (0.09)	−0.00 (0.08)	0.15 (0.09)
Full/Part Age Pension	−0.20 (0.11)	−0.18 (0.09)	−0.11 (0.08)	−0.09 (0.10)	−0.02 (0.11)	−0.02 (0.10)	−0.06 (0.09)	−0.19 (0.11)
Defined Contribution	0.14 (0.09)	0.05 (0.08)	−0.09 (0.07)	−0.11 (0.08)	−0.04 (0.09)	0.06 (0.08)	0.16* (0.08)	0.11 (0.08)
Product-specific knowledge	−0.15*** (0.03)	−0.12*** (0.02)	0.02 (0.03)	0.13*** (0.03)	−0.04 (0.03)	0.04 (0.03)	−0.06 (0.03)	−0.02 (0.03)
Constant	5.88*** (0.89)	6.22*** (0.76)	5.84*** (0.67)	5.34*** (0.76)	5.74*** (0.86)	4.71*** (0.78)	4.87*** (0.74)	3.28*** (0.81)
Observations	24,168	24,168	24,168	24,168	24,168	24,168	24,168	24,168
Adjusted R²	16.85%	12.77%	3.07%	6.49%	6.50%	5.31%	7.82%	8.28%

OLS: ordinary least squares; ABP: account-based pension; DA: deferred annuity.

Product knowledge measured by the number of correct answers (out of 6) for product-specific knowledge questions. Perceived understanding is rated on a Likert-type scale ranging from 1 to 7. Perceived risk is rated on a Likert-type scale ranging from 1 to 7. Perceived control is rated on a Likert-type scale ranging from 1 to 7. Treasury-Table is the baseline treatment group. Standard errors are in parentheses and clustered at the participant level.

***

p < 0.001, **p < 0.01, *p < 0.05.

Compared with participants in the Treasury-Table treatment group, participants in the Treasury-Graph treatment group score 0.4 less (out of a maximum score of 6) on questions about the characteristics of the Annuity and the ABP (p < 0.001 and 0.001), and 0.3 less on questions about the ABP/DA (p = 0.013). Participants in the Text-Only treatment group know more about the Annuity and the ABP/DA than those in the Treasury-Table treatment group (0.32, p = 0.003; 0.25, p = 0.013). There were no significant differences in objective product knowledge between participants in the No-Rating treatment group and the Treasury-Table group. Overall, understanding of the content of the Fact Sheets (information item knowledge) and financial competency is associated with understanding of the retirement income products. In particular, numeracy is positively correlated with knowledge about all products, with coefficients ranging from 0.31 for the Annuity to 0.14 for the Annuity/ABP. Understanding of the Annuity/ABP and the ABP is also positively correlated with information item knowledge, graph literacy, superannuation knowledge and financial literacy. There was also a gender difference with males scoring 0.22 and 0.35 less (out of 6) than females on questions about the Annuity/ABP and ABP, respectively.

4.2. Retirement income product perceptions and product Fact Sheets

Next, we analyse the association between the alternative product Fact Sheets and perceptions of the four retirement income products in terms of perceived understanding of each product, how risky they perceive each product to be and how much control they perceive they would have with each product. To do so, we regress each of the three concepts of product perception on the Fact Sheet treatments (again with Treasury-Table as the baseline treatment group) and include the control variables used in Table 4, Panel A, with the addition of product-specific knowledge.

Table 4, Panel B, reports the regressions of the perceived understanding of each retirement income product. In contrast to the results for objective product knowledge, the Fact Sheet treatments are not correlated with perceived understanding of any of the four retirement income products. The number of Don’t Knows for information item knowledge – as a proxy for confusion about the Fact Sheet information items – is statistically significant at the 99.9% significance level, and negatively correlated with perceived understanding. The more confused/less confident participants are in their knowledge of Fact Sheet information items, the lower their perceived understanding of the retirement income products. Product-specific knowledge is positively correlated with perceived understanding of all products except for the ABP/DA, statistically significant at the 99% significance level. Graph literacy, superannuation system knowledge and financial literacy are not correlated with perceived understanding, nor is the number of Don’t Knows for financial literacy. Being a member of a Defined Contribution plan is positively correlated with perceived understanding of all products. Participants with a Defined Contribution superannuation scheme might have already sought or have had access to information about retirement income products and therefore perceive their understanding to be good. Participants who are likely to receive a full or part Age Pension have less understanding of the ABP and ABP/DA (significant at the 95% significance level) which they are less likely to purchase.

