4.1 Coverage

Coverage in South Africa is divided between a large tax-funded public health system and a regulated market for private health insurance (Table 1).

Eligibility for access to the public sector exists on two levels. For ambulatory care, services are universally free to all residents. For in-hospital care, aside from maternity services which are universally free, eligibility for subsidised care is based on a means test. Higher-income groups and those with medical scheme cover are therefore required to pay for public hospital services. The relevant tariffs are based on those charged to medical schemes less value-added tax. Any person falling outside the means test without medical scheme coverage would therefore face a severe barrier to accessing public hospital services.

The population covered by medical schemes largely includes families with breadwinners earning more than the tax threshold who take out cover either as an individual or through an employer (Ministerial Task Team on Social Health Insurance, 2005). To encourage membership, or at least to mitigate disincentives, a tax credit in the hands of the principal (paying) member is available in respect of each beneficiary covered. The value of this tax credit does not vary by income and is set lower than the implicit per capita subsidy available to people eligible for free public health services (Department of Social Development & Wits School of Governance, 2021).

The private system covers around 15 per cent of the national population with expenditure equivalent to 3.9 per cent of GDP (Table 1) with out-of-pocket expenditure less than 1 per cent of GDP (not included in Table 1) or 7.1 per cent of total health expenditure (van den Heever, Reference van den Heever2016).

Both sectors offer a similar package of services, from primary care to all the hospital-based sub-specialties. The medical schemes system therefore offers ‘substitutive’ coverage to the public health system.

While the scope of coverage is superficially similar between the two sectors, there are important differences in the staff and service to population ratios, with far better ratios in the private sector (National Department of Health, 2011a).

4.2 Regulation of the private health system

The private health system is supervised by a specialised regulator, the CMS in terms of the Medical Schemes Act No.131 of 1998 (Republic of South Africa, 1998b).

Both the council members and the chief executive of the CMS are political appointments, which renders it vulnerable to conflicts of interest that operate through the governing party (van den Heever, Reference van den Heever, Dhai, Ballot and Vellor2021). Over time, the strategic role of the CMS is argued to have diminished for this reason (Settas, Reference Settas2022).

Private health facilities are technically regulated by provincial health authorities, but are argued to be self-regulated due to weak and inconsistent supervision, with new facilities and expansions approved without much scrutiny and/or conditions (HMI, 2019).

The regulatory framework for medical schemes has the following features:

• • Medical scheme membership is voluntary, although many employers require medical scheme participation as a condition of employment. Two types of medical schemes are permitted. First are those established by employers, trades union or industries, referred to as restricted schemes. Second are those that compete commercially, referred to as open schemes. In 2020, there were 57 restricted schemes covering 4.1 million beneficiaries and 17 open schemes covering 4.8 million beneficiaries (CMS, 2005–2020).

• • Medical schemes are non-profit, with a prohibition on the distribution of surpluses. In practice, however, for-profit intermediaries are argued to capture the surpluses through over-priced administration agreements or through risk transfer arrangements (reinsurance or capitation agreements) permitted by weak or conflicted boards (McLeod et al., Reference McLeod, Slattery and Van den Heever2001; CMS, 2008; van den Heever, Reference van den Heever2012; HMI, 2019).

• • Medical schemes are required to have independent trustees, 50 per cent of which must be elected from amongst the members. The remainder are appointed in terms of rules made by each scheme at their discretion. The independence of these trustees is however questioned, with the boards heavily influenced by the for-profit third-party administrators that operate them (HMI, 2019).

• • All medical schemes are required to community rate their offerings or plans, referred to as ‘options’ within schemes. Differentiation is permitted only based on income, principal member (the contributor), adult dependent (a spouse or life partner) and child dependents. While this regime permits a degree of broad age rating (between adults and children) discrimination based on individual health status is strictly prohibited.

