Solving the Retirement Income Covenant

Having led the world in the 1990s in embracing defined-contribution (DC) retirement plans, Australia now - through its new retirement income covenant - is redesigning the system to meet the needs of people living in and approaching retirement.

Given the now widespread acceptance that the goal of your superannuation system should not be maximising a pot of wealth at retirement but generating a sustainable inflation-protected income for life, the question has shifted to how this is done.

Australia’s progress on this challenge is close to my heart. Over more than a decade, in my role as Resident Scientist with Dimensional Fund Advisors and as an academic who has studied lifecycle finance for half a century, I made numerous trips to your country to help lead the debate on retirement income reform.

Your Treasury-designed covenant will require super fund trustees from 1 July 2022 to document a retirement strategy, one that strikes a balance between maximising expected retirement income, managing risks to the sustainability and stability of that income, and ensuring members retain flexible access to their funds.

While this task is proving a challenge for some, Australia is not alone in struggling with how to fund retirement. Significant numbers of middle-class workers in many countries, including my own, are now living longer and confronting rising inflation. What’s more, many face retirement without the reassurance of an adequate pension.

Below, I propose a framework that greatly improves the engineering of the products to deliver retirement income, while leaving the complexity ‘under the hood’ and providing meaningful choices to people planning and entering retirement.

But first, fund trustees in Australia should start by reminding themselves that the primary concern of their members remains what it always has been: ‘Will I have sufficient income in retirement to live comfortably?’

This is broadly in line with my own definition of a good retirement goal: “An inflation-protected income for life that allows you to sustain the standard of living you enjoyed in the latter part of your working life.”

In some ways, the “problem” we are trying to solve is no problem at all, but a blessing. People in developed economies like Australia are living longer and the time they spend in retirement is increasing. According to your official statistician, life expectancy at birth is now 81.2 years for males and 85.3 years for females.¹

We also see these patterns around demographics and longevity in Europe, Japan and North America. We are all living longer. That’s a good thing! The big issue, as you have been discussing here, is how are you going to pay for all that extra time.

My view is that insofar as addressing the retirement shortfall is a problem, it is one of engineering, not science. We already have the tools to fix this issue. We know how to make the system sustainable and increase people’s chances of a good retirement.

Essentially, this new system comes down to making smarter products rather than trying to make consumers smarter about finance in order to use those products.

No country has got retirement income perfect. Like the Netherlands and Singapore, Australia is rated highly, with your three pillars of the age pension, the compulsory superannuation guarantee and voluntary private savings.

But most of your super funds have been built around a lump-sum goal. Under DC systems like Australia’s, individuals carry the risk for both funding their retirement and making the complex decisions required about generating the income they will need for the rest of their lives.

These decisions range from how much to save, how to invest, and how to manage the pay-out phase. To add to these challenges, there is a large and growing class of workers in the informal sector or “gig economy” who do not fall under the superannuation guarantee and who do not earn regular retirement savings. Few of these people have access to financial advice.

Canberra is well aware that having a large segment of the population retire with limited resources does not bode well for the future fiscal position, as the government will have to support the many inadequately funded retirees.

This dilemma is what your retirement income covenant, after more than a decade of enquiries, is designed to resolve.

My view is the ideal solution would allow investors to invest toward retirement income over time and seek to protect those investments from inflation and market risks in low-cost, diversified strategies that offer meaningful engagement.

But if an individual’s superannuation savings are invested to maximise capital value at time of retirement and her personal goal is to achieve a reasonable level of retirement income, there is a clear mismatch involved.

In the view of the superannuation fund, the relevant risk is portfolio value. But the relevant risk for the individual is uncertainty around retirement income.

The answer to this quandary is to adopt a liability-driven investment strategy that is equivalent to how an insurer hedges an annuity contract or how pension funds hedge their liabilities for future retirement payments to members. In this case, however, the liability is the retirement standard of living they owe to themselves.

Essentially, the task is to grow members’ savings during accumulation and manage income uncertainty, using long-term, duration-matched, inflation-linked bonds as the risk-free component, leading up to and during their decumulation phase.

The portfolio allocation between risk-free and risky assets would be dynamically adjusted based on each member’s age, projected future contributions and “funded ratio”. This latter term measures the member’s assets divided by their liabilities – assets being how much income the member’s lump sum buys and liabilities being their target income goal. The funded ratio is key as it provides in a single number an immediate measure of how close members are to their retirement goal.

Some international perspective might be helpful here. In the US, the 2019 SECURE Act requires that DC plans display the amount of retirement income able to be purchased, using the current account balance based on market annuity prices. This can be added to the old-age income benefit to easily calculate the funded ratio and does so without members having to understand compound interest or other technical concepts. This way, members get a clearer view of the consumption currently available in retirement relative to their goal.

In contrast, a traditional fund that focuses on maximising account balances within risk constraints will use short-term fixed income as a hedge. But if the goal is to convert a portfolio into a stream of income, holding short-term bonds actually increases risk and has low expected returns.