Table 4, Panel C, reports the coefficients for regressions of the perceived risk of each retirement income product. While the Text-Only treatment is associated with higher product-specific knowledge, it is also associated with higher perceived risk of the Annuity, a product that actually insures against (longevity, investment and inflation) risk. This result is likely because of the reduced salience of the Product Rating information item which summarises income variation risk. However, answering ‘Don’t know’ to the Fact Sheet information item questions is negatively correlated with perceived risk at the 95% significance level for the ABP/DA and ABP. Product-specific knowledge is negatively correlated with perceived risk at the 99.9% significance level for the Annuity and the Annuity/ABP, and positively correlated at the 99% significance level for the ABP. This is unsurprising as the ABP is the riskier product in terms of its failure to insure against the standard retirement risks (longevity, investment and inflation). Numerate participants also understand this, as numeracy is negatively correlated with perceived risk for the Annuity and the Annuity/ABP.

Table 4, Panel D, reports the coefficients for regressions of the perceived control with each retirement income product. Compared with participants in the Treasury-Table treatment group, participants in the No-Rating treatment group perceive that they have more control with the ABP/DA product; participants in the Text-Only treatment group perceive they have more control with the ABP than those in the Treasury-Table treatment group. Information item knowledge is negatively correlated with perceived control, but the impact varies with the type of retirement income product. In contrast to expectations, the smallest effect is found for the annuity products, while the effect is nearly double for the ABP. The number of Don’t Knows for information item knowledge questions is negatively correlated with perceived control. Participants who seem to be confused by the Fact Sheet items are less likely to perceive any product as providing them with control over their finances. Financial literacy is negatively correlated with perceived control for all products except the ABP, a finding that is in line with expectations as none of these products provide complete drawdown flexibility. The number of Don’t Know responses to the financial literacy questions is negatively correlated with perceived control for all products as well. Interestingly, the effect is less pronounced for the ABP, a product that provides more control than the Annuity and hybrid products. Compared with females, males perceive they have more control with the ABP/DA and the ABP.

We do not find a difference between Fact Sheet perceptions of transparency, trust in product provider and feeling of control between the within-subject Potential Income Shape treatments.¹⁵

4.3. Fact Sheets and product choices: Fact sheet information items, product perceptions and product knowledge

Finally, to analyse how the Fact Sheet information items, product perceptions and product-specific knowledge impact retirement income product choice by Fact Sheet treatment, we run logit regressions where we estimate the probability of choosing productj over producti for each Fact Sheet treatment. We cluster the standard errors at the participant level. Table 5 shows the marginal effects for these logit regressions. Annual Income, Product Rating¹⁶ and Access to Capital refer to the Fact Sheet information items and are measured as differences between the respective information items for productj and producti. As the items Death Benefit and Access to Capital are quite similar due to government regulation (and the pairwise correlation between the differences is therefore 0.99), we only use Access to Capital as a covariate. Potential Income Shape enters the model as a dummy variable, being 1 when the Fact Sheets presented include the Potential Income Shape and 0 otherwise. Perceptions of product understanding, product risk and control with the product as well as product knowledge are also defined as differences between productj and producti. We also control for Fact Sheet information item knowledge as well as the number of Don’t Knows for Fact Sheet information item knowledge. To be concise, we do not present coefficients of the covariates for being male, age, having a long-term relationship/being married, having a bequest motive, financial risk preference, financial literacy, number of Don’t Knows for financial literacy, numeracy, being (potentially) eligible for full or part Age Pension, being a member of a DC super fund and trust in superannuation funds and government here. For the full set of covariates, see Appendix E (Online Appendix).

