Designing Default Retirement Plans

A Framework for Super Funds

By Geoff Warren, Laura Ryan, Craig Roodt, Kirsten Wymer & Edwin Lung

The following article featured in Investment Innovation Institute [i3], 17 March 2021.

Academics, practitioners, consultants and fund managers have come together to develop a new framework for designing default retirement plans. In this article, they give an insight into how to implement such a framework and what the implications for members are.

Disclaimer: The views in this document are those of the authors and do not represent the views of their employers.  This document is provided as general information only and does not consider any entity’s specific objectives, situation or needs.

Fiduciaries presiding over default pension plans must determine an “appropriate” investment strategy for their members. While considerable effort is expended on the technical aspects of developing and modelling such strategies, it is critical to first make assumptions about the members including such as their needs or wants, investment objectives, risk profile and other attributes required to specify the strategy.

This article summarises a framework we have proposed for parameterising these assumptions, which has been recently been published in the Journal of Retirement.

Characterising Diverse and Often Disengaged Member Bases

Decisions about how members are characterised and modelled have significant implications for the investment strategies that are embedded within pension plans. Since the purpose of pension plans is to provide funds to be used in retirement, any mismatch between actual needs of members and perceived needs as acted on by plan providers will clearly impact on member welfare.

Minimising the mismatch is made difficult by:

• Incomplete information about members
• Diverse nature of the membership

Plan providers must not only address the usual difficulties of decision making under uncertainty about investment returns, but also uncertainty regarding the members and their characteristics.

For default pension plans, the problems are further heightened as frequently not even the most basic information is furnished by the member. This is analogous to the difficulty faced by an insurer when attempting to build an investment strategy to satisfy liabilities where key parameters that influence the liability are absent. The criticality is similar to the need to define the system or operational objective before even specifying the technical requirements of a new computer system.

Asking the Right Questions

This issue links with the debate about the purpose of the retirement industry. In Australia for instance, there have been ongoing arguments about the purpose of superannuation (produce income in retirement, or something broader), the desired outcome (e.g. the ASFA retirement standards,[1] or a replacement ratio of working income), and impacts on different groups of members by income, balance, homeownership status, and so on.

The challenge faced by both plan providers and those entrusted with setting public policy can be summed up as asking the right questions. The issue is whether those questions can be answered. For example, how can a plan provider measure income or homeownership status if the member fails to provide that information?

The Proposed Framework

Plan providers need a structure to resolve the key assumptions that underpin product design. Our framework offers such a structure. It consists of four elements, supported by guidance for what type of issues or items need to be addressed under each:

1. Needs versus wants – Whether the strategy is directed toward what the plan provider assumes that default members need, or aims to understand and cater for member self-perceived wants. For instance, the provider may view the member as needing a retirement solution, whereas some members may be focused on their balance and could be risk-averse over short-term losses.
2. Investment objectives – An investment objective or set of objectives must be specified. The three main choices are: (a) providing outcomes in retirement, in particular income, (b) delivering a lump sum at retirement, and (c) maximizing the increase in balance while avoiding drawdowns in the interim.
3. The member – The member for which the default is being designed needs to be defined. This involves establishing the attributes by which members are characterised; and the extent to which member heterogeneity is addressed via segmenting the member base to deliver tailored defaults. Key attributes are balance, income, age and assets outside of the pension plan including any home.
4. Risk appetite – Risk appetite encompasses both risk aversion and risk capacity. It might be stipulated by referring to theory, existing literature, or empirical analysis such as surveying members or observing their actions.

The Elements Are Interrelated

The four elements are interrelated. The needs[2] versus wants element faces the difficulty of defining needs and wants when the membership is not homogeneous, and available information is limited.

This in turn feeds into deciding what objectives to adopt. e.g. whether to create an income stream and/or deliver bequest in retirement, produce a lump sum at retirement for members to do with as they please, or maximise returns while limiting near-term risk.

This links through into what needs to be known about members in order to characterise them, including their risk appetite. Specifying risk aversion and risk capacity faces the difficulty of understanding the members’ true risk appetite and setting the risk measures to be used, which in turn circle back to objectives.

Each of these choices have clear implications for the nature of the investment strategy and its potential outcomes.

Member Segmentation and Assignment

Whether to offer multiple default options is a key decision. The more segments, the higher the likelihood of “matching” member needs, but the greater the complexity.[3] There is also the challenge of assigning members to the segment that most closely resembles their characteristics.

The question arises as to what characteristics are required. In addition to balance, income, age and other assets including any home, other attributes might include health, household status, risk appetite, contribution levels, or perhaps more.

Some of this information could be harvested from the static data provided on commencement with the plan and some from the employer.[4] Additional data could be sourced by surveying members or other information gathering exercises. But all are not without their limitations, which we explore further in the paper.

Implications for Member Outcomes

Our paper provides basic examples to highlight the potential implications of these decisions.

For instance, consider the choice to address longevity risk by maximising the expected length of time that the funds last by taking more investment risk, rather than minimising volatility to reduce the risk of short-term losses[5] through a conservative strategy[6]. This choice alone, taken at 55, leads to a 25% higher expected retirement balance in exchange for the higher volatility. Furthermore, the likelihood of funds lasting until age 90 doubles.

For 55-year old members concerned with balance at retirement, the choice between a growth fund, a target date fund and a conservative fund depends on risk aversion. Assigning a highly risk-averse member to the growth option, or a risk tolerant member to the conservative option, leads to modelled welfare losses of approximately 10%-15% of the balance at retirement.

Plan Provider Choices Do Matter

The decisions made by providers of default pension plans matter. The choice to cater for member needs or wants, the investment objectives that are pursued, how the membership is characterised, and assumptions about risk appetite are each of importance. While it can be difficult to obtain the desired level of insight into default members, the plan provider must explicit address these for elements as the choices will ultimately impact on the outcomes experience by plan members.

This article was co-written by Edwin Lung, Head of Quantitative Analytics, Investments at BT Financial Group; Craig Roodt, Director, Investment and Wealth Advisory at Deloitte Touche Tohmatsu; Laura Ryan, Head of Research at Ardea Investment Management; Geoff Warren, Associate Professor in the College of Business and Economics and Kirsten Wymer, Head of Risk Strategies and Research at BT Financial Group. To access the full paper, please visit the Journal of Retirement.

__________

[i3] Insights is the official educational bulletin of the Investment Innovation Institute [i3]. It covers major trends and innovations in institutional investing, providing independent and thought-provoking content about pension funds, insurance companies and sovereign wealth funds across the globe.

[1] ASFA provides both Modest and Comfortable income targets, and suggests that a balance of $640,000 is needed at retirement for a couple to enjoy a comfortable lifestyle.

[2] In Australia the sole purpose test, which states that funds exist for the purpose of “providing income in retirement”, could be an example.

[3] To return to the computer analogy, the greater the specification for individual enterprise requirements away from the ‘off the shelf’ the greater the likelihood of meeting the enterprise’s needs, but at the cost of greater system complexity.

[4] May depend on applicable employment and privacy laws

[5] The initial approach on commencement of NEST in the UK and KiwiSaver in New Zealand were invest conservatively to prevent people capitulating if there were negative returns in the early years.

[6] This is comparing a 70-30 strategy with a 30-70 strategy.