Introduction
This paper examines economic issues associated with aged care insurance. The background to this assessment is the recent Final Inquiry Report of the Productivity Commission on Caring for Older Australians (2011a) setting out an approach to the future funding of aged care that, among other elements, involves a larger direct financial contribution by care recipients. This policy debate in Australia fits into an international landscape of increasing awareness regarding the materiality of aged and long-term care (Costa-Font, Karlsson and van den Berg Reference Costa-Font, Karlsson and van den Berg2012).
The key concern of this paper is that, given costs of care that are uncertain and likely to vary greatly within the population and in time, increasing recipient co-contributions impose significant risks on potential users of aged care even with a large coverage by the government scheme.Footnote 1 We discuss whether these risks are insurable, and how such an insurance scheme might compare in terms of achieving policy objectives for alternative options, including those proposed by the Commission. The applicability of the insurance option is supported by the fact that a dual disability–injury insurance model was preferred by the Productivity Commission as a foundation of an equitable and sustainable system capable of responding to the care and support needs of people with disability (Productivity Commission 2011b).
Welfare gains can be attributed to insurance under some preconditions in both demand and supply. Demand benefits from insurance because individuals are assumed to be risk averse and prefer a smaller but certain loss (the premium) over the possibility of an uncertain but substantial loss. Therefore, individuals transfer some wealth from the ‘good’ to the ‘bad’ state of the world (low to high marginal utility). Such transfers of wealth are the prime economic function of insurance (Nyman Reference Nyman2003). However, for individuals to be aware of these potential gains, they need to take into account the expected loss and the timing of purchase.
The supply benefits from insurance through underwriting contracts with many individuals. We assume that insurers are indifferent to risk or less risk averse than individuals and know the probabilities associated with the ‘good’ and ‘bad’ outcomes, those probabilities being independent of insurance. Therefore, there are no problems of asymmetric information such as adverse selection or moral hazard. We also assume low transactions costs in devising and implementing the contract so that its terms hold and each part fulfils their obligations. Finally, we consider risks as independent, so that the probability of everyone experiencing either the ‘good’ or the ‘bad’ state at the same time is very low, given a sufficiently large population.
Policy objectives are those of economic efficiency (the best use of society's resources), equity (universal and affordable access to care at acceptable guaranteed quality levels), stability of finance (financial security of older people with limited ability to respond to adverse changes in aged care funding) and fiscal sustainability (balance of public finances). There is, in particular, a risk that pay-as-you-go (PAYG) approaches will prevent sufficient savings from occurring now to allow future burdens to be met, leading to an unjustifiably large sacrifice of consumption, eroding the living standard of future taxpayers and lowering their commitment to the intergenerational contract. Therefore, availability of insurance in this setting provides a significant welfare gain – or equivalently, the lack of insurance imposes a substantial welfare loss relative to the maximum value the society can obtain.
Keeping these objectives in mind, the second and third sections explain, respectively, whether demand and supply prerequisites of welfare gains are met and the aged care insurance market is commercially viable. The fourth section compares the mandatory insurance option with its alternatives in terms of policy objectives. Finally, the concluding remarks posit that while all policy options involve trade-offs, a transition to mandatory insurance presents an attractive venue.
Demand-side preconditions for aged insurance
The extent of loss
The magnitude of the risk of loss is an important component to be aware of in understanding the role of aged insurance. The losses associated with long-term care have a probabilistic character: while the likelihood of ever requiring long-term care is relatively high, the duration of care, and hence its cost, varies greatly within the population in ways that are amenable to statistical characterisation. Moreover, given individual costs of care as high as Aus $50,000 and expected to increase in time, the severity of this eventuality should be high enough to make risk sharing between risk-averse individuals attractive.
Aged care costs are mainly a function of longevity and morbidity (having both a cohort and an individual component) and the price of aged care services. Uncertainty is attached both to the average level of costs for a cohort and within that cohort to each individual's longevity and morbidity (Cutler Reference Cutler1996). This inter-temporal risk makes insurers become possibly risk averse and they neutralise the risk by providing lump sum indemnities instead of service benefits in their contracts, transferring risk to individuals (Brown and Finkelstein Reference Brown and Finkelstein2007).
