20/2/2011

IS THERE AN ECONOMIST IN THE HOUSE?: AN ECONOMIC APPRAISAL OF THE US HEALTHCARE SYSTEM IN LIGHT OF THE 2010 REFORMS AND ALTERNATIVE STRUCTURES.

Anonymous Student Number: Z0905156

EXECUTIVE SUMMARY.........................................................................2 I. INTRODUCTION................................................................................5 II. THE ECONOMIC THEORY OF HEALTHCARE.........................................7 II.I IMPERFECT KNOWLEDGE..........................................................................8 II.II UNCERTAINTY..................................................................................11 II.III SOLUTIONS FOR IMPERFECT INFORMATION IN HEALTHCARE..................................16 II.IV EXTERNALITIES................................................................................17 II.V EQUITY........................................................................................19 III. THE US HEALTHCARE SYSTEM: STRUCTURE AND PROBLEMS...........20 III.I EQUITY AND THE US.........................................................................20 III.II THE STRUCTURE OF THE US HEALTHCARE SYSTEM........................................21 III.II.I THE PRIVATE SECTOR......................................................................26 III.II.II THE PUBLIC SECTOR.......................................................................35 IV. THE AFFORDABLE CARE ACT.........................................................39 IV.I THE ACA AND ADVERSE SELECTION........................................................39 IV.II THE ACA AND MORAL HAZARD............................................................43 IV.III THE ACA: IMPERFECT KNOWLEDGE AND EXTERNALITIES.................................46 V. ALTERNATIVE HEALTHCARE SYSTEMS............................................47 V.I ALTERNATIVE METHODS OF FINANCE AND PROVISION........................................48 V.II INTERNATIONAL QUALITY COMPARISONS......................................................54 V.III LESSONS FOR THE US.......................................................................57 VI. CONCLUSION...............................................................................59 BIBLIOGRAPHY.................................................................................61 APPENDIX A. INSURANCE MARKETS...................................................70 APPENDIX B. THE ADVERSE SELECTION DEATH SPIRAL........................74

Executive Summary
2

The US health care system is currently undergoing a period of extraordinary change due to the introduction of the Affordable Care Act (ACA) of 2010. This paper has two aims: first, to provide a comprehensive economic appraisal of the reforms by first evaluating the pre-ACA health care system and analyzing the proposed reforms. Secondly, this paper aims to consider alternative health care systems to contextualise the discussion and find lessons for the US system that may be incorporated in future reforms. The paper begins by providing the economic framework upon which the discussion is based. It is found that the private market for health care requires some form of Government intervention to compensate for substantial market failures, in particular imperfect knowledge, imperfect information and external effects of vaccination consumption. As the US system gives a low weighting to redistributive concerns, redistribution is only briefly considered. It is found that the US system suffers from the market failures outlined in the framework, despite attempts to curtail information problems via use of employer-provided health insurance and managed care organisations. In addition, cost-containment is found to be poor. It is concluded that the ACA should comprehensively address these information imperfections and increase the range of insurance choices provided by employers. The ACA is found to be largely an improvement on the status quo, however it does not fully confront a root cause of spiralling costs instead, it relies heavily on new tax revenue from Medicare, leading to
3

questions regarding its sustainability. Regarding alternative systems, it is found that the US already incorporates characteristics of other systems in Medicare and Medicaid, and the reforms appear to move it towards a Switzerland-style model; it is also noted that use of cost-containment systems from alternative systems could be adapted for use in the US system.

I. Introduction
On March 23rd 2010 the President of the United States, Barack Obama, signed the Patient Protection and Affordable Care Act, which was to be shortly joined by a package of amendments - the Healthcare and Education Reconciliation Act - on March 30th 2010. The final aggregate, called the Affordable Care Act (ACA), has proven to be one of the most divisive documents of social policy in the US in recent memory 1, and realises an ambition held by numerous presidents, from Theodore Roosevelt to Bill Clinton, to provide a substantial, systemic reform of the US Healthcare system. The ACA is fascinating on two levels. In one consideration, it attempts to tackle the failings of a system that is aimed at providing a vital service in the Worlds largest economy; in a 2009 report by the US Census Bureau, 16.7% of Americans were found to be without healthcare insurance in 2009, a 1.3% rise on 2008. As this paper explains, this problem is a manifestation of the failings of a privately funded healthcare system, and problems continue to those fortunate enough to have insurance. The other source of interest comes from the anomalousness of the US structure amongst the OECD countries. A paper by Anderson and Poullier (1999) finds that in 1997 the median percentage of populations covered by a Government-assured health insurance scheme in the OECD nations was
1

A USA Today/Gallup Poll conducted on March 22nd 2010 found that, nationally, 49% of adults felt the passing of the ACA would be a good thing, with 40% claiming it would be a bad thing.

100%; by stark contrast, the US was at the bottom of the list with 33.3%. This betrays an implicit value judgement pertaining to the role of Government in the healthcare system, which has been enforced to a unique degree. At such a crucial time in the evolution of the US healthcare system, a comprehensive study is therefore warranted to understand what has put so many Americans in such a dire state vis--vis their healthcare coverage, to analyse a unique structure of fundamental importance, and confront the logic of reforms that threaten to change the rules of the game. This paper will attempt to provide answers to three distinct, yet intimately related questions. Firstly, how did the pre-ACA structure function and how did it perform? This must surely provide the base for any discussion of such a complex and powerful topic, and is the consideration of chapter III. From this base, we will confront the ACAs main features to answer our second question from an economic perspective, how relevant and potent are the reforms in the ACA? This question shall be the focus of chapter IV. Chapter V answers how do alternative healthcare systems compare with the pre-ACA system? This will contextualise the discussion, and provide solutions to the failings of the pre-ACA system that can, to a degree, be contrasted with those outlined in the ACA. A theoretical framework underpins the analysis, and chapter II will provide a brief summary of the most significant characteristics and failings of a private healthcare market. The conclusion ties the paper together and projects a statement
6

pertaining to where the US healthcare system is currently and where it is likely to go.

