In part two of this three-part research brief series, we focus on the public element of retirement income provision in Australia, primarily related to the Age Pension. This research brief discusses pension policy trends, the design of pension access, benefit level, and means testing, as well as poverty and fiscal outcomes. 

Highlights include:

• Australia’s Age Pension is broadly affordable and adequate for most but not all Australians: Unlike the main public pensions in many countries, the Australian Age Pension is flat-rate and means-tested so benefits reduce as means increase. It costs less than 3% of GDP and keeps most out of poverty, but not renters.

• It is an important source of income for most retirees: The typical pensioner is a 74-year-old woman born in Australia and living in coastal New South Wales. She did not receive welfare immediately before claiming the pension (though over a third of pensioners did), owns her home, receives the full pension, doesn’t receive a regular private income stream, and has under $50,000 in assessable assets.

• It is also subject to constant reform: To balance the competing objectives, policymakers can adjust a range of Age Pension policy levers, including: (1) eligibility (e.g., via residency and pension age); (2) benefit (e.g., maximum by household type and its indexation); and (3) means testing (e.g., which private resources are tested, how much of these is disregarded, and the rate of benefit withdrawal as such resources increase).

• One area of policy attention has been eligibility age: The Age Pension eligibility age is associated with the decision to retire and raising it is likely to increase mature age labour force participation rates. Eligibility age increased for women (from 60 to 65) and is now increasing for both sexes (to 67 by 2023). Further increases to 70 have been abandoned probably because of: (1) a low fiscal imperative; (2) unequal impacts across groups; (3) labour market concerns (e.g., discrimination); (4) a lack of popularity; and (5) a schedule of increase that was faster than in most other countries.

• There is occasional debate about the level of maximum pension: The choice of benchmark to which the Age Pension is linked remains subjective and arbitrary and will ultimately be based on political consensus but a link to wages in some form is essential to maintain pension benefits in line with standards of living.

• Means testing is the lynchpin of the Australian approach: It is based on policy about (1) which resources to assess and how to measure them; (2) the permissible threshold of resources beyond which pension is withdrawn; and (3) the taper or rate of reduction. It is a useful policy tool, but its potential remains unexplored.

• Future reforms of means testing could look at equalising the treatment of assets: Attempts have been made to equalise treatment of different assets in the Age Pension means test (e.g., deeming) but some assets, such as the family home are excluded. Including the family home would allow Treasury to claw back much of the $2b or about 5% of the pension budget per year that is paid to households with total assets of over $1m.

• Policy decisions about means testing (and age-based taxes) need to consider behavioural outcomes: Means testing drives the share of the population receiving the pension as well as working and saving behaviour. Research suggests that there are economic gains from a greater level of means testing and may result in higher labour supply of prime-age workers. And aggressive means testing of assets along with a more lenient treatment of annuity-type products is justified to disincentivise tax-financed bequests.

• Current and projected expenditure suggests that the Age Pension costs are sustainable: It cost about $45b or 2.4% of GDP in 2018 (2.7% including Service Pensions), a fraction of what other countries spend on public pensions. Recent projections see spending remaining low.

• Australian old age poverty is low when taking account of housing, but the system fails renters: A quarter of pensioners who rent alone spent on average less than $6 on food per day. An increase in rental assistance payments of 40% would reduce lone renter poverty by almost 20 percentage points, at a cost of $380m.