The Productivity Commission provides projections for age- and sex-specific mortality rates from 2007 to 2050 under low, medium and high survival scenarios, all declining and implying higher life expectancies at birth (ranging from 88.2 for men and 90.8 years for women in the medium scenario to 93.8 for men and 95.8 for women in the high scenario). However, the distribution around the modal age at death remains wide in Australia and other low mortality countries. Thus, any given cohort life expectancy – say, 85 years – implies that some 55–56 per cent of the cohort will live longer and so premature mortality is offset. Research by the United Kingdom (UK) Pensions Institute shows that 25 per cent of current 65-year-old British men will live to the age of 90, although the actual figure could lie anywhere between 16 and 33 per cent (Dowd, Blake and Cairns Reference Dowd, Blake and Cairns2010). Consequently, a significant individual longevity risk will prevail in the future.
Regarding morbidity, the prevalence of chronic disease and related health problems is increasing in the older population, but not only because of ageing, as it increases in successive cohorts of 65–69 (Jagger et al. Reference Jagger, Matthews, Matthews, Spiers, Nickson, Paykel, Huppert and Brayne2007). However, this trend might be partially explained by improved detection and earlier engagement with medical care, and extended survival could expose individuals to an increasing risk of disability and age-related functional loss, which in some instances will prove lengthy and severe.
Hence, as the Productivity Commission suggests, significantly greater shares of the costs are borne by care recipients as a result of: rising cohort life expectancy but with high individual variation; an increase in the prevalence of chronic conditions with widespread individual incidence, severity and duration; and rising costs of aged care services relative to other goods and services.
Risk aversion
The degree of risk aversion determines the demand of aged insurance in order to avoid the exposure to the future financial losses. At the same time, morbidity life scenarios affect the potential welfare gains of age insurance. Consumers taking into account state-dependent utility might rationally choose not to purchase insurance because they place a low value on consumption while receiving care (Brown and Finkelstein Reference Brown and Finkelstein2009). Dementia provides an extreme example of a reduced marginal utility of income. However, most people in contemplating the possibility of dementia would likely value the assurance of a reasonable comfort of living for themselves or to avoid a lower living standard of their partner, should they have one. Steadily rising joint life expectancy (Paolucci and Ergas Reference Paolucci and Ergas2011; Swan Reference Swan2010) creates a type of in vivo bequest motive that reinforces the conventional welfare gains expected from aged insurance. Moreover, the risk of suffering other cost-causative aspects of senescence (some degree of disablement or impairment) increases costs without limiting the ability to enjoy most of the activities of life.
The timing of purchase
On the demand side, welfare gains depend on whether individuals perceive the value of insurance sufficiently early in life as to be able to secure its purchase. Aged insurance is unattractive for those close to retirement because premium prices are high in order to accumulate a claim over a relatively short period of time, and to counterbalance negative selection. Rather, long-term care insurance is most affordable for young individuals, who may nevertheless consider it unattractive as they under-estimate the value of income claims that may take place in a distant future. Hence, individuals may act with some sort of state-dependent utility, with erratic rationality, or myopically, delaying the purchase of insurance until it is no longer affordable or the insurable event has already occurred. In such cases, the potential welfare gains will not materialise.
Different reasons explain the low participation in aged insurance: limited consumer knowledge or rationality makes individuals ‘mistakenly’ give up insurance (Brown and Finkelstein Reference Brown and Finkelstein2009), with this myopia stemming from subjectively underestimated probabilities of old-age dependency; ignorance and difficulties in measuring beliefs and processing information; trust in family solidarity; excessive price; and lack of liquidity in assets such as housing (Finkelstein and McGarry Reference Finkelstein and McGarry2006; Pestieau and Ponthiere Reference Pestieau, Ponthiere, Costa-Font and Courbage2011). The existence of public programmes may also distort individual choices, as the availability of means-tested subsidies for aged care lowers the attractiveness of voluntary aged insurance (The Society of Actuaries 2011). In the latter case, the liability for aged care costs hinges not only on attaining old age and an occurrence of need, but also on failing the means test.