II. The Economic Theory of Healthcare

To start, it is integral to reaffirm that the focus of the discussion is indeed about healthcare and not health status. Health status derives from multiple sources, of which healthcare is but one; Arrow (1963) notes the importance of necessities such as clothing and food, whilst Phelps (2003) makes the distinction that consumption goods can be split into those that will add to, or detract from, your stock of Health. This paper focuses on healthcare to keep discussion concise, however the reader is urged to remember that the economic perspective of the healthcare system is only a single perspective of a single factor in the goodness of the health of a nation. Healthcare is subject to onerous Government intervention, as noted by Phelps (2003), in all aspects of the market; inter alia, on the demand side by backing a substantial proportion of individuals, and on the supply side by rigorously testing physicians and technologies to ensure a high degree of quality. Should the market work in a Pareto optimal fashion, according
7

to what Arrow (1963) calls the First Optimality Theorem, the Pareto optimal point should be a competitive equilibrium; the presence of heavy Government intervention is thus evidence that something is

fundamentally wrong in the private market for HC. This naturally leads us onto the question what are the conditions that a pure market would need to meet in order to satisfy the First Optimality Theorem and which of these does the HC market fail? Barr (2004) names these conditions the standard assumptions2 although failure to meet one or more of these is not necessarily to preclude the market from being able to function with relative freedom from the state, due to different degrees of market failure and Government intervention. Healthcare suffers from serious information problems and externalities which mean that any HC system must be constructed to ensure that these failures are assuaged rather than proliferated. We shall proceed by thus considering these failures and discussing equity, which plays an integral part in the construction of any healthcare system. This paper does not consider the remaining standard assumptions in any length as they do not, relative to the main problems, pose an insurmountable threat to the functioning of a private healthcare system. II.i Imperfect knowledge

To begin with the consideration of imperfect knowledge, it is first important to understand the nature of healthcare demand, and therefore

Perfect competition, perfect information, no public goods, no externalities and no increasing returns to scale.

the Grossman (1972) model. The models key conclusion is that rational, fully informed consumers will invest in healthcare until the marginal benefit of that investment is equal to the marginal cost. Contentious empirical results3 notwithstanding, the Grossman model is an important pillar of health economics in that it allows us to model the link between demands for good health status and healthcare. The model assumes perfect information, however it is more plausible that there is imperfect information and hence non-optimal consumption due to incorrect evaluation of the benefits and costs. To address this, an individual will seek assistance from a physician; however, the physician has a conflict of interest between what Arrow (1963) notes as the social expectation of the physician providing accurate advice, and supplierinduced demand (SID). SID occurs when a physician encourages superfluous consumption of healthcare; Roemer and Shain (1959) found that, in insured populations, there was a positive correlation between the number of hospital beds available and the number of hospital days used in a population. Therefore, due to imperfect knowledge there are two factors encouraging a sub-optimal level of demand ignorance regarding the benefits and costs associated with each procedure, and the SID effect.

See Wagstaff (1986, 1993)

Kenkel (1990) explores the empirical validity of the ignorance effect and concludes that there is a positive correlation between consumer information and demand thus, the ignorance effect appears to be a reality. The SID effect, however, is far more contentious; In a survey, Wilensky and Rossiter (1983) find that aggregated data tends to be supportive of SID, whilst disaggregated data are more conflicted patientbased studies find little evidence of any SID but physician-based studies find abundant evidence. They conclude that SID is likely to exist, but the magnitude of any such effect is unlikely to be large. Figure 1 shows the ignorance and SID effect4. Let the equilibrium under perfect knowledge be where MB = MC, as the Grossman model predicts, equating to a demand of Q*. However, under imperfect knowledge we have seen that an under informed consumer is likely to consider a lower MB curve of MB. The new demand is Q and Q*-Q is the ignorance effect. Following the conclusion of Wilensky and Rossiter (1983), the consumer may perceive the MB to be slightly above MB due to SID. MB is one possible curve with a corresponding demand of Q, where Q-Q can be denoted as the SID effect. In net, consumption of healthcare is below the optimal level as dictated by the Grossman model.

We assume that the consumer faces a constant marginal cost over time.

10

Figure 1: The Ignorance and SID Effects in HC Demand

II.ii Uncertainty Uncertainty of the timing of when we need treatment, what we are going to need treatment for, and what specific treatment we will need is, as proposed by Arrow (1963), perhaps the most debilitating problem for any private healthcare system. However, certainty is a sufficient but not a necessary condition for a well functioning private market, as the market itself provides an answer to uncertainty via insurance markets. Appendix A provides a brief theory of insurance markets; we shall start here with the equation that links the supply and demand sides of the market:

( 1 + ) pi L = W 1 W *
Equation 1: Supply and Demand Sides of an Actuarial Premium

11

(1+) is a loading factor, p is the probability of illness, L is the monetary loss due to illness, W1 is wealth if no illness is contracted, W* is the certainty equivalent. For insurance to be viable, p must be known or estimatable to allow uncertainty to be transformed into risk. p must also not be zero or one, the latter constraint having serious implications concerning pre-existing, chronic or terminal conditions. Finally, Barr (2004) notes that the probabilities amongst the insured population must be independent of another this is generally true in healthcare except during epidemics in which case the Government will intervene. Crucial to the ability of any insurance markets functioning is the absence of information asymmetries, and it is here where the problems of healthcare insurance become apparent. Healthcare insurance falls prey to both manifestations of information asymmetry, moral hazard and adverse selection, with severe results.

12

Figure 2: Moral Hazard and the Third-Party Payer Problem Moral hazard has two potential manifestations: under-consumption of preventative care and the third-party payer problem. The former stems from the observation by Pauly (1974) that, in the absence of substantial costs, there may be an incentive for the individual to under-invest in preventative goods Morris et al. (2007) notes that an additional repercussion of this is that pregnancy costs are often not covered by insurance plans in the US, as pregnancy can be induced by the insured at any point due to gains from such a contingency; hence, insurance firms act as if p = 1. The third-party payer problem arises due to the endogeneity of L and the structure of the insurance market; a fully insured patient (and physician) will act as if their MPC is 0, when it is likely that MSC > 0. Figure 2 shows that,
13

if

demand

is

price-elastic,

an

individual

will

consume

(D=MPB=MPC), above the optimal level of Q* (D=MPB=MSC), leading to a welfare loss of XZY. However, if demand is price-inelastic (D), there is no scope for moral hazard. An equally pertinent problem is that of adverse selection, highlighted by Akerlof (1970) to be of critical importance in the health insurance market. The inability of an insurer to differentiate between good and bad risks leads to what Cutler and Zeckhauser (1998) term the adverse selection death spiral. Because bad risks will always be better off mixing with good risks, any plan that caters to both risks will become continually more expensive, until it is no longer feasible and either everyone is uninsured or they end up on a single plan that caters specifically to good risks. Whilst a fuller derivation is provided in appendix B, it is sufficient to state that, due to the community rating of premiums based on the expected average risk of a population, good risks (who incur lower premiums) subsidise bad risks (who incur higher premiums). As the presence of bad risks push the groups premium up, good risks drop out in search of a lower cost plan. Their departure further raises the average risk and premium, forcing further dropouts this cycle continues until either of the conclusions stated above. Cutler and Zeckhauser (1998) use the example of such a spiral occurring in a Harvard health insurance plan overall, the plan went from covering 20% of Harvard employees to disbanding in the space of three years. There is a vast empirical literature on both forms of information asymmetry and the third-party payer problem has been observed in a
14

