1 The Geneva Papers on Risk and Insurance Vol. 26 No. 1 (January 2001) 114±125 Ideas for the Funding of Healthcare in the Context of the Ageing of the Population by V. W. FitzGerald 1. Introduction This paper draws upon a paper entitled ``Refocusing and Reinvigorating Retirement Policy'', which I presented to the Conference of Major Superannuation Funds in Australia in It approaches the topic of health funding reform from a ``retirement policy'' perspective, noting that healthcare needs, like retirement income needs, are heavily concentrated in old age. In Australia, as in many other countries, the ageing of the population implies large future rises in health costs, a large proportion of which is publicly funded out of current taxes. This poses, inter alia, intergenerational equity issues, implying a signi cantly heavier tax burden on the next generation for the support of the present one. How can these rising future costs be met more equitably and ef ciently? Encompassing both income and health needs in old age in a single perspective also gives a fuller view of what makes up security in retirement for individual retirees. Security in retirement has a number of distinct and important elements. For example, the situations of two retirees with the same income will be very different if one is a home-owner and the other pays a market rental. So housing status is an important factor. Another very important factor to individuals is their health status, and how well they are covered for health costs, differences in which can imply very big differences in retirees' exposure to nancial risk, as well as in their quality of life. Those aspects are part of the motivation for exploring the question posed above, but the main focus is on the public policy dimension: how to ensure the nancial viability of the health system. In Australia, providing healthcare for the elderly in the future is now a considerably bigger issue for public policy, and a bigger challenge to future public budgets, than providing taxpayer-funded old age income support, through the age pension. The age pension is noncontributory, paid for out of contemporaneous general taxation, but is subject to a tight income and assets test. The full pension for a single person is about 25 per cent of average earnings. The key reason that the future cost of the age pension no longer poses a major challenge to public budgets is that Australia has a long tradition of funded occupational pensions, called ``superannuation'', in past decades covering about 40 per cent or so of the workforce, but now covering nearly every employee. Beginning in the mid-1980s superannuation was extended to virtually the whole workforce, on a de ned contribution model. Paper delivered at Australian International Health Institute Symposium Melbourne, 11±12 November 1999 on ``Reform, Redesign or Revolution: Health Agendas for the 21st Century''. Executive Director, The Allen Consulting Group Pty Ltd, Melbourne, Victoria, Australia. Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK.
2 IDEAS FOR THE FUNDING OF HEALTHCARE IN THE CONTEXT OF THE AGEING OF THE POPULATION 115 Initially this proceeded via industry-wide wage negotiations, but in the early 1990s it was enforced by legislation. The minimum contribution under the Superannuation Guarantee legislation is increasing by 1 per cent of salaries every two years, until it reaches 9 per cent of salaries in The contribution is paid by employers. Many complying funds exist, but the collection mechanisms are both ubiquitous, extending into virtually every workplace, and ef cient. The total assets in the system at mid-2000 were A$447 billion, growing at almost 17 per cent per annum. This system is expected, through strong contribution ows and the compounding of returns (lightly taxed), to deliver substantial real additional saving, capital resources and investment income for retirement income support in the future. It has already signi cantly reduced expected future dependence on the age pension. The cost to the public budget of the age pension is now projected to rise by less than 1.5 per cent of GDPover the next few decades. Although the problems of future funding of health and retirement income needs have strong parallels, nothing like the superannuation system yet exists in Australia for meeting future health costs. Of course even without the major changes in the age composition of the population that lie ahead, the elderly are disproportionately dependent on the healthcare system and healthcare resources even now (relative to younger people). This suggests that there may be value in exploring the possible extension of the superannuation system to health funding for the medium term as well as the longer term. The ideas canvassed in this paper could also be extended to other forms of aged care (other than healthcare per se). It is important to note that this paper assumes that Australia's present mixed system of health funding will broadly continue (a matter about which there is political, as well as academic, debate). Under this mixed system, roughly 60 per cent of the population rely on the public hospital system for acute care and 40 per cent are privately insured for such care ± the latter tending to use private hospitals primarily, but relying on major public facilities in complex or catastrophic situations. Everyone is covered for routine medical care, generally with a co-payment or ``gap'' to pay for each medical service. The public ``Medicare'' system thus provides quality healthcare to all. It is funded from general taxes, mainly Commonwealth (Federal), plus an earmarked levy on incomes (the ``Medicare levy''). There is signi cant taxpayer support also for private health insurance, primarily through a recently introduced government rebate on premiums. Private health insurance is regulated in Australia to be ``community rated'' rather than risk-rated, although now on a lifetime basis. That is, premiums rise with age of entry and insurers may decline to cover pre-existing health conditions, but explicit discrimination is otherwise not allowed. A consequence of these arrangements is that costs of long-term care, an important component of rising future health costs, are spread across both the insured population and (predominantly) taxpayers. In Australia, as elsewhere, there are of course ongoing debates over the ``grand architecture'' of the health system, but this paper focuses on just one aspect: how to fund the projected large rise in costs driven by ageing. The mechanisms discussed are essentially independent of the precise proportion of the population covered by private health insurance, but the dominant motivation for the paper concerns the public (Medicare) system and, particularly, the viability of public budgetary support for it. Intergenerational equity (the balance of burdens as between contributions and taxes, present and future) is thus an important concern. Ef ciency considerations arise also, and bring into the picture the economics of saving and accumulation of capital for future needs. Although the paper is obviously set in the Australian institutional context, it is hoped that the ideas canvassed are of wider interest.
3 116 FITZGERALD 2. Drivers and future growth of health costs While the present and increasing future demands of the elderly for quality healthcare are a valid starting point for examining the robustness of health funding, it is important to acknowledge at the outset that there are important drivers of healthcare costs other than ageing. In a system in which providers dominate, price signals are almost non-existent and new technologies and treatments are rapidly adopted into community expectations, ageing has not, over the recent past, been a very important factor in increases in health costs in Australia. Those other drivers (including increased utilisation of services per capita) have dominated, so far. Moreover, analyses of the pure effect of ageing (on its own) over future decades seem to suggest that it may be a ``manageable'' factor for future Budgets. For example, a recent study in depth by the Australian Federal Department of Health and Aged Care concluded that Given the size of the Australian population as it is today, if its demographic composition were the same as it is projected to be in 2051 then, very nearly, an extra $17.07 billion in today's dollars would be needed to maintain the same level and quality of the three health services as they exist today. This view of the costs due to ageing is isolated, deliberately so, from the tricky issues that have been side-stepped by the assumptions underlying the analysis presented in this paper, particularly those issues related to possible increases in costs due to a larger number, greater variety and complexity of future procedures. 1 It might be noted, however, that the above estimate of an increase of $17 billion is over 3 per cent of Australia's GDP, while the projected future increase in cost to the Australian Federal Budget of providing the age pension is now put at 1 to 1.5 per cent of GDP by the Australian Government's Retirement Income Modelling Unit. 2 In other words, the pure effect of ageing alone on health costs is more than twice as big an issue as the prospective rise in the public costs of providing retirement income, thanks largely to the extent of private provision through superannuation that is now in place. An important component of this private provision preceded the advent of the Superannuation Guarantee; but the combined effects of voluntary and obligatory superannuation provision have certainly now reduced the budgetary issue of retirement income provision to ``manageable'' proportions. Can that be said to be true of the ageing effect on health costs? In considering the issue of future health costs, it is very important to appreciate that the real issue is not what effect ageing will have on its own, but what effects ageing and the other strong drivers will have in combination and interaction. Work by the former Economic Planning Advisory Council (EPAC) by the Retirement Incomes Modelling (RIM) Unit and by the 1996 National Commission of Audit has highlighted this, but although their respective assessments imply major budget pressures ahead (rather bigger than the $17 billion estimate above), these assessments have received little public discussion so far. There are uncertainties in any such assessments, especially about the extent to which healthcare costs per capita will continue to rise faster than the general CPI, as they have 1 Commonwealth Department of Health and Aged Care, ``The Ageing Australian Population and Future Health costs: 1996±2051'', Canberra, mimeo, Available from the Department's website, accessible via 2 See G. Rothman, ``Projections of Key Aggregates for Australia's Aged'', Retirement Income Modelling Unit Conference Paper 98/2, The Treasury, Canberra, Available from the Treasury website, via above entry portal.