Under the new generation solution, each member would still get a pot of money at retirement and would still have the same choices over their savings that they have now. The difference is the value of the pot would be obtained through a strategy meant to maximise the likelihood of achieving the desired income stream.

Moving to this income-focused strategy would require changes not only to the way super funds invest their members’ money but also to how they engage and communicate with savers.

Instead of being asked complex questions about asset allocation, engaged members would be asked three simple questions – their retirement income goal, how much they can contribute from current income and how long they plan to work. Of course, the asset allocation is important, but this is only a factor for achieving success. It is not a meaningful input for the choices the consumer makes.

Once these variables are known, the fund need only regularly communicate to the member the funded ratio for progress to the nominated goal and the probability of reaching it. To increase that probability, the fund member has only three choices – save more, work longer or take more risk.

The aim here is to make DC retirement plans as easy as the old defined-benefit (DB) plans, with a comprehensive default offering that requires no decisions by members at all. Answers to the three questions for default members would be inferred from their age, salary level, and the default contribution rate.

Members who do become engaged should have access to a decision tool that offers only meaningful choices based on what they already know and without their having to become financially literate. The solution provides a smooth transition from accumulation to payout phase, customised at retirement.

Alongside saving more, working longer or taking more risk, members could be offered a fourth choice. This option, particularly significant for a country like Australia where most people’s wealth is tied up in housing, is to generate more benefits from their existing assets using reverse mortgages or downsizing the family home.

As for the ‘tail risk’ of living far longer than the expectations at retirement, there is the option of adding a deferred annuity which only pays after age 85, although in Australia your means-tested old age pension would likely satisfy that need.

Another suggested component for the retirement solution in Australia, and one that is being looked at by other nations such as Brazil, is one that I put to your Treasury back in 2018 when it first sought feedback on improving retirement income security.

Treasury could issue a new low-cost, liquid and safe ultra-long bond instrument which I and my research partner Arun Muralidhar have dubbed SeLFIES (‘Standard-of-living indexed, Forward-starting, Income-Only Securities’).

Unlike a typical government bond, which pays semi-annual coupons and a principal payment at maturity, SeLFIES make no payments until a set date in the future (typically the expected retirement date) and then make level periodic payments for a fixed number of years equal to the average life expectancy at retirement. The payment stream mirrors the payout pattern from contributing to a pension plan.

SeLFIES payments are indexed to aggregate per capita consumption, which hedges both inflation and standard-of-living growth from time of purchase and vastly simplifies the investment decision process. Individual savers can use current standard of living as their retirement income goal without having to forecast either future inflation or standard of living growth. In short, SeLFIES ensure constant income, access to capital and longevity risk-management.

This single, liquid, low-cost, low-risk instrument is easy to understand, even for the least financially confident individual. This is because they embed accumulation, decumulation and compounding with standard-of-living and inflation adjustments, and are expressed in terms of their annual income payment in retirement just like the age-pension and annuity benefit. SeLFIES are well-designed for individuals with no formal pension plan to save for retirement and for those using personal saving to add to their plan benefits.  Pension plans and insurance companies can also use SelFIES to hedge more effectively their annuity and payout liabilities.

These bonds are good for governments as well, by improving balance sheet management with stabilising, reliably-supplied domestic long-term funding, while reducing the risk of individuals retiring poor.

Even without the supplement of SeLFIES, there are still four broad options in the drawdown phase that trustees can use as components for members to mix and match blends depending on their needs and circumstances.

These start at a guaranteed income for life, delivered through a traditional annuity, the age pension or the now rare DB plan. This may be the optimal solution for someone who has no bequest motive or urgent liquidity need.

The second option is a conservative drawdown plan, built around government bonds. This provides secure income and deals with the bequest and liquidity needs, but the member sacrifices the annuity’s longevity protection and its extra income from the ‘mortality dividend’.

The third option, as set out above, targets an income goal for retirement backed by an asset allocation to manage retirement income risk using robust, scalable and low-cost investment strategies. The allocation should make efficient use of all dedicated retirement assets, including housing.

The fourth option, and one that can supplement the others, is longevity insurance using a deferred life annuity purchased at retirement and with payments starting at age 85. In practical terms, your means-tested age pension could satisfy this longevity protection need.

Meeting the retirement income framework of maximising retirement income, managing relevant risks like inflation and market risk and preserving flexibility to access savings will require a combination of all of these tools.

From my experience and lifetime of studying funding retirement, a better DC solution starts with the goal of generating sustainable retirement income. If that is our goal, we need an appropriate risk metric, which is volatility of that future income.

Our strategy, then, is managing that income risk, using an integrated approach which matches the retirement-goal liability with all sources of income, including future contributions and housing.

As no member is average, we are going to need to customise the solution across a range of metrics. If the member is falling short, we need to give them meaningful information and actionable choices.

Most of all, we want to build optionality into the system by giving them flexibility in their drawdown choices.

This is all doable within your retirement income framework and I will be greatly interested to see how the industry progresses in meeting this challenge in the coming years.