• • All scheme plans must cover a set of mandatory minimum benefits, typically referred to as prescribed minimum benefits (PMBs). The PMBs are specified as a comprehensive positive list of 271 major medical conditions and associated treatments.Footnote ¹ In addition, there are 26 chronic conditions with associated treatment algorithms. Coverage must include, diagnosis and treatment, whether in- or out-of-hospital. All conditions are identifiable using an ICD10 code, which by law must accompany all claims made to medical schemes. No co-payments or deductibles are permitted for PMBs and all benefits must be covered out of the scheme risk pool. The PMBs focus on conditions that have catastrophic health and financial outcomes. Outside of PMBs, medical schemes can implement discretionary co-payment requirements, deductibles and benefit ceilings.

• • Member movements between open medical schemes are governed by open enrolment requirements qualified by limited penalties for those instances where anti-selection risks are high. Open schemes are not permitted to distinguish between group and individual membership. Restricted schemes can however reject applications from anyone outside of the employer or industry. They must nevertheless accept everyone from their employer or industry.

• • As membership of medical schemes is voluntary, a system of anti-selection penalties are in place. People who remain for long periods outside of the medical schemes system face contribution penalty surcharges up to a maximum of 75 per cent of the underlying contribution rate. The extent of the penalty varies in accordance with the number of years they have not been a member of a scheme over the age of 30.

• • The prices for healthcare goods and services are not regulated, apart from medicine prices, which are subject to a single-exit pricing regime.Footnote ² Market participants therefore set prices in private negotiations.

• • Income-related cross-subsidies are provided to contributors through a system of tax credits administered through the tax system. However, this subsidy excludes non-taxpayers, leaving around 5 million people who earn too much to qualify for means-tested subsidised hospital services and too little to receive the tax subsidy (Ministerial Task Team on Social Health Insurance, 2005; van den Heever, Reference van den Heever2016).

• • There is no social mechanism in place to subsidise continuation coverage in the post-retirement period (McLeod, Reference McLeod2007). Therefore, many pensioners either reduce their coverage or drop out of cover. Continuation coverage is nevertheless protected, provided pensioners can afford the contribution. There is no conclusive information on the numbers that completely drop out of coverage. Nevertheless, a sizable group of people over the age of 65 remain in coverage, amounting to 8.97 per cent of all beneficiaries in 2020 (CMS, 2005–2020).

5.1 Age trends

The demographic make-up of the medical schemes system has not changed significantly over the reform period from 2000 (Figure 1) although there are changes based on scheme type (open or restricted). The average age of beneficiaries increased only slightly from 31.7 in 2005 to 33.5 in 2020. Open (commercial) medical schemes did, however, increase materially from 31.5 to 35.3. This can be attributed to the introduction of the Government Employees Medical Scheme (GEMS) as a restricted scheme for civil servants, which started taking members from 2006.

Figure 1. Medical schemes: average age by scheme type for the years 2005–2020.

Source: Derived from Council for Medical Schemes (2005–2020).

GEMS grew rapidly from zero members in 2005 to 1.5 million by 2020 (CMS, 2005–2021), drawing much of its membership from open schemes formerly offering individual cover to civil servants. By 2020, GEMS had reached 2 million beneficiaries. The aggregate demographic outcome for restricted schemes therefore shows a decline in average age from 32.2 in 2005 to 31.5 by 2020 as GEMS attracted only active employees, effectively excluding civil servants who retired before 2005.

5.2 Medical scheme consolidation

The medical schemes system has consolidated significantly over time, with the number of schemes declining from 131 in 2005 to 74 by 2020 (Table 2). The consolidation has been more dramatic for open schemes, which reduced from 47 in 2005 to only 17 by 2020. Restricted schemes reduced from 84 in 2005 to 57 by 2020, with the more muted decline attributable to their lower exposure to demographic risk.

Table 2. Number of schemes by type of scheme from 2005 to 2020

Source: Derived from Council for Medical Schemes (2005–2020).