4 IDEAS FOR THE FUNDING OF HEALTHCARE IN THE CONTEXT OF THE AGEING OF THE POPULATION 117 persistently done for some years now, and if so by how much. On the range of RIM analyses put forward by the National Commission of Audit: 3 Healthcare costs (public and private combined) in Australia are likely to rise by 6 to 8 per cent of GDP over the next 30±40 years. In 40 years' time, on these analyses, healthcare costs could reach 17 per cent of Australia's GDP, or about double the present gure, with the aged healthcare component being about 10 per cent of GDP. While substantial healthcare costs are met privately, currently around 70 per cent are met by government. Healthcare costs to the Commonwealth Government alone could rise by 4 per cent of GDP. It must be said that projections of future health costs are considerably more uncertain than those of future age pension costs, depending in particular on assumptions about future growth in relative health costs due to the various drivers mentioned above, such as new treatments, rising community expectations, more intensive use, etc. Nevertheless, plausible estimates for total health costs in the 2030s and 2040s range from 13 per cent of GDP upwards, and they imply an increase in public aged healthcare costs of at least 3 per cent of GDP to the Commonwealth Budget alone. 3. How to provide for future health costs? The looming rise of public costs for healthcare in the future clearly poses issues for intergenerational equity similar to those involved in the retirement income issue. The appropriate response in respect of retirement income is now widely agreed to be a balance between support from the taxpayer of the day, compulsory self-provision through superannuation beforehand, and voluntary saving beforehand. Is a similar response appropriate for healthcare? Leaving aside the intergenerational dimension, should there be some obligatory user contribution now (going beyond the Medicare levy), rather than maintaining a public system which is ``free'' to all at point of use, although very costly to the taxpayer? Given the heavy concentration of healthcare usage in old age ± and hence the link to retirement provision in general ± should such a contribution be separate, or closely linked with superannuation? At a very broad level, the answer is that policy needs to strike a balance between the three main potential sources of funds to meet future health costs for an older Australian community. They are basically the same sources as for retirement income: First, current taxes raised from the taxpayers of the day; Second, out-of-pocket contributions by those receiving healthcare at the time, paid out of their (then) current income; and Third, contributions to the costs of that care funded by advance provision now, i.e. through 3 National Commission of Audit, Report to the Commonwealth Government, Canberra, AGPS, June 1996, especially ch. 6.
5 118 FITZGERALD earmarked saving, accumulating in savings funds (both individual accumulation accounts and risk pools) ahead of the future need. 4 It is an issue for debate whether some such contributions should be mandatory (either the second or third types above, or both). It is important to note one difference vis-aá-vis the issue of ensuring provision for future income. In respect of health costs, at least part of any advance provision through saving should desirably be accumulated in risk pools to deal with unforeseen catastrophic health costs for those covered. Of course there is some risk-sharing inherent in the fact that future budgets will inevitably still meet a substantial part of the costs. In effect, much of the risk of catastrophic health costs for retired people will inevitably, and desirably, be borne by the future Australian taxpayer (collectively). In the absence of funds built up in advance, the sharing of the costs between taxpayers and the retired will be essentially a ``zero sum game''. The retired who are members of health insurance funds will likely receive some cross-subsidy from younger people, but in the absence of a radical reversal of the trend of recent years for younger people not to take up private health insurance, 5 the older members (and other groups prone to high health costs) would mainly insure each other, such risk-sharing still having advantages over self-insurance. If substantial funds were built up in advance, however, the taxpayer share and the share paid for out of retirees' other nances could both be moderated. Should there be mandatory contributions for health costs in retirement? Just as with income provision through superannuation, a case can ± as outlined earlier ± be made for mandatory self-provision ahead of time to meet health costs later on, particularly in retirement. However any consideration of such a policy needs to take a number of factors into account, including what people are already being required to contribute (via their employers) to their superannuation, as well as the modest present contribution to health costs via the Medicare levy. Before forming a view on those matters (and on whether any contributions for healthcare, mandatory or otherwise, should in fact be linked to the superannuation system), it is instructive to look at what other countries are doing, or debating. 