range of healthcare systems. Barros et al. (2008) finds limited evidence for it in a new health insurance scheme in Portugal, but makes two caveats: one has to control for the presence of adverse selection and the SID effect is difficult to empirically differentiate from the third-party payer problem. Coulson et al. (1995) finds prescription drug refills are higher amongst elderly insured populations in the US than those without insurance, suggesting price elastic demand (and therefore moral hazard), whilst Chiappori et al. (1998) find moral hazard for home visits when faced with a 10% copayment5 , but not for visits to the hospital; hence, transport and transaction costs may be a substantial proportion of total costs such that MPC is significantly non-zero even under insurance. Concerning adverse selection, Cutler and Zeckhauser (1998) note the problem of allocative efficiency losses due to the insurer limiting the coverage of a plan as a response to the presence of bad risks. Hence, adverse selection, as established by Rothschild and Stiglitz (1976), leads to two types of equilibria; the failure of a pooling equilibrium (based on community rating) means either the market fails entirely (no one is insured) or an attempt at a separating equilibrium (where good and bad risks are attracted to different plans via differences in coverage meant to cream skim the good risks) leads to poor coverage. Empirically, adverse selection is evident; in addition to the Cutler and Zeckhauser (1998) example, Akerlof (1970) uses the observation of very limited or nonexistent private health insurance plans for the elderly as an example of adverse selection leading to total market failure. Browne (1992)
5

Copayments make the patient pay a certain percentage of the cost of treatment.

15

considers the difference between individual and group insurance markets. He finds that good risks tend to purchase more insurance in the group market6 than in the individual market, and that in the individual market low risks effectively subsidise high risks, as theory predicts. II.iii Solutions for Imperfect Information in Healthcare A wider agenda of health-related education initiatives helps abate the ignorance effect; in stemming superfluous consumption they may also impact on SID. However, a more targeted source of the SID effect is through the payment scheme of doctors; the two predominant physician payment schemes are fee for service (FFS) in which a physician is paid per service provided, and fixed salaries. A study by Hickson et al. (1987) tested the differential impact of payment schemes by segregating the doctors in a paediatric care centre into FFS and salary earners. They found that, subject to a benchmark of the correct level of paediatric care, FFS physicians over provided care whilst salary earners under provided. Hence, switching physicians from FFS to salary payment schemes, and implementing regulations to prevent under provision of care, would be effective in countering SID (and moral hazard by the physician). Because moral hazard arises from the implausibility of observing an individuals actions, attempts to curb moral hazard must focus on incentives to affect their actions. Coinsurance is a popular incentive system, involving the patient facing a portion of the MC of their treatment, and comes in two broad forms: a deductible (a fixed dollar amount that

Group insurance is less susceptible to adverse selection, as discussed in the next chapter.

16

the insured is required to pay for any treatment requested) or a copayment (a percentage of the cost of treatment). Both instruments are common in health insurance packages, however, as the individual does not face the MSC of their actions, they cannot fully compensate for moral hazard. Alternatively a premium that rises disproportionately with the degree of cover could be used, although this is may lead to poor coverage. Insurance companies also make use of incentives, such as noclaims bonuses, to increase MPC. Perhaps most troubling of the information problems, adverse selection has only one full solution mandating health insurance to prevent low risks dropping out of the market and instigating the death spiral. Due to the inherent committal in social health insurance systems 7, the problem of adverse selection does not arise with considerable potency in West European systems. However, whilst such a solution may assuage adverse selection, it may be seen as inequitable that good risks are effectively subsidising bad risks. Costly and invasive medical procedures to ascertain risk are an alternative solution, but impractical in many cases. Finally, there are examples of group insurance acting as a basis for stability in the market; such an example is that of employer provided health insurance (EPHI), explored in detail in the next chapter. II.iv Externalities Although the main failings of the private health insurance market lie in informational problems as described above, there are notable

externalities in the consumption of healthcare. Again, reflecting on the

7

Discussed in chapter V.

17

wider causes of health is important as there are negative consumption externalities associated with certain commodities (e.g., cigarettes) and positive externalities for the research of new treatments. Restricting the analysis to healthcare in particular, there are clear positive externalities associated with the consumption of preventative healthcare specifically, the consumption of vaccinations. Phelps (2003) notes that the more individuals that vaccinate themselves, the lower the expected cost of any individual from contracting the disease. Will the private market account for this? Individuals will not factor in such benefits from herd immunization when deciding to have a vaccination, implying aggregate under-consumption of vaccinations. Jansen et al. (2003) and Boulier et al. (2007) both consider mumps; the former study notes that declining take up of the MMR vaccine in the UK is associated with increasingly large measles outbreaks, whilst the latter estimates large, increasing, MSB associated with the vaccine (more than one case prevented for every vaccination). While Brito et al. (1991) agree with the conclusion that a private market leads to under-consumption, they argue that the Government should use the tax system to provide incentives for vaccinations as opposed to crude mandates. Whilst this may be an attractive proposition for vaccinations against low-fatality diseases, the expected cost against potentially highly fatal and infectious diseases (e.g., Ebola, swine flu) are too high to allow for anything less than total Government control of prevention. such a fact was evident in the stockpiling of swine flu vaccines by Governments during the peak of the pandemic in 2009.
18

II.v Equity Whilst the efficiency arguments outlined above provide a strong case for some form of Government intervention in the healthcare market, the precise flavour of that intervention will be partly dictated by notions of equity. However, there is considerable confusion in the economics literature pertaining to what equity may mean and how one may measure it; a society may consider a distribution of resources equitable even if the distribution is unequal. A useful distinction to make, therefore, is between horizontal and vertical equity: the former is concerned with equal opportunity for those with equal need, whilst the latter is focused on the redistribution of resources within the income spectrum. What is meant by opportunity and need? Culyer and Wagstaff (1993) consider four alternative definitions for both terms and note numerous conflicts between them whilst Mooney (1983) considers seven plausible alternatives for the variable being equalised for individuals of equal need. Broader equity principles also determine the desired magnitude and direction of redistribution via the healthcare system as Williams and Cookson (2000) note, this would ideally lead to an equitable distribution of health as opposed to healthcare. These definitional problems reappear in empirical studies. Horizontal equity is investigated by regression analysis on variables influencing healthcare consumption, the results of which depend on definitions of a need or non-need variable. Vertical equity studies often consider the financing structure to see if it is progressive however, this only considers redistribution in cash, rather than outcomes. For the purposes of this
19

paper it is sufficient to simply note these problems, as we are focused on efficiency arguments; we do, however, contextualise the discussion by considering the broader equity principle of the US system.