Table 5.

Marginal effects of logit regressions on the choice of productj over producti.

Pr(Choice_j = 1)	Treasury-Table	Treasury-Graph	Text-Only	No-Rating
Annual Income, per US$100	0.02***	0.01***	0.03***	0.02**
Annual Income, per US$100	0	0	0	−0.01
Product Rating	0.06***	0.05***	0.03***	0.03***
Product Rating	0	0	0	−0.01
Access to Capital, per US$100,000	0.01***	0	0.02***	0
Access to Capital, per US$100,000	0	0	0	0
With Income Shape	0.01	0.03*	0.01	0.01
With Income Shape	−0.01	−0.01	−0.02	−0.03
Perceived Understanding	0.03	0.12***	0.07**	0.11**
Perceived Understanding	−0.04	−0.03	−0.03	−0.04
Perceived Risk	−0.03**	−0.03*	−0.08***	−0.01
Perceived Risk	−0.01	−0.01	−0.01	−0.03
Perceived Control	0.05**	0.07***	0.06***	0.11***
Perceived Control	−0.02	−0.02	−0.01	−0.03
Product Knowledge	−0.01	0	0.01	−0.01
Product Knowledge	−0.01	−0.01	−0.01	−0.01
Info. Item Knowledge	0.01	0	0	0
Info. Item Knowledge	−0.01	−0.01	−0.01	−0.01
Info. Item Knowledge – Don’t Knows	−0.01	0.02	−0.01	−0.03
Info. Item Knowledge – Don’t Knows	−0.01	−0.01	−0.01	−0.03
Other covariates	Yes	Yes	Yes	Yes
N	3,588	3,576	3,660	1,260
Pseudo R²	20.91%	32.55%	13.74%	22.92%

Standard errors are in parentheses and clustered at the participant level.

***

p < 0.001, **p < 0.01, *p < 0.05.

We find that the marginal effects of Annual Income and Product Rating are quite robust between treatment groups. If productj promises US$100 per year more than producti, the likelihood of productj to be chosen increases by 1–3 percentage points. A difference in Product Rating of one leads to an increase in retirement income product choice by 3–6 percentage points. The marginal effects of Product Rating in the Treasury-Table and Treasury-Graph treatment groups are similarly high, as are the effects in the Text-Only and No-Rating treatment groups. A higher Access to Capital of US$100,000 means an increase in likelihood to be chosen of 1–2 percentage points in the Treasury-Table and Text-Only groups (p < 0.001). Perceived understanding is positively correlated with the product being chosen in all treatment groups except the Treasury-Table group. The higher a product’s perceived risk, the less its likelihood to be chosen (though this is not significant for the No-Rating treatment group). Perceived control has an overall positive correlation with the product being chosen. Product-specific knowledge, Fact Sheet information item knowledge and the number of Don’t Knows for information item knowledge do not affect product choice. The full model, reported in Tables E1 to E4 in the Online Appendix, shows that almost all the control variables, including trust in superannuation funds and trust in the government, are not significantly associated with the choice of retirement product.

5. Discussion and conclusion

We designed and implemented an online survey with an embedded choice task to assess the impact of variations of product Fact Sheets proposed by the Australian Treasury on retirement income product knowledge, perceptions and choices. Our key findings can be summarised as follows:

First, there is reasonable understanding of four of the five information items on the Fact Sheet, with no significant difference in understanding of the information items by Fact Sheet version. Access to Capital is least understood across all Fact Sheet versions.

Second, the alternative Fact Sheet versions are associated with different degrees of knowledge of the retirement income products. Overall, the annuity products are most understood and the ABP, which is most prevalent outside the experiment, is least understood. Participants in the Treasury-Table treatment know less about all products, while participants in the Text-Only treatment group know more. This raises questions about the benefit of providing tables, graphs and the Product Rating graphic if an aim is to enhance product knowledge.

