یارانه های حق بیمه و بیمه سلامت اجتماعی: جایگزین ها یا مکمل ها؟

Premium subsidies have been advocated as an alternative to social health insurance. These subsidies are paid if expenditure on health insurance exceeds a given share of income. In this paper, we examine whether this approach is superior to social health insurance from a welfare perspective. We show that the results crucially depend on the correlation of health and productivity. For a positive correlation, we find that combining premium subsidies with social health insurance is the optimal policy.

rivate health insurance markets discriminate according to risk of illness. Those with a higher risk of illness usually have to pay higher premiums than those with a lower risk. In many countries, this price discrimination is regarded as unjust, violating equity principles such as ‘equal access’ or ‘solidarity’. A common solution are social health insurance schemes which establish transfers from low risks to high risks by forcing all citizens into one health insurance contract with a uniform premium. In a recent paper, Zweifel and Breuer (2006a) fundamentally question this equity argument in favor of social health insurance. They maintain that being a high risk does not necessarily imply that a person should receive transfers: “[Uniform premiums] result in a cross-subsidization of high-risk by low-risk, low-income individuals. This can result in counter-productive effects. For example, a healthy young worker subsidizes a wealthy older manager who is a heavy user of medical services. Equity considerations seem to call for redistribution from everyone else to the double disadvantaged, viz. the high-risk, low-income individuals.” (Zweifel and Breuer, 2006a, p. 172) Based on the above argument Zweifel and Breuer propose to substitute social health insurance by “premiums subsidies”. These subsidies are targeted to individuals whose expenditure on health insurance exceeds a given share of income. By this policy, they want to focus transfers on high-risk, low-income individuals. Zweifel and Breuer also advance efficiency arguments for risk-based premiums, stating that these allow cost sharing to be tailored to the individual risk type, thereby dealing better with moral hazard. In addition, they point out that risk-based premiums avoid possible costs due to risk selection induced by uniform premiums. This applies if social health insurance is provided by competing insurers. The case for social health insurance also depends on the severity of risk discrimination in private health insurance markets. For the individual health insurance market in the US, Pauly and Herring, 1999 and Pauly and Herring, 2007 find that premiums are not proportional to risk, pointing to a substantial amount of risk pooling. However, risk pooling is only partial because higher health risk is significantly related to higher premiums overall and to lower coverage rates in unregulated states (Pauly and Herring, 2007, pp. 775–776). Social health insurance is also defended by its effect on the income distribution. Empirical studies show that poverty and ill-health are positively correlated.1 For this reason, McGuire (2006) argues that social health insurance may well be optimal from a second-best perspective. It redistributes not only to those with higher health risks but also tends to make the poor better off. Formally, this line of reasoning has been analyzed by Cremer and Pestieau (1996) who show that a positive correlation of health and income provides a strong argument for social health insurance. van de Ven (2006) criticizes the concept of premium subsidies advocated by Zweifel and Breuer. He points out that high-risk, low-income consumers have little incentive to shop around for a well-priced health plan if their premiums are subsidized. Furthermore, the fact that they receive a subsidy at the margin creates a moral hazard problem. These individuals will tend to over-insure.2Zweifel and Breuer (2006b) also mention a negative incentive effect of their proposal. Low-income individuals who receive a premium subsidy effectively face a higher marginal tax rate as the subsidy is decreasing in income. Whether premium subsidies in combination with risk-based premiums are an alternative to social health insurance is therefore an open issue. The fact that in Switzerland premium subsidies go along with social health insurance raises the general question whether premium subsidies are substitutes or complements to social health insurance. Overall, an assessment of the ZB proposal calls for an analysis of optimal government interference in health insurance markets to advance equity objectives. In the following, we develop a theoretical framework for this purpose which considers social health insurance as well as premium subsidies. We allow for heterogeneity in productivity and risk types. The government maximizes a social welfare function and uses a linear income tax to redistribute between high and low-productivity individuals. To support high-risk individuals, it can pay premium subsidies if expenditure for health insurance exceeds a given share of pre-tax income or introduce social health insurance. Since Zweifel and Breuer want to target transfers to the worst-off in society, we pay particular attention to the solutions for a maximin social welfare function. In addition, we present results for the utilitarian welfare function. We examine three schemes in detail. The benchmark is social health insurance combined with optimal linear taxation, a scheme which has been analyzed in detail by Cremer and Pestieau (1996).3 The second is the proposal by Zweifel and Breuer with risk-based premiums and premium subsidies. The third scheme combines premium subsidies with social health insurance, an approach which is taken in Switzerland. Building on these results, we extend the analysis and examine whether different combinations of social health insurance and premium subsidies can increase welfare. Our model takes explicitly into consideration the incentive effects on labor supply, in particular, those due to changes in the marginal tax rate induced by premium subsidies. Furthermore, we allow for different degrees of correlation of health and productivity. To keep the analysis tractable, we abstract from further moral hazard problems. For private health insurance markets, we assume that premiums are actuarially fair given an individual's risk type. We do not consider partial risk-pooling in the private health insurance market. The paper is structured as follows. In Section 2 we present the model. Section 3 introduces premium subsidies and examines when these will be claimed by individuals. In Section 4 we present the general problem of choosing premium subsidies and social health insurance and analyze different solutions. Section 5 concludes.

