Can private health insurance fulfil its societal role when it comes to ensuring coverage for the elderly? Hamza Hanbali shares some lessons from Australia and Belgium
In many developed countries, private health insurance (PHI) plays an important role in the healthcare system, relieving basic state-backed universal coverage and providing access to more comprehensive and timely healthcare. However, economic and competitive forces can obstruct PHI accessibility, and at the heart of this is risk classification.
Ignoring the heterogeneity between policyholders can lead to adverse selection spirals. Low-risk individuals who pay higher than their ‘fair’ premiums may choose not to insure themselves, which could threaten a system’s sustainability and affordability.
On the other hand, the economic solution to this – risk classification – may hinder the fulfilment of PHI’s societal role, making health protection unaffordable for vulnerable individuals. This is particularly relevant when age is the classification criterion. As Ian Atkinson wrote in the March 2017 issue of The Actuary, “we discriminate against a group we all hope we’ll one day join”. What are insurers to do?
Australia: community-based rating
Australia adopts a community-based rating system with open enrolment to ensure universal PHI accessibility. Policyholders cannot be denied coverage, and pay the same premium regardless of health status, age and sex. However, this requires a complex set of interconnected rules.
On the demand side, there are incentives to encourage young policyholders to take out PHI, such as loadings payable by those who enrol after age 31. The goal is to dampen the effect of adverse selection generated by community-based rating, which were reported in the 1980s and 1990s.
On the supply side, a risk equalisation scheme is in place so insurers that attract riskier policyholders share the burden with those that attract healthier policyholders. The scheme plays an important role in the Australian system, with estimated transfers of more than 40% of hospital and medical costs between insurers.
It is, however, unclear whether this solution has real merits in terms of sustainability and affordability. Recent data shows early signs of another wave of falling participation from young policyholders, likely due to the upward trend in premiums. Population ageing contributes to this trend, especially given the increase in healthcare utilisation. Another potential factor is risk equalisation itself. Although designed to support the system, it can be a source of moral hazard from providers, and a barrier to innovation and disease prevention.
Belgium: lifelong contract designs
Some countries attempt to address the issue by redefining certain features of health insurance products. In Belgium, PHI contracts (excluding disability covers) are lifelong, and the pricing basis must be fixed at policy issue. Policyholders know, in principle, the amount of all future premiums at policy issue. Contracts are priced based on entry age, which again incentivises individuals to take out cover earlier.
Compared with the Australian system, this is less vulnerable to adverse selection and does not rely on a risk equalisation scheme. However, the lifelong nature of the contracts, combined with fixed pricing bases, has other technical consequences, especially for level premium policies. Insurers need to build reserves and forecast expected future medical costs with great precision. Medical inflation makes this challenging, although not impossible. Medical inflation refers to systematic and unpredictable increases in medical costs and is, by definition, hard to project.
In its basic form, such a system addresses the accessibility, and to some extent affordability, of PHI for the elderly. However, it also generates long-term risk for insurers and may threaten their solvency, which ultimately would also impact policyholders. The Belgian regulator has consequently allowed insurers to update premiums over time using a medical index. In theory, insurers cover the traditional diversifiable insurance risk and policyholders bear the medical inflation risk.
Both insurers and consumer representatives support this system, provided it is standardised at market level. To this end, the system requires two components: first, the construction of a medical index that captures inflation; second, an updating mechanism for future premiums that allows for fair risk-sharing between insurer and policyholder.
The method currently in use for the medical index compares the aggregate claims (for five different age groups) net of payments from the universal cover over two consecutive years. Some refinement may be required to account for product differences in the construction and the application of the index.
For the updating mechanism, although only future premiums are updated, these updates should reflect the effect of medical inflation on both premiums and reserves. In a system where policyholders bear all medical inflation risk, premiums may need to be increased by more than what the medical index suggests.
To illustrate this, consider a simple contract that pays an estimated benefit of 100 at time 10 in exchange for level premiums of 10 in arrears for 10 years. At time 1, the estimated benefit is revised up by 50% to 150, due to inflation. Increasing the remaining premiums by 50% would make total premiums payable 145, which falls short of the updated estimated benefit by 5. Therefore, if the policyholder were to truly bear all medical inflation risk, future premiums would need to be updated by more than 50% – effectively applying an ‘updating factor’.
The size of this factor will depend on the policyholder’s entry age and duration-in-force, even if the medical index itself does not. The dependence on entry age arises because the updating factor depends on the level premium. To understand the dependence on duration, suppose the update occurs after the fifth premium payment, not the first. Increasing the remaining five payments by 50% would bring total premiums to 125, which would now lead to a shortfall of 25. Therefore, the updating factor would need to be higher when the adjustment is performed at time 5 compared to time 1.
This example highlights three key points. First, premium increases might be higher than medical inflation. Second, the updating factor depends on the entry age. Third, policyholders with contracts of longer duration might face higher increases than those with contracts of shorter duration.
Could this mechanism open the door to age-based discrimination? Consumer representatives in Belgium have pointed out that insurers could offer teaser rates to new young entrants and apply high adjustments later. One way to avoid this is to set up age-independent premium adjustments. These can be derived by pooling the reserves of a group of policyholders, for instance by policy inception year or even the entire portfolio. The adjustments would then depend on the pool’s age distribution, and not on the age of individual policyholders.
Ignoring secondary effects such as selection and lapses, these methods are actuarially equivalent for insurers but not for policyholders. Under the age-dependent method, policyholders who started their contracts at younger ages require higher adjustments than those who started later. Conversely, due to pooling, age-independent adjustments would lead to later-starting policyholders paying for the medical inflation of earlier-starting policyholders. Therefore, although pooling could help avoid high premium increases for the elderly who started contracts at younger ages, it may also be unfair to those who started later. In fact, these policyholders would face two penalties: one due to the higher initial level premium, and the other due to pooling.
Reconciling economics and ethics
Australia and Belgium both provide partial answers to PHI accessibility and affordability for the elderly, but in both cases there remain technical limitations – even in addition to those discussed above.
All the incentives and penalties involved in the Australian PHI market mean it has started to resemble a compulsory system, which can have consequences on product quality. The market has witnessed an increase of junk products targeting policyholders who do not want to pay penalties, and as a result, even insured individuals can face large out-of-pocket expenses. In Belgium, despite regulation, some insurers have reportedly applied sharp increases of up to 9% over the medical index. The transferability of reserves is another issue. Policyholders who want to switch providers lose their reserves, unlike in Germany, and are therefore tied to their initial insurer.
These challenges show that reconciling the economic and ethical constraints of risk classification requires a sophisticated and well-planned system. Sound scientific comparisons between different systems certainly are timely.
Hamza Hanbali is a lecturer in actuarial science at Monash University, Australia. He previously worked in risk management at AXA Belgium, and on the design of updating mechanisms for private health insurance contracts
Image Credit | iStock