The consolidation of open medical schemes appears to result from demographic features, using below or above market average age profiles in 2005 as a proxy for beneficial or adverse risk profiles (Table 3).

Table 3. Open schemes: average age and scheme consolidation from 2005 to 2020

Source: Derived from Council for Medical Schemes (2005–2020).

Consolidation over the period 2005–2020 was slower, although significant, for schemes with beneficial risk profiles in 2005, reducing from 23 to 11. By comparison, those with adverse risk profiles declined from 24 to 6. Overall beneficiary numbers, which remained generally flat over the period, grew for schemes with better risk profiles in 2005 (from 3.7 to 4.1 million) while those with adverse risk profiles declined (from 1.2 to 0.7 million).

The average age for schemes with beneficial risk profiles in 2005 also remained significantly below that of schemes with adverse risk profiles for the full period from 2005 to 2020.

Schemes with historically beneficial risk profiles are therefore more likely to remain in the market and grow, while those with historically adverse risk profiles are more likely to go into decline and exit. It is furthermore quite probable that the establishment of GEMS contributed to the exit and increased vulnerability of the open schemes with adverse risk profiles in 2005.

5.3 Medical scheme design

Medical schemes are permitted to offer multiple plans, provided they are financially viable, community rate, offer PMBs and comply with open enrolment. From 2005, the number of plans per open scheme has increased from just over five in 2005 to around seven by 2020 (Figure 2). The larger schemes offer from 10 to 17 plans (CMS, 2005–2021).

Figure 2. Open schemes: medical schemes and plans per scheme from 2005 to 2020.

Source: Derived from Council for Medical Schemes (2005–2020).

Restricted schemes, by way of contrast, have averaged only two plans per scheme (CMS, 2005–2021). This is attributable to their different operating environments. Open schemes face the possibility that their risk profile could change rapidly, unlike restricted schemes, with consequences for the cost of coverage and pricing.

Open schemes adapted to this operating environment by offering multiple plans with different levels of coverage and associated pricing (Table 4). Lower-priced plans, with reduced cover, tend to be attractive to younger and healthier families, while more comprehensive plans are important to families with older and sicker members. Comprehensive plans are also attractive to high-income groups regardless of their health status.

Table 4. Average age by monthly medical scheme contribution bands for the years 2005–2020 (2020 prices) aggregated for all medical schemes

Source: Derived from Council for Medical Schemes (2005–2020).

As schemes are not permitted to underwrite or restrict the movement of members between plans, a level of anti-selection is mitigated through this design. Older and sicker people effectively pay more for their coverage over time, while high-income groups stabilise costs by ‘buying up’ (pay more to access better coverage) into comprehensive cover regardless of their health status. Members therefore largely self-select by risk group rather than being actively grouped (risk-selected) by schemes.

Table 4, which provides the distribution of average beneficiary ages according to plan contribution bands (plan contributions per member were adjusted to constant 2020 prices for all years and allocated into the relevant bands), shows that average ages are structurally lower for the lower bands and higher for the higher bands. Although there are slight structural changes over the period 2005–2020, the broad approach has been retained over time for both open and restricted schemes.

Although restricted schemes tend to have fewer plans than open schemes on average, the largest restricted schemes tend to behave like open schemes, with similar designs. This is for two possible reasons. First, where employees have a choice between the restricted scheme and open schemes, as occurs with GEMS, benefit designs need to be kept attractive by risk group to avoid anti-selection against the scheme. Second, regardless of whether a restricted scheme is voluntary or mandated by the employer, employees may prefer better priced open schemes and evade the mandate where loopholes exist. For instance, families can, inter alia, divide their participation between the restricted scheme and open schemes.

Smaller restricted schemes are however designed around a single employer and typically have only one or two plans. Participation is often supported by cross-subsidies for post-retirement continuation members. Because membership in the restricted scheme sample is heavily weighted toward the few large schemes, when aggregated, the age distribution by contribution band approximates that of open schemes.