In many countries, equity principles such as ‘equal access’ or ‘solidarity’ are regarded as important in the context of health insurance. One way to achieve these objectives are social health insurance schemes which establish transfers from low to high risks by forcing all citizens into one health insurance contract with a uniform premium. Zweifel and Breuer (2006a), however, argue that the best way to redistribute to the double disadvantaged, i.e., high-risk, low-income individuals, is to combine risk-based premiums with premium subsidies. As Switzerland shows, premium subsidies can also be used in combination with social health insurance. This raises the question whether premium subsidies are substitutes or complements to social health insurance. In this paper, we assessed the merits of social health insurance and premium subsidies in a theoretical framework. First, we characterized the optimal solution for the ZB proposal as well as for social health insurance. We found that the correlation of health and productivity is crucial. The ZB proposal is more likely to be superior, the less positive this correlation. Second, we compared the ZB proposal with a social health insurance scheme that also contains a premium subsidy. This changed the results strongly in favor of social health insurance and shows that a considerable part of the welfare advantage is due to the fact that premium subsidies introduce an element of non-linear taxation. In particular, we found that premium subsidies complement social health insurance if health and productivity are positively correlated. Our results apply for both maximin and utilitarian welfare. Numerical simulations which allow for general combinations of social health insurance and premium subsidies confirm our findings. Our main insight is that premium subsidies are complements to social health insurance if health and productivity are positively correlated. Only if there is a considerable negative correlation between productivity and health, premium subsidies substitute for social health insurance. The result by Cremer and Pestieau (1996) that a positive correlation of health and income provides a strong argument for social health insurance can therefore be extended to include premium subsidies. This supports the argument by McGuire (2006) that social health insurance can be optimal from a second-best perspective if poverty and ill health are positively correlated. A limitation of our analysis is that we abstracted from moral hazard effects. This would shift the argument in favor of the ZB proposal if there are considerable benefits of making optimal cost sharing dependent on risk types. On the other hand, premium subsidies reduce the incentives to search for a well-priced health plan. Furthermore, individuals who receive a premium subsidy will tend to over-insure. Our analysis did not include the retired who cause a large part of health care expenditure. Typically, social health insurance is organized on a pay-as-you-go basis and therefore redistributes between age groups. The impact of premium subsidies then depends on how premiums vary with age. Future work could consider these effects in a model with overlapping generations. Finally, we assumed that risk-based premiums are actuarially fair given an individual's risk type. Although a common assumption in the literature, studies by Pauly and Herring, 1999 and Pauly and Herring, 2007 indicate a substantial amount of risk pooling even if risk-based premiums are allowed. An encompassing evaluation of the merits of premium subsidies in combination with risk-based premiums would have to take into account this aspect as well. From a policy point of view, our analysis shows that on equity grounds a strong case can be made for social health insurance. In particular, the positive correlation of health and income makes social health insurance attractive. Whether efficiency gains in form of better control of moral hazard can outweigh this equity argument still needs to be demonstrated. This also holds for alternative approaches of targeting based on the risk types of individuals (Pauly et al., 1992 and van de Ven et al., 2000). Subsidies based on risk-based premium expenditures only, however, do not seem to be a viable alternative to social health insurance.