Withdrawal Capacity in the Face of Expected and Unexpected Health and Aged-care Expenses During Retirement


31 December 2014

We examine the impact on retirement income levels, income stability and longevity risk of accounting for costs associated with age-related health treatment and aged-care services during the retirement phase. To measure the impact of these costs on income sustainability and longevity, we derive asset return data using historical bootstrap simulation to determine an optimal withdrawal income during retirement using dynamic optimisation techniques. We show that the greatest risk to income sustainability occurs when unexpected health costs translate into greater longevity, particularly for conservative investors. Paradoxically, this means that high costs associated with health treatment may result in a longer life, however, without commensurate adjustment in asset allocation towards assets with a greater risk-return profile it also risks premature wealth depletion. We further show that the optimal withdrawal rate is highly sensitive to the timing of health costs and moderately sensitive to later-life aged care costs. We find that for a broad set of circumstances the risk of premature ruin can be mitigated through a dynamic lifecycle strategy during the retirement phase.

It is well documented that the profile of the population is ageing within most developed countries. Ageing naturally affects individuals across a number of domains including physical and mental health, housing, income security, opportunities for social and economic participation, including labour force participation and wealth management priorities.

The decline in defined benefit (DB) pension plans and the transition to defined contribution (DC) plans is gradually expanding liability management from specifically funding retirement income to the financing of a wider range of social responsibilities including age-related health treatment, appropriate housing and aged-care management services and facilities.

Typical individual lifecycles comprise a period of employment followed by a period of retirement. Increasingly it falls to individuals to reallocate consumption from their working life to retirement if they wish to enjoy financial security and avoid poverty in old age. DC pension plans can achieve this reallocation in a way that is consistent with the preferences of the individual plan member.

Arguably, investors don’t tend to think of retirement consumption as a liability at all. Investors may count their superannuation portfolio as an asset but often forget to count the liabilities for which the asset is held.

Explicit consideration of potential costs incurred towards the end of one’s life tends to be an unsavoury reality and so remains largely ignored by financial advisers and retirees until the immediate need arises to meet such expenses. DB plan members obviously face the same risks but income stability means they are at least marginally better able to systematically plan for such costs, where this planning takes place.

This study examines the impact on income level, income stability and longevity risk of anticipating the costs associated with age-related health treatment and aged-care services during the retirement phase.

To measure the impact of such costs on income sustainability and longevity, we simulate asset return data using historical bootstrap simulation to derive an optimal withdrawal income during retirement for a range of confidence levels. This allows us to test the sensitivity of income sustainability in relation to the retirement horizon, the magnitude and timing of health and aged-care costs, unexpected longevity and the interplay between risk aversion and asset allocation during retirement.

We derive a series of ruin probability profiles that quantify the impact of both the timing and magnitude of health and aged-care costs on the safe withdrawal rate for a typical retirement portfolio.


The stochastic optimisation/dynamic goal-oriented investment methodology has a number of attractive features:

  • The model is extremely flexible and can accommodate almost any set of assumptions or features relating to existing types of pension arrangements. The model therefore has considerable practical potential.
  • The methodology allows us to develop sensitivity and ‘what if?’ experiments by changing key assumptions and observing how these changes affect our results. These exercises are obviously useful because they identify the key factors affecting results and gauge the response to particular assumptions.
  • The model is naturally extended beyond the accumulation phase (the period up to retirement) to deal with the distribution (or post-retirement) phase. This is a necessary element of retirement modelling that has historically been disaggregated from accumulation phase modelling by retirement planning scholars.

We examine the probability of ruin for a range of investment strategies for investors who face expected and unexpected health and aged-care costs during retirement.

Broadly, investors who anticipate health and aged-care costs suffer a lower probability of ruin over the retirement horizon compared with investors who fail to account for such liabilities. However, investors who fail to anticipate health and aged-care costs may be able to avoid ruin, and indeed outperform a static investment strategy, if they adopt a form of drawdown dynamic lifecycle (DDLC) investment strategy. This naturally requires a higher risk tolerance than they may be able to bear, but it may also be the only way to avoid ruin.