5.4 Cross-subsidies

Although the self-selection of plans by level of coverage reduces cross-subsidisation between risk groups, both open and larger restricted schemes cross-subsidise the more comprehensive plans through some over-pricing of the most popular plans. This is indicated by the underwriting surpluses in the most populous options, with deficits in the more comprehensive, and less populous options (Table 5 and Figure 4 seen together). These cross-subsidies are seen across the market and appear necessary to avoid death spirals in the more comprehensive options – thereby also protecting the pricing of the most commercially important options.

Table 5. Aggregate surplus/deficit by monthly risk contribution band (Rands) for the years 2005–2019 (R'million) (2020 prices)

Source: Derived from Council for Medical Schemes (2005–2020).

From 2005 to 2020, the most populous options have increased in cost for both open and restricted schemes (Figure 3). Although not identifiable from the data, it is unlikely that this movement represents a ‘buy-up’ phenomenon. A probable explanation is that the most populous options increased their contribution costs in real terms with no benefit improvements. These increased costs are potentially attributable to real increases in provider costs that are then passed on to members. This suggests that the demand for medical scheme coverage in South Africa is both income and price inelastic.

Figure 3. Distribution of beneficiaries by monthly contribution band (based on risk contributions) for the years 2005–2020 (2020 prices).

Source: Derived from Council for Medical Schemes (2005–2020).

5.5 Corporate group consolidation

The consolidation of medical schemes offers only a partial picture of the probable market dynamics that influence competition between schemes and provider purchasing. For-profit intermediaries that contract to medical schemes to provide administration services, which includes provider contracting, influence both consumer and scheme decisions.

These intermediaries in many instances form part of large corporate groups with subsidiaries that: provide financial product advice to consumers and employers (brokers); sell health insurance products; manufacture, distribute and retail medical products; and provide health services (HMI, 2019).

Treating the main holding company in a corporate group as the overall strategic decision-maker, only two of these groups, Remgro Limited (Remgro) and AfroCentric Investment Corporation Limited (AfroCentric) account for 83.9 per cent of the market share of schemes based on membership (Figure 4). Of these two, Remgro controls 65.4 per cent of the market, up from only 28.0 per cent in 2005. As at the time of writing, Remgro is also the largest shareholder in the largest hospital group in South Africa, MediClinic (HMI, 2019).

Figure 4. Consolidation of third-party administrators grouped by corporate ownership 2005–2020.

Source: Derived from Council for Medical Schemes (2005–2020).

Administrators outside these two groups, identified as ‘other’ in Figure 4, show a substantial decline in market share from 57.7 per cent in 2005 to a mere 16.1 per cent by 2020. The consolidation has been rapid, with most of the change occurring in the years 2015–2017.

The various ownership and incentive-related relationships (essentially some form of remuneration for bringing business) between the corporate groups and brokers is used to direct the choices consumers and employers make concerning medical scheme participation (CMS, 2008). This confers significant market power on the corporate groups in a position to deploy brokers, as they can direct consumers to schemes they control and retain them (CMS, 2008; van den Heever, Reference van den Heever2012; HMI, 2019). Schemes with less influence on brokers are in a weak position when it comes to attracting and retaining members.

The market advantage of the corporate groups also extends into provider purchasing. When acting for the medical schemes they operate, the overall purchasing power of the corporate groups, together with the various vertical relationships, permit them to obtain preferential contract terms (confirmed by the HMI (2019)), resulting in higher provider costs for schemes falling outside the two corporate groups. The apparent death spirals for many schemes are therefore potentially exacerbated by these market disadvantages, as the provider cost disadvantages feed into anti-selective decisions by members, which can in-turn be supported by brokers tied in one way or another to the main corporate groups.