4 An extremely important reason for considering advance provision is the sheer power of compound interest. For example, to deliver $1 in real terms in 30 years time (roughly the interval between mid working life and mid retirement) requires only a single contribution of about 8 to 13 cents now, if well invested ± i.e. so as to earn around the historical rate of return on business capital. In Australia this has been roughly 8 to 10 per cent per annum real. It might be noted that years of work (and potentially of accumulation) are still about twice as many as years of retirement for males in Australia, further improving the ``leverage'' held out by this type of arithmetic. 5 ``Life-time community rating'', recently introduced in Australia, was designed to make health insurance more attractive toyounger people. Coupled with the 30 per cent rebate on premiums, it has been successful in bringing about a signi cant, although not radical, reversal of the decline which has occurred in membership, from a high of around 70 per cent of the population two decades ago. In response to the present Australian Government's initiatives, the privately insured proportion of the population has recently risen from a low of around 30 per cent to about 37 per cent in mid That in ux of mainly younger and healthier people will take some pressure off premiums in the future, including for older people.
6 IDEAS FOR THE FUNDING OF HEALTHCARE IN THE CONTEXT OF THE AGEING OF THE POPULATION Other countries' models, particularly that of Singapore Although most countries' retirement income provision systems do not themselves cover health costs ± leaving that to separate public schemes or to private health insurance ± there are some instructive examples and various proposals are being debated in particular countries, e.g. in the United States. There have been calls in the U.S. for ``medisave accounts'' treated for tax purposes like the familiar 401(k) retirement savings accounts, with proponents looking to features such as choice (and use of own funds) as helping to contain future health costs. The most comprehensive model incorporating provision for healthcare costs (not only in old age) into a system which also provides for retirement income is Singapore's Central Provident Fund (CPF) scheme. Singapore is just one of a number of countries which have ``provident fund'' arrangements providing for a range of welfare needs, but Singapore's CPF is among the most comprehensive, funded by 20 per cent contributions from each of the employer and the employee (although the employer component is to be reduced). Retirement income provision is actually not, as some outside comments seem to imagine, the major destination of savings accumulated through the CPF. The largest need met through it by far is housing. Singaporeans in fact have a much higher level of home ownership (around 90 per cent) than do we or, say, Canadians or Americans (all around 70 per cent). Nevertheless retirement income provision is one of the CPF's major purposes. Singapore's government recognized early in the life of the scheme, in 1984, that health costs were very important not only to the security of Singapore's citizens in their retirement, but also to their welfare throughout life. Singapore has chosen to keep provision for most such needs off-budget. It was recognized that to meet unforeseen major health costs, particularly hospital-related costs, savings needed to be built up, both in individuals' accounts and in risk pools, and this was deemed to be best done through integration within the CPF. The CPF contains three elements to provide security against health costs: Medisave: Six per cent of salary rising to 7 per cent at age 35 and 8 per cent at age 45 is paid into a Medisave account until the balance reaches S$16,000, after which the excess can be transferred into the ordinary account providing for housing, retirement, etc. These funds, possibly together with some out-of-pocket contribution, may be used to pay for hospital and attending doctor expenses. Medishield: Medishield provides the risk pool or insurance element, covering Singaporeans against extraordinary hospital costs. It is funded by a system of premiums which is essentially age-phased lifetime community rating in its character. There are some exclusions from the scheme (e.g. normal pregnancies) and a lifetime individual coverage limit of S$70,000. In practice, only about 20 to 25 per cent of hospital costs are funded by it, the bulk coming from Medisave accounts. Medifund: This is the government safety net to support those without enough resources in their Medisave accounts or in other personal savings to meet their health expenses. It is particularly important for lower income people and for elderly people whose resources have dwindled. Singapore also allows out-of-pocket medical expenses in a wider category than the hospitalrelated costs primarily covered by these three elements to be deducted from tax, up to a maximum of 2 per cent of salary. The government also subsidises the health system to a substantial degree from its budget. Singapore does not encourage its citizens to take out thirdparty health insurance that could remove too much of their exposure to out-of-pocket costs.