III. The US Healthcare System: Structure and Problems

III.i Equity and the US Before one may analyse the reforms included in the ACA, it is pertinent that the current US healthcare system is analysed using the economic logic developed in the previous chapter such that we can list the failings of the system. In turn, it is necessary to note the wider notion of equity that the US subscribes to vis--vis healthcare, as this will have a fundamental bearing on both the redistributional goals of the system and the breadth of any Government intervention. Examination of the US healthcare system betrays a libertarian notion of equity. Figure three shows that private healthcare expenditures still constitute the largest proportion of total healthcare expenditure despite a slowly growing public share8; combined with Anderson and Poulliers (1999) earlier stated conclusion that the US has the lowest proportion of citizens with Government backed healthcare in the OECD, it is clear that there is some restriction on the level of Government intervention in
8

The rise in public expenditures circa 1965 is due to the introduction of Medicare and Medicaid.

20

healthcare. Further credence to this claim is provided by two companion studies by Wagstaff and van Doorslaer (1992a, 1992b) in which equity in the delivery and finance of a selection of developed nations is examined and compared using sample data ranging from 1980-1987. In the former study, the primary concern is horizontal equity, loosely defined as equal treatment for equal need; definition problems notwithstanding, the US is found to have unequal treatment for equal need (although the ultimate direction is indeterminate). The latter study investigates vertical equity ; they find that the US is the most regressive country in the sample, in both the Kakawani and Suit progressivity indices9. The system, therefore, has an implicit limit on the role of Government and is relatively unconcerned with the ultimate distribution that arises. These are classic libertarian traits; it is a procedural notion of equity that is primarily concerned with the manner in which the healthcare is obtained, as opposed to distribution of outcomes whilst libertarians do also allow for a limited role of Government to prevent acute destituteness, the Government strictly facilitates the free market to realise its potential gains. It would be naive to suggest that this is a definitive statement of equity in US healthcare, but for the purposes of this paper it will suffice to note that the ACA reforms should minimise Government presence in the sector and not be constrained by distributional goals except in acute cases. III.ii The Structure of the US Healthcare System

The US scores a -0.145 in the Kakawani index and a -0.160 in the Suit index.

21

The US has a complex healthcare system that provides the first problem in analysis; figures 5 and 6 draw on information from B. S. Klees et al. (2009) and J. K. Iglehart (1999) to show a simplified structure of the private and public sectors respectively, and this paper considers them as largely separate entities for reasons of simplicity and conciseness. Note that the analysis is being conducted is based on the broader structure of the systems, as opposed to their finer detail.

Figure 3: Change in the Composition of US Health Care Expenditures Over Time, Source: CMS NHE Data

22

Figure 4: Comparison of Health Care Expenditures (as % of GDP) Amongst Select OECD Countries, Source: OECD

23

Figure 5: Structure of the Private Sector of the US Health Care System

Populatio n
If in employment, pays via

Pay Premiums to
Direct

Insurance Firms
Premiums are Tax Subsidised

Employ er

If no insurance, pays full cost; with insurance will pay coinsurance Physician and Provides health care

Reimburses according to insurance plan

Physician Practices

24

Figure 6: Structure of the Public Sector of the US Health Care System

Populatio n

Mandatory Tax Contributions (Part A); Voluntary Premiums (Parts B + D)

Federal Governm ent Revenue/ Expendit ure State Revenue/ Expendit ure

Medicare Part A Medicare Parts B + D

Provide s health care

25

Physician and Physician Practices

Pays coinsuran ce

General Tax Revenue

Medicaid

Populatio n

Mandatory Tax Contributions (Part A); Voluntary Premiums (Parts B + D)

Federal Governm ent Revenue/ Expendit ure State Revenue/ Expendit ure

Medicare Part A Medicare Parts B + D

Provide s health care

Physician and Physician Practices

III.ii.i The Private Sector Following Iglehart (1999), we shall follow the healthcare dollar through the private sector, starting with a healthy population and ending with the patient in receipt of care. The first transaction to consider is the payment of the premium to the insurance firm. Figure 5 shows that there are two primary channels through which this transaction can occur the individual can pay the insurance firm directly, or they can purchase health insurance through their employer. The former channel is the traditional model of health insurance upon which the economic analysis of chapter II was based; it is therefore theoretically susceptible to adverse selection10. However, this channel of premium purchase is small relative to insurance plans purchased via an
10

Pays coinsuran ce

General Tax Revenue

Medicaid

Moral hazard, imperfect knowledge and externalities arise at a later stage in the system.

26

employer. A report by the Employee Benefits Research Institute (2010) shows that, in 2009, 91.0% of Americans with either EPHI or individual insurance had EPHI. Two questions naturally follow: why is EPHI so popular? Are there any economic benefits of EPHI? In an EPHI plan, the employer will typically cover a portion of the cost at a subsidised rate due to the tax exemption of any income of the employee that goes towards paying the premium. Indeed, the portion of the premium that the employer covers is substantial; Phelps (2003) calculates the share for individual insurance to be 73% in 1970 and a report by the Kaiser Family Foundation (2010) finds that this percentage had risen to 86% by 1999, although it suffered a decline to 81% by 2010. Therefore, answering the first question, the explicit premium paid via EPHI is smaller than in directly purchased insurance. It is contentious whether such a channel produces a net economic gain. The main benefit comes from mitigating adverse selection; as health status isnt generally a relevant factor in deciding an individuals employment, there will be a random mix of good and bad risks that are unlikely to drop out11. This allows for the insurer to adopt a pooling equilibrium solution based on average risk with relative security from the adverse selection death spiral. A factor contributing to this security is the lack of choice facing employees: Jensen et al. (1997) found that whilst the percentage of employees with two or more insurance choices had increased in the two years from 1993 to 1995, in 1995 the average was

11

Due to EPHI providing cheaper premiums.

27

only 62%. Most of this rise had been concentrated in large firms12 (a rise from 75%84%), a stark contrast to small firms13 (7%10%). This lack of choice may be beneficial in preventing the death spiral, however it is clearly sub-optimal in that it is unlikely to be allocatively efficient, possibly resulting in poor coverage for certain risk bands. The other source of security, the link between employment and a subsidised premium, gives rise to two sources for economic inefficiency. First, a phenomenon known as job lock occurs, where EPHI leads to labour immobility. Monheit and Cooper (1994) report that there is typically a 20%40% loss in mobility for a small proportion of workers due to job lock, meaning that whilst it likely exists, its magnitude is unlikely to be large. Second, Olson (2002) notes that both theory and empirical evidence show that the inclusion of an EPHI plan induces the worker to accept a wage lower than their marginal productivity and finds that those with health insurance coverage accept, on average, pay that is 20% below those who have no such coverage. The combination of these factors would suggest that EPHI is a potential source for net inefficiency. Focusing on the flow of funds and services between insurance firms and patients we observe two further channels: the first is the traditional indemnity insurance plan which the analysis of the previous chapter assumed whilst the second is a more recent channel that has been termed managed care. Reminiscent of the split between EPHI and directly purchased insurance, managed care dominates. Jensen et al. (1997)
12

200<employees. 149 employees.