Third, the alternative Fact Sheet versions impact participants’ perceptions of the retirement income products differently. The perceived riskiness of the products generally follows the Product Rating (or the dot point summary in the case of the No-Rating treatment). However, participants in the Treasury-Graph treatment are more likely (relative to the Treasury-Table treatment) to (correctly) perceive that the Annuity and hybrid annuity products are less risky, while Text-Only treatment is more likely to be associated with (incorrect) perceptions that such products are risky. Moreover, participants in the Text-Only and No-Rating treatments (correctly) perceive that they have more control with the ABP and hybrid ABP products. This suggests that the Treasury-Table and Treasury-Graph Fact Sheets, where the Product Rating is more salient, better communicate the risk management attributes of annuity-type products. Since ‘perceived control’ is important to Australian retirees, regulators should pre-test retirement product Fact Sheets to ensure that they do not facilitate incorrect perceptions of control. Care should be taken in selecting a default retirement product (if that were policy or superannuation fund practice) as some retirees may stick with unsuitable products as a result of their misperception of degree of control offered.

Fourth, we find that financial competence is significantly associated with perceptions of the retirement income products. People with good knowledge of the products and developed numeracy skills are more likely to (correctly) perceive the annuity and hybrid annuity products to be risk management products rather than risky products. However, those who are confused about the Fact Sheet information items (proxied by answering ‘do not know’ to the information item knowledge questions) are less likely to consider the ABP and hybrid ABP products to be risky.

Fifth, we find that when participants are presented with a retirement income product Fact Sheet to assist with product selection, the annuity and hybrid annuity products are most preferred. In aggregate, the Annuity/ABP (a hybrid lifetime annuity/phased withdrawal product) is chosen 34% of the time, closely followed by the Annuity (an indexed lifetime annuity) which is most preferred in one third of decisions. The ABP (an account-based pension) is preferred only 13% of the time and is least preferred in all treatment groups except the No-Rating treatment, which excludes the income security Product Rating. This is contrary to real-world experience in Australia where the ABP accounts for around three quarters of all income streams, compared with only 6% for lifetime annuities and no or negligible take-up of the hybrid annuity/ABP products which are not widely available. Overall, preference for the Annuity and hybrid Annuity products is higher in Fact Sheet versions where the income security Product Rating is more salient (particularly the Treasury-Table and Treasury-Graph treatments), the information on Access to Capital and Death Benefits is more detailed and the Potential Income Shape is included on the Fact Sheet.

Sixth, in terms of product choices, we find a positive, statistically significant association between the likelihood of choosing a retirement income product and the information items (income security) Product Rating and Average Annual Income for all Fact Sheet versions. Potential Income Shape is also relevant for the Treasury-Graph version of the Fact Sheet. Not surprisingly, the role of the income security Product Rating is stronger in the Treasury-Graph and Treasury-Table versions, where the Product Rating is a prominent, salient feature. However, it still has a significant (but smaller) effect in the Text-Only treatment where the Product Rating is far less visually prominent and in the No-Rating treatment group which excludes the numeric Product Rating but includes the underlying risk information.

Finally, perceptions about the retirement income products are also important for product selection. Perceived control has a positive, statistically significant association with the likelihood of choosing a retirement income product across all Fact Sheet versions; perceived understanding is positively associated with choice of a retirement income product in most versions of the Fact Sheet, while perceived risk is negatively correlated with choice of a retirement income product for all except the No-Rating version which excludes the numeric Product Rating. Our findings suggest that perceptions of the retirement income products are more important for stated preferences than actual understanding of the Fact Sheet information items or product knowledge. This is in line with studies on subjective financial literacy, such as Fernandes et al. (2014) and Anderson et al. (2017). However, the relationships could be more subtle as suggested in Bateman et al. (2018) who conclude that financial competence in terms of financial literacy, numeracy and product knowledge is key to formation of the (correct) perception that annuities are risk management products and are associated with less control than ABPs and their hybrid varieties.

