One of the key risks Australians face in retirement is longevity risk; the risk they will outlive their wealth. While many factors contribute to this risk, including uncertain investment returns and unexpected spending requirements, a major component is an individual's unknown life expectancy. In defined benefit (DB) pension plans, longevity risk is managed across the entire plan. Those who die earlier than expected help 'pay' the pensions for those who live longer than expected. In some sense, all a large DB pension plan must do is ensure it has sufficient assets for the 'average' life expectancy of its members. In defined contribution (DC) schemes there is no pooling of risk and individuals are responsible for managing their own longevity risk.
One way for an individual to ensure they will have a retirement income stream throughout retirement is to plan for their 'maximum' life expectancy, for example, they may believe this to be 105). However, the difference between funding retirement for average life expectancy, as in a DB plan, and funding retirement for maximum life expectancy, as in a DC plan, can be significant.
If real interest rates, that is, the rate of interest in excess of the rate of inflation, are expected to be 2.5 per cent a year forever, a male retiree aged 65, who requires $50,000 in the first year of retirement, rising at the rate of inflation until death, will require about $718,000 to fund retirement for their average life expectancy of 18 years until the age of 83, compared with $1.255 million to fund the same retirement income for their maximum life expectancy of 40 years, that is, until age 105. An alternative perspective is that a 65-year-old male retiree who has $718,000 could fund an income in retirement, growing with inflation until age 105, starting at $28,600, rather than $50,000.
Hence, retirees that plan around their maximum life expectancy will limit spending in retirement and reduce their standard of living compared with those that plan for average life expectancy. Life annuities offer the risk-pooling benefits of DB plans through private markets. If these could be purchased at "actuarially fair" prices, that is, $718,000 for a life annuity starting at $50,000 in the above example, the vast majority of retirees would be better off than they are through self-insuring. The benefits of pooling risk are so great that even at prices well below actuarially fair prices, life annuities can offer significant benefits to DC plan retirees.
There are many factors that limit the demand for private market life annuities, including lack of awareness, product fees, the desire to leave bequests, lack of liquidity and moral hazard. Moral hazard refers to the fact buyers of private annuities tend to be people who expect to live longer than average and providers protect themselves against this by raising the price and reducing the promised benefits. In addition, all Australian residents, subject to some qualifications, have access to a basic, taxpayer-funded, means-tested life annuity through the Federal Government's age pension. The single rate of age pension is set to be a minimum of 25 per cent of male total average weekly earnings and is indexed quarterly to the Consumer Price Index.
Despite all these factors, it is surprising how little demand there is for private life annuities globally, given the pooling benefits they offer retirees. A March 2007 working paper by Wei-Yin Hu and Jason S Scott of Financial Engines Inc in Palo Alto, California, titled "Behavioural Obstacles to the Annuity Market", provides a behavioural explanation for this lack of demand. They show how well-documented biases in individual behaviour can affect the annuity purchase decision.
Behavioural finance recognises that people tend to evaluate risky financial outcomes narrowly, relative to some reference point, rather than more broadly in terms of total net worth. This is referred to as mental accounting. As a result, rather than seeing a life annuity as increasing retirement income from $28,600 to $50,000 for life, as in the above example, they focus on the possible gain or loss in the annuity payments compared with the cost of the annuity. What if they were to die after two or three years? They would have paid $718,000 and received back very little. Life annuities viewed in this way are seen as a gamble on life expectancy and can be viewed as increasing risk in retirement rather than reducing it.
Overlaid on the mental accounting is the fact individuals exhibit risk aversion, having a greater dislike for negative outcomes from a financial gamble than the pleasure they get from positive outcomes. This further reduces the perceived attractiveness of life annuities. Finally, individuals tend to overweight low probability events, particularly those low probability events that can be more easily imagined. Since there are many ways for a retiree to imagine their imminent demise, they tend to overweight the probability of dying soon after purchasing the annuity, making life annuities even less attractive.
There is no doubt in my mind the private market for life annuity types of products will grow in Australia in coming decades and will, on average, be enormously beneficial to Australian retirees. While product providers must increase the value, quality and features of such products, there is a significant role for financial advisers in overcoming behavioural biases of the type outlined in Hu and Scott's paper and ensuring their clients maintain a broad frame in all their financial decision making.