As a result, even though there would be welfare gains from risk sharing, only a small proportion of young adults would be so risk-averse as to purchase voluntary aged insurance. Thus, it is not apparent that a voluntary demand for long-term care insurance would be sufficient to make this market attractive for insurers and induce them to bear the costs of entry.
Supply-side preconditions
Adverse selection
Adverse selection imposes formidable constraints on the nature and functioning of voluntary insurance markets. It occurs when individuals taking out insurance have a better knowledge of their risk exposure than the insurer (asymmetric information). At the same time, those with above-average exposure to risk can be more inclined to insure. If adverse risks do auto-select for insurance, estimates of the average risk exposure of the population will understate the risk to which the insurer is exposed, because part of the information in consumers' possession is ‘hidden’ from the insurer. Hence, adverse selection is said to arise when consumers have ‘hidden attributes’. As a consequence, it may reduce the incentive for insurers to engage in price competition. Brown and Finkelstein (Reference Brown and Finkelstein2007, Reference Brown and Finkelstein2009) find that in the long-term care insurance market, prices are substantially above the break-even contract, with substantial differences in prices by gender that do not translate into coverage.
All told, adverse selection would not be a serious problem in markets for aged care insurance if it had to be purchased early in life, when only a few individuals are likely to know more about their life expectancy than what is statistically observed by insurers. At late life stages, individuals are more likely to have private information on their risks of aged care need and adverse selection can be a serious obstacle. Therefore, there is a relationship between the age at which cover is obtained, the risk of adverse selection and the inherent affordability of insurance.
Evidence of adverse selection in the aged insurance market has been found in the United States of America (USA) through correlation between the subjective probability of entering a nursing home and the probability of purchasing long-term care insurance (Sloan and Norton Reference Sloan and Norton1997) and France, through correlation between high risk of dependency and purchasing long-term care insurance (Courbage and Roudault Reference Courbage and Roudault2008). Interestingly, heterogeneity in preferences plays a role that is difficult to be accounted for, as there is also evidence of correlation between wealth and expectancy of life, or having life insurance and living longer (Cutler, Finkelstein and McGarry Reference Cutler, Finkelstein and McGarry2008; Finkelstein and McGarry Reference Finkelstein and McGarry2006). This finding is important because, besides the typical effect of adverse selection, some low-risk individuals with a high level of risk aversion will also buy insurance, compensating to some extent the negative effect.
Moral hazard on the insured side
Moral hazard refers to any change in behaviour in consequence of becoming insured. Ex ante moral hazard refers to the tendency of insured individuals not to invest sufficiently in precautionary measures, because the costs of an incident will not fall entirely on them. Ex post moral hazard refers to the individual's mitigation efforts after the insured event occurs. Therefore, it corresponds to the consumer undertaking actions that are ‘hidden’ from the insurer and implies that estimates of the inherent risk of the population will understate the risk to which the insurer is actually exposed.
Markets for medical insurance provide abundant evidence of moral hazard, both ex ante and ex post. However, aged insurance does not allow individuals to directly transfer income between states of the world where its marginal valuations are different; instead, it provides the right to access services at no cost or reduced cost, access being controlled by health professionals and based on need. At the same time, most (and the most costly) aged care services have little value for those without need – no healthy person would choose to reside in a high-care facility. For that reason, there is a common agreement in the literature that moral hazard is not a major concern in this market (Pestieau and Ponthiere Reference Pestieau, Ponthiere, Costa-Font and Courbage2011). However, the problem would raise a concern for specific aged care services that are consumed also by healthy individuals, such as assistance with household chores and transportation, provision of which has a substantial discretionary component. Evidence of this type of moral hazard in community care is provided by Grabowski (Reference Grabowski2006) for the USA, Karlsson (Reference Karlsson2004) for Europe, as well as Bell and Bowes (Reference Bell and Bowes2006) for Scotland. Consequently, these services may be subject to deductibles or co-payments without otherwise efficient decisions being distorted.
Although the appearance of moral hazard depends on the use of discretionary services also consumed by healthy individuals, insurers may find it attractive to adopt them for the purposes of mitigating adverse selection (hidden actions providing signals for hidden attributes).