In terms of an overall evaluation of the proposed Fact Sheet, our conclusions are as follows. First, we find that with the exception of Access to Capital, the information items on the proposed Fact Sheet are quite well understood, across all versions of the Fact Sheet. Second, when using the Fact Sheet to choose retirement income products, two factors drive choice: (1) the information items Average Annual Income and the income security Product Rating and (2) perceptions of risk and control associated with the products. Perceptions that annuity products provide income security are associated with the Treasury-Table and Treasury-Graph Fact Sheets where the Product Rating is more salient, while perceptions that the ABP products allow control are associated with the Text-Only and No-Rating Fact Sheets where the Product Rating is less salient or omitted. This finding, in conjunction with poor understanding of Access to Capital, suggests that future versions of the Fact Sheet could provide a clearer explanation of the liquidity features of ABP products. Third, when using the Fact Sheets to select retirement income products, the survey participants strongly prefer annuity products across all versions of the Fact Sheet. This effect is greater for Fact Sheet versions where the income security Product Rating is salient and where participants ‘correctly’ perceive annuities and their hybrid varieties as risk management products (the Treasury-Table and Treasury-Graph which include the Product Rating graphic). This outcome should not be surprising since the Fact Sheets tested present the products in a consumption frame, which has been shown in previous literature to strongly encourage preference for annuities over non-annuitized retirement income products (Bockweg et al., 2018; Brown et al., 2008, 2013).

Finally, one might wonder why we find so much demand for annuity products when real-world demand is weak in Australia and globally. We advance several reasons. First, once participants agree to take part in the experiment, they are ‘forced’ to think about all four retirement income products and their key features. We financially incentivised participant’s understanding of the information items included in the Fact Sheets and the key characteristics of the four retirement income products. In the real world, decision-making is influenced by financial advisers, regulations and decisions, and advice of family and friends who are unlikely to advise annuity products due to a lack of experience or awareness. Furthermore, the two hybrid retirement income products included in the choice task have only recently been endorsed by Australian regulators, and a standard annuity is little known. It is not unusual to find these effects in online stated choice experiments (Bateman et al., 2018; Beshears et al., 2014). Second, in hypothetical decision-making, we assume away frictions in product distribution. In the Australian setting, regulations and the absence of hedging possibilities have led to limited supply of annuity-type products that insure longevity risk. Third, it is well known that a consumption frame increases the demand for annuities (Bockweg et al., 2018; Brown et al., 2008, 2013). In this context, the information items on the retirement income product Fact Sheets highlight the ability of a product to cover spending needs in retirement – through the information items Average Annual Income and Potential Income Shape, and in the presentation of risks in terms of income security for the Product Rating – but provide little emphasis of the liquidity feature of ABP products.

Overall, our findings will inform information format decisions by regulators in those countries currently considering the introduction of disclosure formats for retirement income benefits. In the Australian context, we contribute to the policy discussion of the shape of multiple avenues of guidance to be provided to retirees by superannuation funds under the Retirement Income Covenant to commence in 2022.

Supplemental Material

sj-pdf-1-aum-10.1177_03128962221132720 – Supplemental material for How Fact Sheets affect retirement income product knowledge, perceptions and choices

Supplemental material, sj-pdf-1-aum-10.1177_03128962221132720 for How Fact Sheets affect retirement income product knowledge, perceptions and choices by Hazel Bateman and Inka Eberhardt in Australian Journal of Management

Footnotes

Acknowledgements

The authors thank Karen Cong from the Institute for Choice for excellent survey programming. They also thank Rebecca McCallum, Penny McCormack, Nicole Mitchell, Erin Murray, Jordan George, Darren Kennedy, Matthew Laing and Chris Timotheou from Treasury for help in designing the versions of the Fact Sheets for testing and for feedback, and participants of the Netspar International Pension Workshop 2020 in Leiden, specifically Hans Hoeken, participants of the 28th Colloquium on Pensions and Retirement Research, Australian Securities and Investments Commission (ASIC) and Behavioural Economics Team of the Australian Government (BETA) for valuable feedback.

Final transcript accepted 5 September 2022 by Chelsea Liu (DE Finance).