7 120 FITZGERALD The government does not wish people to see healthcare as ``free''. In this regard it might be noted that when they do require hospital care, Singaporeans do have choices ± as between private room or larger wards and the like ± which carry different costs, to be met from their Medisave balances or out-of-pocket. The Singapore system is instructive in: Effectively addressing the health component of retirement security while avoiding (via mandatory minimum provision) excessive costs to the future taxpayer; Combining individual and pooled savings, plus a safety net; Leaving people with incentives not to overuse costly health resources (they have choices; and must bear some out-of-pockets); and Integrating provision for health costs with retirement income provision. Of course Australian institutions are quite different to Singapore's. Nevertheless, since the basic issues being addressed are the same, Singapore represents an instructive point on a spectrum. Other models on the spectrum like the U.K.'s system, which leaves the great bulk of the load to the future taxpayer, or the American system relying more on the private insurance market, but leaving parts of the community poorly covered, seem less instructive for us. 5. Possible ideas for integration of health provision into the superannuation system in Australia As for retirement income provision, Australia has a mixed system of funding ± about 70 per cent of total health costs being borne by public budgets, mainly through the Medicare system, and the other one-third being borne privately ± including a still substantial private health insurance (PHI) component. The proportion of the population covered by PHI has increased appreciably (to around 40 per cent in mid-2000) in response to the recently introduced 30 per cent rebate and lifetime community rating system. Unlike the age pension, publicly funded healthcare is not means-tested, and it is utilized by people across the income range, as indeed is private health insurance. In fact self-insurance for discretionary hospital procedures and reliance on the major public hospitals for complex surgical procedures and the like is not an uncommon strategy among better-off people. Older people, even on quite modest means, are, for fairly obvious reasons, among the keenest to maintain private health insurance cover if they can afford it. Presumably this is because it brings choice of doctor, there is less likelihood of having to wait for such important procedures as hip replacements, and, with the recent trend to reducing uninsured extra costs to the insured hospital patient, it provides relative security against downside nancial risk from unforeseen health costs. Overall, the situation resembles that in respect of superannuation coverage 15 or 20 years ago, before the push to extend coverage and contribution levels. Looking ahead 30 or 40 years, and considering the projected substantial growth in public healthcare costs, principally due to the ageing of the population, it must be realized that there is no effective way to prevent an unbalanced share of the burdens falling on future young Australian taxpayers unless either: Reliance on the ``free'' public health system shrinks and the private sector share increases, which seems unlikely to happen on any scale as long as the public system remains ``free''; or Some kind of patient contribution is phased in, presumably over an extended period,
8 IDEAS FOR THE FUNDING OF HEALTHCARE IN THE CONTEXT OF THE AGEING OF THE POPULATION 121 within the public system. Such a contribution could be met by people from funds built up via some ``add-on'' to the superannuation system, or separately (as discussed below). There would appropriately be ``safety net'' exemptions and the contribution could be capped at an annual limit. The limit could be broadly income-related (e.g. a basic amount, subject to safety net provisions, 6 plus one or two steps applying to people on higher incomes). Such a contribution scheme will only work without shifting too much of the burden to those, notably the retired, actually using healthcare in the future if there is pre-accumulation of adequate extra funds for the purpose. For reasons already discussed, these savings would comprise both individual accumulation balances and risk pools. Eligible uses of healthcare accounts In that regard, one eligible use of an individual accumulation could be to pay one's patient contributions directly from one's own accumulated ``healthcare account'', or to pay for hospital treatment in the private health system on a self-insurance basis. Alternatively, premiums for private health insurance could be paid from the account. With private health insurance now operating under a form of lifetime community rating, with premiums phased by age of entry, it would be very valuable to many retirees to have the means to pay those premiums through savings built up before retirement. Thus the accounts would serve as a ``front end'' for either the public or the insured route. It would clearly be reasonable for funds built up, and earmarked, to meet hospital and related