13

28

report a surge in managed care enrolment during the mid1990s, where 72.6% of insured working Americans were in a managed care plan. However, there has been a decline in recent years while a more recent study by Draper et al. (2004) finds enrolment in health maintenance organisations (HMO), the dominant managed care organisation (MCO) model, to have peaked in the late 90s at around 77 million insureds this has gradually fallen to around 75 million in 2002. Once again, we must ask two questions: what is managed care, and why was it suddenly so popular before falling out of favour? Are there any economic benefits to such a model? Managed care14 is an attempt to confront the problem of moral hazard in the healthcare system through strict controls on both the demand and supply side. In particular, the third-party payer problem, previously discussed, creates a significant problem for healthcare costs; figure 4 shows that the US spends far beyond the OECD average and that this divergence is increasing. Moral hazard is not the sole contributor to such large expenditures, however theory suggests that moral hazard will contribute to Aaron and Ginsburgs (2009) finding that healthcare spending is likely to be excessive. Managed care plans were highly popular due to the lower premiums they were able to offer as a result of the strict controls in place; Jensen et al. (1997) note that not only were the premiums smaller than conventional plans in absolute terms, they also

14

MCOs attempt to merge the insurer, the physician and the hospital. The classic MCO is the HMO, where a group of physicians are the insurance firm; the largest rival structure is the Preferred Provider Organisation (PPO) which drives patients towards selected physicians in return for a lower FFS.

29

had a slower rate of growth. In addition, figure 4 shows a stabilisation in US healthcare expenditures during the peak of managed cares popularity. On the demand side, MCOs use a mix of incentives and control measures. As with traditional insurers, coinsurance is the primary incentive measure used. They also attempt to directly control patient choice by forcing patients to get a second opinion15 or see a primary care physician (gatekeeper) before being admitted to a specialist; the problem with the former method is that second opinion physicians are likely to agree with their colleagues due to the likelihood that they will ask their colleague for a second opinion in the future. Recalling that Chiappori et al. (1998) find that transport and transaction costs are a significant portion of the total cost of healthcare receipt, the latter measure may instigate under consumption of healthcare as it may raise the MPC above MSC. On the supply side, the main incentive through which physicians are reigned in is through the payment mechanism. FFS and salary payment mechanisms have already been compared prior, however a key feature of HMOs is a prospective payment system of capitation where the physician is allocated a fixed budget per patient. The physician will now restrict the provision of healthcare to prevent going overbudget; Kerr et al. (1995) finds that physicians respond to this constraint by increasing use of control measures such as gatekeepers, although PPOs still have incentives to overprovide due to remuneration being volumebased.

15

In addition to managing moral hazard, this also mitigates the effects of supplier induced demand (SID) as the second opinion physician incurs no monetary gain/loss from their action.

30

In net, a study by Miller and Luft (1994) finds that HMO plans have lead to greater use of preventative care, shorter inpatient stays and less consumption of expensive treatments and procedures, suggesting that managed care plans have intrinsic economic benefits in mitigating moral hazard and SID. However, the recent decline in the popularity of managed care is, in part, due to the controls which make them economically beneficial; Lesser et al. (2003) finds that in a competitive labour market, good economic conditions meant employers whose health insurance plans offered more choice were at an advantage to employers who focused on managed care plans. Due to this inferior good nature of managed care plans it is necessary to find a mechanism whose efficacy is invariant in economic conditions or the labour market. There are two sources for the direct flow of funds from the patient to the physician/physician practice: coinsurance payments and out of pocket (OOP) payments faced by the uninsured. The measures that have been discussed thus far in this chapter will have little impact on those without insurance, because the dilemma that uninsured individuals face is not necessarily a financial one if the problem were that individuals simply could not afford health insurance then it would be economically optimal for the Government to introduce a lumpsum transfer that allowed everyone to purchase insurance; this is the second optimality theorem that would still allow for a Pareto optimal allocation of resources. Rather, it is the problem of adverse selection that prevents individuals from purchasing insurance. Data from a recent US Census Bureau report (2009) lends support to this hypothesis it shows that insurance coverage
31

substantially rises with income16 and that insurance is higher amongst employed persons17 (who, due to employer provided care, are partially protected from adverse selection). Figure 7 shows that the uninsured population, as a percentage of the total population, is economically acyclical, suggesting it is a structural rather than transitory problem.

Figure 7: Number and Percentage of US Population without Health Care Insurance Over Time, Source: CMS NHE Data

This is not to suggest that the cost of healthcare is insubstantial for low income families, indeed Iglehart (1999) notes that Medicare beneficiaries with incomes below the federal poverty level, and without Medicaid benefits, spent half their income on OOP healthcare costs.

16

The percentage of low income households (under $25,000) that are uninsured is 26.6%; for high income households (greater than $75,000) this is 9.1%.
17

The percentage of those with stable employment who are uninsured is 15.2%; the percentage of non workers who are uninsured is 29.1%.

32

Concerning costs, figure 8 shows that the problem lies in private health insurance as opposed to OOP payments, however it should be noted that the uninsured impose substantial18 costs on the Government via uncompensated care costs where physicians are reimbursed for care provided to fiscally insolvent patients. A report by the CEA (2009) notes that healthcare costs are being driven, in addition to the factors prior mentioned, by poor costeffectiveness in the use of new technologies and high administration costs. The former problem arises in part due to physician practices competing on the basis of technology as opposed to price (due to the lack of traditional free market incentives) and is also a result of SID; new technologies and treatments could mean greater healthcare demand. Administration costs are high due to the fragmented system of multiple insurers and providers, and due to the need to estimate risk and set premiums accordingly; a study by Woolhandler et al. (2003) finds that administration costs in 1999 accounted for 31.0% of US healthcare expenditures19.

18

A report by the Executive Office of the President Council of Economic Advisers (CEA) (2009) notes that uncompensated care costs borne by the Government in 2008 were $42.9 billion.
19

This is likely an overestimate, as administration costs for Public Health, Retail Pharmacy Sales, Research, Medical Equipment and Supplies and Construction were unavailable for the study. 33

Figure 8: Relative Growth of Health Insurance Expenditures. Source: CMS NHE Data

Reforms targeted at the private sector of the US healthcare system should therefore concentrate on three main areas.
i.

Addressing adverse selection in the directly purchased insurance channel, but also with an eye to EPHI.

ii.

Confronting moral hazard without restricting choice to unacceptable levels where individuals switch out of such plans in good economic times.

iii.