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: This research was supported by the Australian Research Council Centre of Excellence in Population Ageing Research (project CE170100005).

ORCID iD

Hazel Bateman

Supplemental material

Supplemental material for this article is available online.

Notes

References

Agnew

Anderson

Gerlach

, et al. (2008) Who chooses annuities? An experimental investigation of the role of gender, framing, and defaults. American Economic Review 98: 418–422.

Agnew

Szykman

Utkus

, et al. (2012) Trust, plan knowledge and 401(k) savings behavior. Journal of Pension Economics & Finance 11: 1–20.

Anderson

Baker

Robinson

(2017) Precautionary savings, retirement planning and misperceptions of financial literacy. Journal of Financial Economics 126: 383–398.

Australian Government Actuary (AGA) (2018) Retirement Income Risk Measure. Parkes, ACT, Australia: AGA.

Australian Prudential Regulation Authority (APRA) (2021) Annual Superannuation Bulletin June 2021 (Issued 29 January 2021). Sydney, NSW, Australia. Available at: https://www.apra.gov.au/annual-superannuation-bulletin

Australian Securities and Investments Commission (ASIC) and the Dutch Authority for Financial Markets (AFM) (2019) REP 632 Disclosure: Why It Shouldn’t Be the Default? Available at: https://asic.gov.au/regulatory-resources/find-a-document/reports/rep-632-disclosure-why-it-shouldn-t-be-the-default/#:~:text=Financial%20services%20disclosure%20has%20traditionally,context%20in%20which%20disclosure%20operates

Bateman

Piggott

(2011) Too much risk to insure? The Australian (non-) market for annuities. In: Mitchell

Piggott

Takayama

(eds) Securing Lifelong Retirement Income: Global Annuity Markets and Policy. New York: Oxford University Press, pp. 81–105.

Bateman

Dobrescu

Newell

, et al. (2016) As easy as pie: How retirement savers use prescribed investment disclosures. Journal of Economic Behavior & Organization 121: 60–76.

Bateman

Eckert

Iskhakov

, et al. (2017) Default and naïve diversification heuristics in annuity choice. Australian Journal of Management 42: 32–57.

10.

Bateman

Eckert

Iskhakov

, et al. (2018) Individual capability and effort in retirement benefit choice. Journal of Risk and Insurance 85: 483–512.

11.

Benartzi

Previtero

Thaler

(2012) Annuitization puzzles. Journal of Economic Perspectives 25: 143–164.

12.

Beshears

Choi

Laibson

, et al. (2011) How does simplified disclosure affect individuals’ mutual fund choices? In: Wise

(ed.) Explorations in the Economics of Aging. Chicago, IL: The University of Chicago Press, pp. 75–96.

13.

Beshears

Choi

Laibson

, et al. (2014) What makes annuitization more appealing? Journal of Public Economics 116: 2–16.

14.

Bird

Foster

Raftery

, et al. (2018) Who starts a self-managed superannuation fund and why? Australian Journal of Management 43: 373–403.

15.

Bockweg

Ponds

Steenbeek

, et al. (2018) Framing and the annuitization decision: Experimental evidence from a Dutch pension fund. Journal of Pension Economics & Finance 17: 385–417.

16.

Boyer

Box-Couillard

Michaud

P-C

(2020) Demand for annuities: Price sensitivity, risk perceptions, and knowledge. Journal of Economic Behavior & Organization 180: 883–902.

17.

Brandt

Day

Deutscher

(1975) Information disclosure and consumer credit knowledge: A longitudinal analysis. Journal of Consumer Affairs 9: 15–32.

18.

Brown

Kapteyn

Luttmer

, et al. (2017) Cognitive constraints on valuing annuities. Journal of the European Economic Association 15: 429–462.

19.

Brown

Kapteyn

Luttmer

, et al. (2021) Behavioral impediments to valuing annuities: Complexity and choice bracketing. Review of Economics and Statistics 103: 533–546.

20.