Supply-side moral hazard
Supply-side moral hazard corresponds to the way providers are compensated for their services. When the compensation exceeds the cost of an additional unit of care, providers have an incentive to over-supply. Thus, there will be inefficiency in the provision of aged care services, for example, in prescribing unnecessary procedures or assistance. This problem occurs while providers are not exposed to financial responsibility for their choices. However, as the insurance literature shows, transferring responsibility to insurers and managed care conveys the incentives for risk selection and under-provision, which in the long run might be more inefficient than the initial problem. Consequences for overall costs have to take into account not only health care, but also availability of informal, often spousal care.
The complexities of insurance are even more pronounced considering the nature of aged care. For example, covering residential care perpetuates the increasing trend in rising costs of this mode of formal care, while disincentivising the efficiency of domiciliary care. In addition, covering different forms of care makes the insurer more exposed to adverse selection and moral hazard.
Insurance also removes or reduces the incentive for consumers to actively seek better providers, which may accentuate the problem of supply-side moral hazard. Consequently, encouraging active patient choice, for instance by requiring co-payments, may induce more intensive provider competition. However, consumer choice in voluntarily private demand of long-term care insurance is constrained by difficulties in assimilating their information on prices and coverage which may create space for overpricing in the market (Brown and Finkelstein Reference Brown and Finkelstein2007). In this regard, controlling supplier moral hazard poses challenges that are not readily resolved.
Transaction costs
The market of insurance rests on the assumptions of reasonably low transactions costs in devising, implementing and enforcing the contract, so that the terms of the insurance contract hold and each party fulfils their obligations. This assumption of minor transaction costs may seem innocuous, but it is not. For example, if the structure of payoffs differs from those matching the states of the world, the parties must not only incur the administrative costs of the contract but also bear basis risk, which is the risk that the value of a hedge may not move in perfect opposition to the position it hedges.
It is also a requirement that the insurer be able to tell at tolerable cost whether it is the ‘good’ or the ‘bad’ state of the world that prevails. However, as mentioned above, deciding upon a person's need for a particular form of care may be contentious and it is an example of costly state verification. Following Grignon and Bernier (Reference Grignon and Bernier2012), we assume that transaction and verification costs are sufficiently low as not to distort the accuracy of decisions.
Correlated risk
In order to achieve welfare gains and thus to provide incentives for offering insurance contracts, it is necessary that individual risks are independent, so that the probability of everyone experiencing either the ‘good’ or the ‘bad’ state at the same time is very low, given a sufficiently large population (uncorrelated risk). If risks were correlated within the population, expanding coverage would not contribute to reducing them and would limit the scope for intra- and intergenerational risk pooling, imposing a need for higher reserves on the insurers through increased prices or lower quantity or value for money of the insurance package. This is related to the typical assumption of insurers' preferences being risk neutral, while they might be risk averse in the long-term care market given the special circumstances and uncertainties. However, risk associated with long-term care may not be independent. For example, health technology assessment increases longevity not for one but on average for all members of current and succeeding cohorts.
The correlated longevity risk could be dealt with through commercial products such as life insurance policies or longevity bonds, transferring longevity risk on to commercial investors, and facilitating the development of the commercial insurance market. These bonds might be issued by the government as recently recommended (Henry Reference Henry2010). In this way, the insurance market would produce efficiency gains with insured individual risks, while the systematic (i.e. undiversifiable) risk component would be shared among the population, attaining the optimum identified in the Borch (Reference Borch1975) theorem.
Comparing policy instruments
Aged insurance has the potential to yield welfare gains, but they are unlikely to be achieved unless the insurance is made mandatory at young ages. It is therefore relevant to first recognise the existing solution, by analysing the key points and challenges of the Productivity Commission (2011a) proposal, and then to discuss our alternative of mandatory aged care insurance.
The Productivity Commission proposal
The Productivity Commission report responds to the challenge of the increasing need of the ageing population and the concurrently decreasing availability of informal care-givers and skilled labour force. The proposal's main goal is to provide comprehensive information for individuals given their needs and financial capacity, and to improve their choice of insurance contracts (providers, home care, additional coverage, quality in accommodation) within a universal co-ordinated health-care system.