Incentivizing small business owners to provide a range of choice in insurance plans to partially improve the efficiency of EPHI.

34

In achieving these goals, they must provide incentives to encourage the use of costeffective treatments and technologies in addition to

decreasing fragmentation and administration costs. III.ii.ii The Public Sector Although this paper is focused on the private sector, due to the ACA reforms being primarily concerned with the failings of that sector, the public sector is, as shown by figure 3, a substantial and growing part of the US healthcare system. The two largest components of the public sector, Medicare and Medicaid, are themselves a response to two of the most catastrophic failures of the private sector: Medicare is a response to the inability of the elderly to purchase insurance due to adverse selection, as noted by Akerlof (1970), whilst Medicaid insures some, but not all, of low income Americans unable to purchase insurance.

Figure 9: Medicare Enrolees and Medicaid Beneficiaries as a Percentage of the Total Population. Source: CMS NHE Data 35

Figures 3 and 9 exhibit the main problem facing the public sector: figure 3 shows an ever increasing proportion of total healthcare expenditures were being accounted for by the Public sector whilst figure 9 reveals that, whilst Medicaid has a seemingly cyclical pattern due to its relationship with low income households, both Medicare and Medicaid coverage have remained relatively stable over time. Further, figure 3 shows that it is the Federal portion of the public sector that exhibited the largest proportional increase; the federal budget is primarily used to fund Medicare 20, implying that Medicare has the most significant cost issues of the two programmes a problem exacerbated when it is noted that Medicare has the most stable coverage of the two. For this reason we shall focus on Medicare; however, it is important to note that Medicaid does not cover all low income families, and the reforms do address this problem. Medicare provides coverage to the over65 population via three channels - based on a contributions record of mandatory taxes, the entire population receives hospital insurance21, whilst the voluntary

supplemental medical insurance22 is comprised of insurance against physician costs and insurance for prescription drugs. Whilst the plan does go a substantial way to mitigate the problems of adverse selection for the elderly, by offering blanket coverage, it also inherits the cost problems

20

However, the Federal Government must also match at least 50% of a states Medicaid expenditure.
21 22

Medicare part A; the funds are kept separate in a Hospital Insurance fund. Medicare parts B (physician costs) and D (drug coverage). Part C is Medicare Advantage, where the usual benefits of Medicare are made available through private health plans (such as managed care plans); Part C does not offer any new content, rather it offers choice in the receipt of care. 36

that highrisk insurance leads to; as figure

10 shows, Medicare

expenditures have largely been above the rate of urban consumer inflation23.

Figure 10: Growth Rate of Medicare Expenditures Compared with CPI-U Inflation. Source for Medicare Expenditure: CMS NHE Data, Source for CPI-U: US Bureau of Labour Statistics.

Due, in part, to the substantial costs inherent in insuring the elderly, McClellan (2000) recognises that OOP spending is notably higher in Medicare than in private plans elsewhere, driven by large coinsurance payments; he considers this large coinsurance to be the result of historical circumstance, however our prior analysis would suggest that high coinsurance is inevitable due to the high expected cost of the elderly populations healthcare, and the potential for moral hazard. Medicare also has gaps in coverage; Medicare does not cover long-term nursing home

23

As accurate data for this period is scarce we use urban-CPI inflation as a second best measure relative to pure health care inflation.

37

care and, prior to the introduction of part D in 2006, there was no remittance for outpatient prescription drugs. In response to the coverage gaps, McClellan (2000) notes that 87% of Medicare enrolees have supplementary medigap private insurance plans. There was an attempt in 198824 to move away from the extensive coinsurance; however it was repealed in 1989. Rice et al. (1990) find that popular resistance was due to fears over the role of the state (equity) and dissatisfaction with a new supplementary tax on the elderly which would allow for the scheme to be funded purely by its beneficiaries. The failure of Medicare reform provides lessons for the ACA; primarily, changes to the tax system are likely to be resisted, and Americans give a high weighting to their notion of equity. Medicare reform is, however, inevitable there are serious cost sustainability issues due to the presence of moral hazard and high cost of the elderly, although there has been a degree of success with small reforms by switching Medicareparticipating physicians and hospitals to a prospective payment system, as noted by Phelps (2003).

24

Under the Medicare Catastrophic Care Act (MCCA) (1988).

38

IV. The Affordable Care Act

This papers contention has been that the major inefficiencies of a private healthcare system arise from two main sources: Information imperfections and externalities. With particular reference to the US case it is the information asymmetries that create the largest problems. With this in mind, this chapter shall analyse the reforms in terms of how they address adverse selection, moral hazard, imperfect knowledge and externalities; due to the evident lack of empirical data, the analysis shall be broadly theoretical. Excellent overviews of the ACA provisions are provided by: The Kaiser Family Foundation25, US Government26 and Harrington (2010). IV.i The ACA and Adverse Selection Concerning young adults, the US Census Bureau (2009) finds that 30.4% of Americans aged 1824 were uninsured, the highest of any age group. Whilst not necessarily a result of adverse selection, it is a gap in coverage fostered by a transient period in young Americans lives, leaving them without their parents health insurance coverage or a form of EPHI, as noted by Collins and Nicholson (2010). This has a detrimental impact on the labour market; the authors find that 31% of young Americans delayed their further education whilst paying off medical debt. The ACA solves this problem by allowing young adults of up to 26 years to stay on their parents insurance plans. In addition, it extends coverage to the very poor

25

http://healthreform.kff.org/timeline.aspx

http://www.healthcare.gov/law/timeline/index.html 39
26

by automatic enrolment in Medicaid for individuals with incomes under 133% of the federal poverty level, and reduces the coinsurance levels for Medicare enrolees27; pertinently, it reduces the non-first-dollar deductible donut hole in Medicare part D drug coverage by offering tax rebates and reducing coinsurance from 100% to 25% by 2020. Finally, it enforces a minimum level of coverage by mandating that any insurance plan offered through a health insurance exchange28 (HIE) offers a Government-defined bundle of essential health benefits (EHB). With respect to small businesses29, there will be tax credits for those offering health insurance plans of up to 35% of employer costs in phase 130, increasing to 50% in phase 2. However, whilst Claxton et al. (2007) found that small businesses were less likely to provide any EPHI31, the reform does not necessarily induce the employer to address the lack of choice in health insurance plans that was noted in chapter III. The phase 2 tax credit is contingent on the insurance being purchased through the HIE, however, which may implicitly help the range of choice due to potentially greater competition in the exchange. A number of provisions deal with high-risk coverage. First, bans on lifetime limits on coverage, a tool used to implicitly deny coverage to those with chronic conditions, and on rescinding coverage of an individual

27

Although not to the same degree as the MCCA.