5.6 Provider costs

The private health system in South Africa has been characterised by substantial real cost increases from the early 1980s (CMS, 2008; van den Heever, Reference van den Heever2012). The early increases were related to a structural movement of providers away from the public health system in the mid-1980s. The later stages, from the end of the 1990s into the 2000s, however, have been attributed to both price and demand increases, with supplier-induced demand (SID) heavily implicated (CMS, 2008; HMI, 2019).

Figure 5 shows that index changes (index = 100 in 2000) in membership over the period 2000–2017 (which is flat from 2013) were exceeded by index changes in the supply of acute hospital beds as well as critical care beds. Claims cost changes for hospital and specialist services also exceed both supply and membership.

Figure 5. Index changes – critical care beds, specialist and hospital per capita claims (in constant prices) and members from 2000 to 2017 (2000 = index value of 100).

Sources: Derived from Hospital Association of South Africa (2000–2017) and CMS (1980–2021).

The HMI found that SID related to hospital beds, critical care beds and specialists, was a major contributor to the claims cost increases experienced by medical schemes.

‘Rates of hospital admission are positively associated with levels of supply of hospital beds, after adjusting for clinical and demographic factors. Where there is a greater proportion of hospital beds to the population, there is higher rate of admissions and greater utilisation. We also find that the supply of ICU beds is significantly positively correlated with ICU admissions, suggesting that excessive utilisation is more likely to be experienced in areas where there is discretion around whether or not to admit a patient. We also find that the supply of practitioners is significantly positively associated with a higher rate of admission in eight to nine out of ten specialties where the level of discretion around admission is exercised’ (HMI, 2019: 95).

Contributing to the SID problem is the apparent absence of any meaningful incentive for medical schemes to replace standard fee-for-service (FFS) reimbursement with contracts that share risk on demand and quality.

In part this may stem from the accumulated market power of hospital groups, which have consolidated considerably over time, with the three main hospital groups now accounting for 83.2 per cent of the acute hospital beds and 90.1 per cent of all admissions (Table 6). However, to date medical schemes have not systematically intervened to move the market away from FFS, suggesting that the problem lies principally with medical scheme incentives.

Table 6. Consolidation of the market for acute healthcare facilities

Source: Health Market Inquiry (South Africa) (2019).

7.1 Precondition 1: free consumer choice of insurer (rating = 10)

Consumers are largely able to choose their medical scheme, except where an employer mandates a restricted scheme. Employer mandates also don't completely prevent an employee from joining another scheme through a spouse.

7.2 Precondition 2: consumer information and market transparency (rating 2)

Consumers do not have good access to relevant information regarding: competing medical scheme benefit offerings due to non-standard benefit representations, the complexity of the benefits, the multiplicity of plans on the market and the conflicted advice offered by brokers who are effectively paid by the schemes; the quality of the coverage on offer, as the implications of excluded benefits cannot reasonably be understood ex ante by consumers; and the quality of services provided, where no performance-related information is made available to the public, either through a regulator or a medical scheme. The ability of consumers to exercise informed choices for medical scheme and provider selection is therefore very low.

7.3 Precondition 3: price-sensitive insurers and providers (rating 2)

Both medical schemes and private healthcare providers can avoid the consequences of cost and quality failures, which constrain the potential for consumer to make informed choices. The precondition of price-sensitive insurers and providers is therefore not met.

The regulatory framework does, however, make it easy for consumers to move between schemes and plans. If better informed, this ease of movement could improve the incentives of medical schemes and providers to respond to the cost and quality of coverage and services. The achievement of this precondition is therefore related to precondition 2.

7.4 Precondition 4: contestable markets, with low barriers to entry (rating 2)

The extent of market consolidation for both medical schemes and hospitals is indicative of steep barriers to entry, which suggests that this precondition is not met.

For medical schemes, new open (or commercial competing) medical scheme entrants face size-related and demographic constraints. Small schemes face the risk that a few large claims and/or an adverse risk profile could arise very quickly, given open enrolment, forcing it to exit the market. Were these to be addressed, despite the hospital market consolidation, schemes would have an incentive to alter provider efficiencies through contracting.