Brown

Kling

Mullainathan

, et al. (2008) Why don’t people insure late-life consumption? A framing explanation of the under-annuitization puzzle. American Economic Review 98: 304–309.

21.

Brown

Kling

Mullainathan

, et al. (2013) Framing lifetime income. Journal of Retirement 1: 27–37.

22.

Chomik

Graham

Yan

, et al. (2018) Retirement income in Australia: Part III – Private resources. CEPAR research brief, November. Available at: https://cepar.edu.au/sites/default/files/retirement-income-in-australia-part3.pdf

23.

Commonwealth of Australia (2014) Financial System Inquiry – Final Report. Available at: https://treasury.gov.au/sites/default/files/2019-03/p2014-FSI-01Final-Report.pdf (accessed 8 December 2021).

24.

Commonwealth of Australia (2017) Minister for revenue and financial services, explanatory statement. Treasury laws amendment (2017 measures no. 1) regulations. Available at: http://classic.austlii.edu.au/au/legis/cth/num_reg_es/tla2017mn1r2017201700704516.html.

25.

Day

Brandt

(1974) Consumer research and the evaluation of information disclosure requirements: The case of truth in lending. Journal of Consumer Research 1: 21–32.

26.

Deetlefs

Bateman

Dobrescu

, et al. (2019) Engagement with retirement savings: It is a matter of trust. Journal of Consumer Affairs 53: 917–945.

27.

Department of Social Services (2019) Means Test Rules for Lifetime Income Streams. Available at: https://www.dss.gov.au/seniors-budget-measures/means-test-rules-for-lifetime-income-streams (accessed 9 June 2020).

28.

Diacon

Hasseldine

(2007) Framing effects and risk perception: The effect of prior performance presentation format on investment fund choice. Journal of Economic Psychology 28: 31–52.

29.

Dohmen

Falk

Huffman

, et al. (2011) Individual risk attitudes: Measurement, determinants, and behavioral consequences. Journal of the European Economic Association 9: 522–550.

30.

European Insurance and Occupational Pensions Authority (EIOPA) (2018) Implementation of IORP II: Report on the pension benefit statement: Guidance and principles based on current practices. EIOPA-BoS-18/328, 13 November. Frankfurt: EIOPA.

31.

Fernandes

Lynch

Jr Netemeyer

(2014) Financial literacy, financial education, and downstream financial behaviors. Management Science 60: 1861–1883.

32.

Galesic

Garcia-Retamero

(2011) Graph literacy: A cross-cultural comparison. Medical Decision Making 31: 444–457.

33.

Gosling

Rentfrow

Swann

(2003) A very brief measure of the Big-Five personality domains. Journal of Research in Personality 37: 504–528.

34.

Hiscox

Hobman

Daffey

, et al. (2017) Supporting Retirees in Retirement Income Planning. Barton, ACT, Australia: Department of the Prime Minister and Cabinet, Commonwealth of Australia.

35.

Howlett

Burton

Bates

, et al. (2009) Coming to a restaurant near you? Potential consumer responses to nutrition information disclosure on menus. Journal of Consumer Research 36: 494–503.

36.

Ikonen

Sotgiu

Aydinli

, et al. (2020) Consumer effects of front-of-package nutrition labeling: An interdisciplinary meta-analysis. Journal of the Academy of Marketing Science 48: 360–383.

37.

Iskhakov

Thorp

Bateman

(2015) Optimal annuity purchases for Australian retirees. Economic Record 91: 139–154.

38.

Janger

Block-Lieb

(2006) The myth of the rational borrower: Behaviorism, rationality, and the misguided reform of bankruptcy law. Texas Law Review 84: 1481–1565.

39.

Koh

BSK

Mitchell

Fong

(2021) Trust and retirement preparedness: Evidence from Singapore. Journal of the Economics of Ageing 18: 100283.

40.

Kozup

Howlett

Pagano

(2008) The effects of summary information on consumer perceptions of mutual fund characteristics. Journal of Consumer Affairs 42: 37–59.

41.