28 29

HIEs are discussed in IV.iii Defined in the ACA as having under 25 employees with an average wage of less than $50,000. 30 The ACA is being implemented in two phases: phase 1 is in effect from 2010 2013 inclusive, phase 2 is 2014 onward. 31 45% of small (3-9 employees) businesses contrasted with 99% of large (200+). 40

come into effect in phase 1. In addition, a ban on denying coverage to both children and adults with pre-existing conditions will come into effect in phase 1 for children and phase 2 for adults. Adults in phase 1 will be able to enrol in a federal-funded high-risk pool which charges premiums equivalent to that for a normal risk population and covers at least 65% of costs with limits on OOP expenditure. The high-risk pool is, by design, temporary and that is reflected in its unsustainable structure; the divorce of premiums and risk mean that the Government has set aside $5 billion to cover administration costs and excess payments. Whilst high-risk pools are currently operated in some states, they are small in magnitude: the Kaiser Family Foundation (2010) estimates the coverage extends to around 200,000 Americans. Coverage of individuals with pre-existing and chronic conditions is a key goal for US reform, due to it being one of the most visible flaws of the preACA system. However, the Phase 1 solution of high-risk pools is not a sustainable one. The crucial component of the ACA, which allows for the sustainability of non-exclusion of high-risks, is the introduction of a mandate for individuals and medium-sized32 businesses in 2014.

Individuals without coverage and employers who have employees on individually purchased insurance plans will pay fines. The mandate could lead to lower premiums in two ways. First, there is greater risk spreading due to a larger pool of insureds. However, equation 733 provides a caveat:

32

50+ employees.

41

(N NH ) N Premium = (1 + ) H PH + PL L N N N ( P PL ) + NPL = (1 + ) H H L N N = (1 + ) H ( PH PL ) + PL L N
Equation 2: Insurance under Community Rating with Mandatory Insurance.

It is shown that the premium is determined by the ratio of the high risk

population to total population:

NH . Therefore, whether premiums are N

cheaper or not will depend, inter alia, on whether the number of higher risk persons that can now access coverage is smaller than the number of low risk persons induced to enrol. Also important is that mandatory insurance is not a commitment technology for any single plan. The market for insurance will, if anything, become more competitive due to the HIE, meaning individuals could still drop out of plans, altering the high risk ratio. Whilst the ACA provides incentives for greater EPHI, which does act as a commitment technology to single plans, the Congressional Budget Office (CBO) estimates that there will be 3 million less Americans covered by EPHI in 2019 due to the incentives of subsidies for purchasing insurance through the exchange. The role of EPHI must also be reexamined, with adverse selection being confronted by the mandate and EPHI propagating labour market inefficiencies, it is questionable whether EPHI should receive continued support through tax incentives. Another
33

Assuming that: the loss L is equal for both groups, there are only two risk groups in a population N and the probability of a high risk becoming ill is greater than that of a low risk PH>PL.

42

channel for lower premiums is the realisation of economies of scale due to a greater N: this would lower, currently substantial, administration costs, manifested via a diminished loading factor (1 +) . There could be a period of calibrating the penalties and enforcement methods, however, in which the benefits outlined above will not be fully realised. Glied et al. (2007) notes that mandatory insurance legislation varies substantially in its efficacy, with conformance rates between 30%99%, attributing the variance to differences in the level of the penalty, the affordability of insurance and the probability of an uninsured person being charged. The US may experiment with different enforcement methods and penalties before finding an optimal infrastructure. IV.ii The ACA and Moral Hazard It would appear that in its attempt to confront adverse selection, the ACA has not dealt with moral hazard in a systematic and comprehensive manner. Recall that the third-party payer problem is inevitable given the US healthcare structure, and that the private sector currently uses coinsurance to limit it; further, recall the failure of MCOs and the need for an alternative organisational structure. The ACA, if anything, exacerbates the third-party payer problem. To increase coverage, it provides subsidies for coinsurance payments, bans annual caps on coverage in phase 2 and bans coinsurance altogether for preventative care. Whilst these extend coverage, and the latter addresses moral hazard leading to under-consumption of preventative goods, it also denies the private market of a crucial tool of raising the MPC of treatment,
43

reducing excessive consumption. In addition, there are no concrete alternatives to MCOs in the plan, meaning the reform has failed to confront two large areas of moral hazard, with the exception of Medicaid now fully funding visits to primary care physicians and Medicare testing new systems of physician payment. For example, bundled payments are to be introduced whereby a hospital is, ex-ante, given a set payment for all the costs incurred in treating a patient (in-patient and out-patient services included) this is, in essence, a form of capitation. As Cutler (2010) notes, the hope is that FFS will no longer be the dominant payment scheme in Medicare. However there are attempts at confronting other areas of moral hazard, although their impact is limited as they are largely confined to Medicare/Medicaid. To encourage consumption of preventative care, states will be afforded grants to encourage Medicaid enrolees to use preventative care, and subsidies will be provided for employers to cover up to a further 30% of an employees healthcare costs, contingent on attaining health-related goals. Reforms also aim to reduce spending on inefficient technologies and procedures, generating excessive demand, by establishing an excise tax of 2.3% on the sale of any taxable medical device, linking Medicare payments to hospital quality, and creating the Patient-Centred Outcomes Research Institute and Centre for Medicare and Medicaid Innovation, which shall evaluate new technologies and delivery systems and provide recommendations.

44

It should be noted that the CBO (2010) estimated that the ACA would, by 2019, reduce the federal deficit by $143bn, meeting the Presidents requirement that the ACA ought to be at least budget neutral. If the reforms do not tackle moral hazard in a comprehensive manner, where do the cost savings arise? They arise largely from Medicare and Medicaid34 and account for $511bn of reduced expenditures in 2010-2019. The remainder, $420bn, accrues through new taxes prominently, $210bn35 will be raised through a higher tax on Medicare enrolees who earn above $200,000 36. $60bn is to be raised through placing a progressively rising annual fee on insurance firms. Insurance firms also have their ability to increase premiums limited, fees imposed, and taxes on those who provide high cost insurance; this would suggest that there will be a drastic cut in the loading factor, particularly due to administration costs, as firms attempt to reduce Xinefficiency in order to remain competitive within the HIE. Indirectly, therefore, the reforms may provide incentives for the creation of a rival organisational structure to the MCO. In net, however, there is space for further reforms to confront the third-party payer problem directly, as opposed to relying on high tax rates to control the budget. The concern over cost sustainability (which stems from problems of moral hazard) has been noted in the, still infant, literature on the reforms; Harrington (2010) and Doherty (2010) both conclude that while there is considerable
34

E.g., $156.6bn is to be saved by reducing annual market basket updates for Medicare providers and $38.1bn to be saved from changes to Medicaid prescription drug coverage.
35 36