Given the concerns of sustainability, the proposal relies on PAYG funding with an increased co-payment from care recipients, capped with a stop-loss mechanism, with means-testing support provided by the Commonwealth on home equity release products such as reverse mortgages. The PAYG public system to subsidise aged care costs is not necessarily inefficient. It redistributes income from the working to the retired population. At first sight, it might raise concerns on intergenerational equity grounds, but Grignon and Bernier (Reference Grignon and Bernier2012) show how benefits from formal aged care reach not only dependents but also their children through the lower need of informal care. Besides, if productivity rises over time, ‘pre-saving’ replaces present with future consumption when the marginal utility of income is lower. Another advantage of PAYG is that it avoids the problematic implementation of an individually funded system, which involves a transitional generation of users who need to both pre-save and respect the phasing out intergenerational contract. However, PAYG produces greater incentives for the moral hazard of claiming benefits without having contributed.
To solve the issue of moral hazard, in line with the Productivity Commission report, Baldini and Beltrametti (Reference Baldini and Beltrametti2006) propose to impose a 30 per cent co-payment rate over the first 20 years. However, more extensive means-testing of aged care assistance is effectively a tax on asset accumulation during working life, reducing incentives for lifetime savings and leading to strategic impoverishment in older individuals with a low recovery rate (Cremer, Pestieau and Ponthiere Reference Cremer, Pestieau and Ponthiere2012).
Therefore, stop-loss or reverse mortgage tools are proposed by the Commission to limit the risk borne by recipients. The stop-loss mechanism would avoid inefficiencies in provision derived, for example, from delays in obtaining the right form of care because of financial problems. However, it only refers to the cost of care but not to other services, such as accommodation, which remain uncapped. Thus, consumers in need and especially those with a spouse still carry the risk. Setting the stop-loss level is also complicated, as it involves a trade-off between risk on potential recipients and public expenditure.
By stimulating the reverse mortgage market, the Commission lowers the costs of rapid asset de-accumulation among older people. However, the mortgagee is exposed to adverse selection by mortgagors who have a high expectation of prolonged tenancy in homes, who are in homes that are particularly vulnerable to price depreciation, or both. At the same time, the owner of a portfolio of reverse mortgages would be exposed to correlated risks such as fluctuations in aggregate house prices. Different forms of reverse mortgage have been implemented in France and the UK, albeit with limited impact so far (Grignon and Bernier Reference Grignon and Bernier2012).
The Commonwealth might be the least-cost insurer for some risks inherent to reverse mortgages. However, unless insurance is mandatory, risks associated with adverse selection or moral hazard would remain as in a private insurance scheme, and it is not clear that this approach would be more efficient – the Commission's proposal imposes unnecessary risk on potential aged care users. This approach would admittedly reduce the short- to mid-term fiscal pressures on the Commonwealth thanks to increased contributions from the ‘baby boom’ generation. However, asset de-accumulation among older people would reduce expected bequests to younger cohorts that currently bear most of the tax burden. Therefore, the Commission's solution places substantial risk on the current and subsequent generations without providing an instrument for more efficient risk sharing. Although the application of the stop-loss mechanism alleviates this problem, it is limited and is sustained through a lower reduction of the fiscal burden than would occur under a long-term insurance scheme.
The above features raise the question of how well the Commission's scheme ranks in terms of sustainability, as consumers could find themselves lacking any clearly defined and legally binding entitlement to aged care services.
Mandatory aged care insurance
Different reports have considered the need of implementing an aged care insurance market. For example, Grignon and Bernier (Reference Grignon and Bernier2012) recommend the implementation of a universal public insurance plan with full coverage based on a standard evaluation of care needs, while Blomqvist and Busby (Reference Blomqvist and Busby2012) consider that the universal programme would be too expensive and their proposal is based on subsidies targeted at those who need them with a safety net guaranteeing minimum coverage. For Australia, we propose to implement mandatory aged care insurance with private (and competitive) provision.
Insurance could and ideally should be provided on a competitive basis, with insurers competing to provide cover within a framework set by public regulation. They would become the primary buyers of care and obtain payments from consumers and the government, the latter compensating insurers for providing cover to high-risk and low-income individuals, in a universal and mandatory set-up. Efficiency incentives would stem from the choice of consumers between competing insurers and the choice of insurers contracting with competing providers. However, the discussed preconditions for welfare gains from aged insurance would pose a serious regulatory and policy challenge.