To address this precondition, four interventions have been proposed (CMS, 2008; HMI, 2019). First, risk equalisation is needed to address the demographic constraints. Second, a version of social reinsurance is required to mitigate the barriers created for small risk pools. Third, hospital licensing requirements need to ensure more diversity in hospital ownership. Fourth, the competition authorities need to avoid mergers that excessively consolidate the market.

7.5 Precondition 5: freedom to contract and vertically integrate (rating 10)

The definition of the ‘business of a medical scheme’ contained in the Medical Schemes Act (Republic of South Africa, 1998b) permits a registered medical scheme to make any arrangement necessary to comply with its obligations to cover the costs of medical benefits. This includes the ability to establish selective contracts with provider networks and for the medical scheme to own their own health services.

While this precondition is largely met, the failure of the market to exploit these opportunities appears to lie with the weak incentives medical schemes have to compete on the cost and quality of healthcare services.

7.6 Precondition 6: effective competition regulation (rating 7)

The supervision of competition in South Africa was enhanced in 1998 with the introduction of the revised Competition Act (Republic of South Africa, 1998a), which established a new supervisor, tribunal and an appeal court.

The private health system has become a specialist area of focus, which resulted in the establishment of a major market inquiry (HMI, 2019) to better understand competition dynamics in the sector. The findings and recommendations now feed into merger investigations and examinations of market conduct. As these decisions improve, based on the accumulation of information on the industry, this precondition is in the process of being met.

7.7 Precondition 7: cross-subsidies without incentives for risk selection (rating 2)

The regulatory framework for medical schemes in South Africa does not have any mechanism in place to pool risks at the system level such that incentives for risk selection are removed. There are some market-related incentives to avoid excessive ‘buy-down’ (pay less for reduced coverage) from comprehensive plans that could be regarded as a cross-subsidy. Nevertheless, a version of self-selected risk selection exists, which means that this precondition is not met.

7.8 Precondition 8: measures to avoid free riding and anti-selection (rating 7)

Membership of medical schemes is voluntary, ostensibly creating the conditions for anti-selection. Measures to mitigate anti-selection, however, include waiting periods and late joiner penalties. These measures have proven adequate to preserve the financial stability of medical schemes from 2000. This precondition, although not entirely adequate, is to some extent in place. The introduction of membership mandates would establish a more complete regime.

7.9 Precondition 9: quality of care supervision (rating 1)

Although medical schemes, in conjunction with the intermediaries that support them, are in a strong position to provide the public with information on health provider performance due to the universal requirement to include diagnostic codes on all billing, they have not done so to date. This may be related to their weak incentives to compete on the cost and quality of health services. There is however no regulatory supervision of quality in the private sector, with the result that this precondition is not met.

The Health Market Inquiry (HMI, 2019) consequently recommended that an information regulator be implemented, with the authority to impose uniform standards of information production and to compel reporting on health provider outcomes. This information can be collected from both schemes and providers.

7.10 Precondition 10: guaranteed access to basic care and affordable out-of-pocket care (rating 3)

The system of PMBs has been implemented without three key features needed to ensure guaranteed access to basic care in the private sector. First, the benefits involve a complex list of diagnosis treatment pairs, making it difficult for members to compare coverage adequacy across schemes and plans. Second, the PMBs are not implemented in conjunction with a risk equalisation scheme, creating incentives for schemes to evade compliance rather than to compete on cost and quality. Third, there is no system of price supervision that can address the prices charged for PMBs and out-of-pocket expenses.

This precondition has therefore not been adequately met.

7.11 Precondition 11: no conflict of interest by the regulator (rating 1)

All the main office bearers of the Council for Medical Schemes, the regulator of medical schemes, are political appointments. This potentially exposes the regulator to interference by vested interests operating through political structures. This precondition is therefore not met.