major health costs to be eligible for the 30 per cent private health insurance (PHI) rebate, assuming it continues under future governments. If used for payment of private health insurance premiums the PHI rebate should clearly also apply, but of course only once. In either case, the rebate would appropriately be accessed only when the funds were actually paid out of the account for one of those purposes. Accumulation under superannuation tax treatment (i.e. a tax of only 15 per cent on fund income) and eligibility for the PHI rebate might even be suf cient incentives to encourage substantial voluntary participation in a health cocontribution scheme. (Here, ``co-contribution'' refers to the concept of employees contributing alongside the amounts that employers must contribute for their superannuation.) More likely, however, mandatory participation would need to be considered, for essentially the same reasons that led to making superannuation contributions compulsory. One important issue to consider is whether the accounts would only be drawn upon after retirement. I do not have a rm view on this, but it would appear appropriate and even desirable to allow pre-retirement access for eligible purposes, especially if participation in the scheme were made mandatory. As a related element of security in retirement, the balances in the accounts could be used by self-funded retirees to ``buy'' the Pensioner Health Bene ts, or PHB card (now more accurately called the ``health concession card''), for an annual payment equal to its average cost to the government for pensioners. The main bene t provided by the card is heavily subsidized access to prescribed medicines. It is believed from attitudinal research that many self-funded retirees would be prepared to pay much more than the card would actually cost the 6 Presumably such safety net provisions would not generally exempt retirees with adequate balances in their healthcare accounts (as discussed in the text following), other than in circumstances of hardship.
9 122 FITZGERALD government in order to gain security against the downside nancial risk posed by unforeseen substantial expenses for medicines. The card should be able to be provided to non-pensioner retirees for $4 to $5 per week ($200 to $250 per year) on a revenue neutral basis. 7 Should healthcare provision be integrated with superannuation? The basic policy case for integration with superannuation is founded on: The central importance of cover for major health costs to individuals' security in retirement; and The need for a universal contribution so as to reach an appropriate sharing of health costs between retirees as a group (funded out of accumulations they have previously built up), and future taxpayers. One practical argument for integration is that the superannuation industry has established systems to collect contributions (related to wage and salary income) from almost every employer in respect of virtually all employees. A further practical argument for a collection and aggregation system distinct from the private health insurance industry is the obvious one ± that, at present at least, only about 40 per cent of Australians are members of private health insurance funds. And as outlined, healthcare accounts residing with superannuation could serve to build up the means (especially ahead of retirement) from which to on-pay premiums to specialized PHI funds, but also to cater for those who do not wish to join such funds. Also, ``excess'' balances could simply add to ordinary superannuation provision. Administrative issues for the superannuation industry are discussed further below. Elements of a model The bones of an approach to adopting these ideas and integrating them with other retirement provision through superannuation funds are as follows: (i) The co-contribution concept ± i.e. the notion, which has previously been proposed, that employees should contribute something for their retirement alongside their employer's superannuation contributions ± could be revived. It could take the form of (say) a 2 or 3 per cent co-contribution phased in and earmarked to a healthcare account within one's superannuation fund. Such a contribution could be phased-in in steps of 1 per cent at twoyear intervals following completion of the phasing-in of the Superannuation Guarantee in There should also be no practical dif culty in incorporating this element into older-style de ned bene t funds; almost all such funds have an accumulation component for members' own contributions. (ii) Subject to consistency with an overall reform of the structure of taxation applying to superannuation, reform which should ideally treat contributions from all sources equally, the healthcare contributions should be: ± treated as contributions in lieu of salary; and ± taxed as for other superannuation amounts. 7 See D. Scho eld, ``Re-examining the distribution of health bene ts in Australia: Who bene ts from the Pharmaceutical Bene ts Scheme?'', NATSEM Discussion Paper No. 36, University of Canberra, October Available from the NATSEM (National Centre for Socio-economic Modelling) website.