The system performance rests on ensuring effective competition at the levels of insurance and care supply; defining instruments for assuring universal affordability of access, determining a minimum level of cover; devising tools for consumer empowerment and information; effective market regulation with prudential obligations; and structuring financial flows to encourage efficient transactions (Paolucci Reference Paolucci2010). Although the implementation of our proposal involves substantial transition costs, it accounts for potential benefits from competition such as gains from risk-pooling, incentives for efficiency and innovation, or the capacity to support and advance equity of access.
There would be efficiency gains from intra- and intergenerational risk sharing and from having informed, risk-bearing purchasers structuring the market. Also, this system would achieve equity and universality of access through targeted income transfers, and would guarantee stable supply, with consumers having legally enforceable claims to cover on the contracted terms. Lastly, the new system would reduce the burden on taxpayers of providing the targeted affordability subsidies, as premiums would largely prefund the required outlays.
Its implementation would need a transition period first because of the complex market design, and second because it would take a number of years for the funds accumulated in the system to fully cover outlays, which would either delay the implementation or require the scheme to be supported with public subsidies. Therefore, there would be a transitional burden on the cohorts that would both pay the premiums and generate tax revenue for securing aged care for older, uninsured cohorts. Relying on the ‘baby boom’ generation might be insufficient to bear this burden.
Although not free of problems, the mandatory aged care insurance model is still workable and desirable, and its financial burden might not prove excessive compared to the Commission's proposal, particularly if some spreading of the transition costs over successive generations could be achieved.
Concluding remarks
The trend in aged care expenses and demographics constitutes a challenge for individuals, as it increases the risk they carry, and for governments, which aim at increasing the equity, efficiency and sustainability of their social systems. In Australia, the Productivity Commission recommends reforms involving increased co-contributions, which suppose the risk of losing a significant portion of individual lifetime savings. We propose an alternative model of mandatory insurance, free from the problem of excessive individual risk exposure, and potentially representing the first-best option for the financing of aged care.
We have reviewed the different preconditions for welfare gains in the aged care insurance market and our results indicate that adverse selection is not a problem for this market if the insurance is purchased at relatively young ages. Moral hazard is likely to occur on the consumer side only for those services that are universally valued and discretionally provided, but co-payments and deductibles would prevent inefficient expansion in demand, and hold transaction costs by reducing the processing costs of low-value claims.
Insurers may have an incentive to postpone entry into high and expensive forms of aged care, and increase access to some community care or home care options to save costs and mitigate adverse selection. Also, they may present incentives for over-servicing when reimbursement exceeds marginal costs. However, selective contracting by insurers of competing providers yields to cost-efficient provision avoiding risk selection. At the same time, insurer competition for informed customers provides incentives to invest in service quality.
Regulation needs to assure effective selective contracting and transparency to help rational individuals. Young individuals would in all likelihood not face enough incentive to rationally buy voluntary aged insurance while other sources of public provision are available, even if they depend on means-tested subsidies. Hence, we propose a mandatory scheme for relatively young individuals, allowing for the consumer choice of insurer and provider, complemented with public subsidies to guarantee universal access.
Constituting aged care insurance on a commercial basis requires overcoming certain market failure-related obstacles. However, economic theory and evidence reveal that those obstacles can be tackled, subject to the right policy efforts. The commercial insurance market for aged care-related financial risks seems a viable option and a welfare-improving alternative to the reform recently proposed by the Productivity Commission. Such an option has been adopted in Japan and the USA where voluntary private health insurance represents about 5–7 per cent of total long-term care expenditures. In terms of population coverage, the USA and France are two of the leading markets, with about 5 and 15 per cent of the population aged 40 and over holding a long-term insurance, respectively. In other countries, e.g. the Netherlands and Switzerland, over the last decade the government has been gradually moving long-term care from national non-competitive insurance towards universal mandatory competitive social insurance (Organisation for Economic Co-operation and Development Reference Colombo, Llena-Nozal, Mercier and Tjadens2011). The Australian government should weigh the pros and cons of each model with great care, as their decision will affect the financial and social security of Australians for generations to come.