Lambregts

Schut

(2020) Displaced, disliked and misunderstood: A systematic review of the reasons for low uptake of long-term care insurance and life annuities. Journal of the Economics of Ageing 17: 100236.

42.

Lipkus

Samsa

Rimer

(2001) General performance on a numeracy scale among highly educated samples. Medical Decision Making 21: 37–44.

43.

Loewenstein

O’Donoghue

(2006) ‘We can do this the easy way or the hard way’: Negative emotions, self-regulation, and the law. University of Chicago Law Review 73: 183–206.

44.

Lusardi

Mitchell

(2011) Financial literacy around the world: An overview. Journal of Pension Economics & Finance 10: 497–508.

45.

Markard

Holt

(2003) Disclosure of electricity products – Lessons from consumer research as guidance for energy policy. Energy Policy 31: 1459–1474.

46.

Navarro-Martinez

Salisbury

Lemon

, et al. (2011) Minimum required payment and supplemental information disclosure effects on consumer debt repayment decisions. Journal of Marketing Research 48: S60–S77.

47.

Office of the Superintendent of Financial Institutions (OSFI) (2016) Disclosure Requirements for Defined Contribution Pension Plans. Available at: https://www.osfi-bsif.gc.ca/Eng/Docs/disclosure.pdf (accessed 8 December 2021).

48.

Organisation for Economic Co-operation and Development (OECD) (2021) Pension Funds in Figures. Available at: https://www.oecd.org/daf/fin/private-pensions/Pension-Funds-in-Figures-2021.pdf (accessed 9 June 2021).

49.

Salisbury

(2014) Minimum payment warnings and information disclosure effects on consumer debt repayment decisions. Journal of Public Policy & Marketing 33: 49–64.

50.

The Treasury (2016) Development of the Framework for Comprehensive Income Products for Retirement (Discussion paper, December). Available at: https://consult.treasury.gov.au/development-of-the-framework-for-comprehensive-income-products-for-retirement

51.

The Treasury (2018a) Retirement Income Covenant Paper: Stage One of the Retirement Income Framework. Commonwealth of Australia, May. Available at: https://treasury.gov.au/sites/default/files/2019-03/c2018-t285219-position-paper-1.pdf

52.

The Treasury (2018b) Retirement Income Disclosure Consultation Paper: Stage Two of the Retirement Income Framework. Commonwealth of Australia, December. Available at: https://treasury.gov.au/sites/default/files/2019-03/181207-FINAL-Retirement-Income-Disclosure-Consultation-Paper.pdf

53.

The Treasury (2021) Retirement Income Covenant. Final position paper, Commonwealth of Australia, 19 July. Available at: https://treasury.gov.au/consultation/c2021-188347

54.

Thorp

Bateman

Dobrescu

, et al. (2020) Flicking the switch: Simplifying disclosure to improve retirement plan choices. Journal of Banking & Finance 121: 105955.

55.

US Securities and Exchange Commission (SEC) (2018) Updated Disclosure Requirements and Summary Prospectus for Variable Annuity and Variable Life Insurance Contracts. Available at: https://www.sec.gov/rules/proposed/2018/33-10569.pdf (accessed 14 January 2020).

56.

Vischer

Dohmen

Falk

, et al. (2013) Validating an ultra-short survey measure of patience. Economics Letters 120: 142–145.

57.

Walther

(2015) Key investor documents and their consequences on investor behavior. Journal of Business Economics 85: 129–156.

58.

Warren

(2021) Design of comprehensive income products for retirement using utility functions. Australian Journal of Management 47: 105–134.

59.

White

Sussman

Beckett

(2019) When shrouded prices seem transparent: A preference for costly complexity. In: Bagchi

Block

Lee

(eds) Advances in Consumer Research, vol. 47. Duluth, MN: Association for Consumer Research (ACR), pp. 185–189.

60.

Wiener

Holtje

Winter

, et al. (2007) Consumer credit card use: The roles of creditor disclosure and anticipated emotion. Journal of Experimental Psychology: Applied 13: 32–46.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

4.80 MB