Estimated by the Joint Committee on Taxation (2010). This may escape the public opposition that the MCCA (1988) received due to the tax base 45 being only a section of the Medicare beneficiary population.

uncertainty associated with long-run spending forecasts, it is likely that costs will continue to be problematic whilst payment system trials are being evaluated, although Baicker and Skinner (2011) find evidence that healthcare spending growth in OECD countries is curtailed by higher taxes, suggesting the ACA taxes may have a similar (although smaller, due to being relatively limited in scale) effect. IV.iii The ACA: Imperfect Knowledge and Externalities Mechanic (1989) notes that whilst individuals make rational choices about their choice of health insurance plan, there are a number of factors which may encourage a sub-optimal plan being adopted, leading to allocative efficiency losses. Namely, the fact that insurance companies compete on a wide spectrum of variables with few services available to help consumers compare plans. Thus, a consumer may have a well defined preference ordering, but matching preferences to variables is not straightforward. HIEs are expected to be installed in every state, with a maximum of one HIE per geographic area if there multiple in a state. They address the information processing problem primarily by standardising insurance plans; each plan offered through the HIE has to include the package of EHB, with plans being ranked by what proportion of the actuarial cost of the EHB they cover37. Additionally, HIE will be able to provide clear comparisons of plans by standardising variables and will use uniform enrolment forms for plans. Whilst this may lead to allocative efficiency

37

Four levels - bronze covers 60% of the cost, through to platinum which covers 90%. All insurers selling via a HIE must provide at least one gold and silver plan.

46

gains

and

efficiency

gains

from

increased

market

competition,

standardisation could incentivize sub-optimal allocation of resources to maximise performance in the variables which are used for comparison. Consequently, the net effect will depend on the precise specification of the EHB and the variables covered. Additionally, by limiting HIEs to one per geographic area and linking HIEs, costs incurred by fragmented markets are reduced. Regarding externalities, Orenstein and Hinman (1999) find that the US uses laws stipulating access to schools being contingent on the consumption of certain vaccinations have been decidedly effective: they find that most coverage rates exceed 80% and have been associated with a decline in the diseases they protect against. However, they also note that such laws are ineffective for consumption of vaccines for those below school age (up to two years old). The ACA eliminates coinsurance for certain vaccinations, and while this is unlikely to be as strong an incentive as the school laws, it could lead to a significant rise in coverage rates if vaccine demand is sufficiently price elastic, or the fall in the marginal cost is substantial. Further, the potential for moral hazard is limited as the full marginal benefit of a vaccine is usually exhausted after one course of the vaccinations.

V. Alternative Healthcare Systems

47

With the ACA, the US healthcare system now moves into an era of near universal coverage due to the individual mandate. Universal coverage already exists amongst most of the OECD, with the average percentage of the total population falling under public health insurance schemes among 22 comparable countries being 98.4% (Docteur and Oxley, 2003). However, the US healthcare system remains anomalous in the manner in which it attains universal coverage; only Switzerland has a comparable system of mandated private insurance. Three questions are of immediate concern in the context of reforming US healthcare: how do these systems operate? How do their outcomes compare to the US? What lessons can be learnt from their experiences? This chapter shall consider these three questions in turn, with particular focus on cost containment and moral hazard. V.i Alternative Methods of Finance and Provision There are two defining features of any healthcare system: how that system is financed, and how care is delivered. In the US system, the private sector that predominantly finances and delivers healthcare; however in the rest of the developed and developing38 World, it is public expenditures that dominate the finance of care. Regarding provision, a number of systems either use public provision or (more commonly) a mix of public financing and private provision. An important point to make is that whilst Wagstaff (2009) is correct in asserting that the method of finance does not restrict a system to a certain method of provision, some

38

Wagstaff (2009) notes the emerging popularity of SHI systems in developing countries.

48

form of regulated private provision is generally observed. Even in the publicly provided British National Health Service there are quasi-markets to encourage competition, because adverse selection arises between the population and the insurance firm (or analogous organisation), and moral hazard predominately arises due to incentives embedded in the method of financing, rather than the method of provision. For this reason, discussion is focused on financing methods. There are two dominant models of healthcare financing that operate in developed nations: Full-Tax Financing (FTF) and Social Health Insurance (SHI). Whilst precise details vary by system, the former finances healthcare by drawing from a pool of general tax revenue, similar to how Medicaid is financed, and the latter finances healthcare by way of mandatory contributions, similar to Medicare contributions, deducted from wage earnings, being paid into a sickness fund analogous to an insurance firm bound by strict regulation. Before the specifics of each system are debated, it is crucial to note that whilst there may be analogues of the US system, both FTF and SHI fundamentally differ from a private finance system in that the premium (i.e., the tax that is paid) is based on ability to pay, not risk. This has substantial implications for redistribution, as found by Wagstaff et al. (2000). They find that, contrary to the US and Switzerlands highly regressive systems, FTF systems tended to be either proportional or slightly progressive39 whilst SHI systems tended to be mildly regressive, although France was proportional suggesting that such a system does
39

The UK had a Kakawani index of 0.051, the highest in the sample.

49

not necessarily prevent redistributional goals. However, because the US does not give a high weighting to redistributive concerns, such findings are unlikely to influence any further reforms; what are the efficiency arguments for either system? Concerning adverse selection, both systems can be used to provide universal coverage, although the existence of a contributions record in SHI systems can be used to exclude non-contributors from care. Fortunately, this is not the case in West European SHI systems. However, van de Ven et al. (2007) find significant incentives for a sickness fund to risk select amongst a selection of West European SHI systems as each fund is liable for the costs of an insured and is unable to ask for higher contributions or turn down an applicant based on risk. SHI systems use a system of exante and ex-post risk adjustment transfers to redistribute from a fund with a lower expected cost to a fund with a higher expected cost, in order to mute any potential gains from such selection. However, the possibility of imperfect transfers mitigates their efficacy, and funds were observed to be selecting through avenues such as selective advertising. Adverse selection is of no tangible problem in FTF systems as the entire population pays into, and draws from, the same pool of funds. Solving imperfect transfers by making them all ex-post is not viable as sickness funds will be fully reimbursed for all expenses, and the thirdparty payer problem becomes prominent. This moral hazard is part of the reason why van de Ven et al. (2007) found two of the four countries studied reducing the level of ex-post transfers. Both FTF and SHI systems
50

face moral hazard problems, and both use co-insurance as an attempt at mitigating it, as can be seen in table 1. Also note that, whilst Switzerland drastically increases the standard deviation of SHI OOP payments, both systems are compatible with a range of equity principles. Additionally, a factor contributing to higher mean OOP payments in a FTF system could be greater use of supplementary private insurance by the wealthy to achieve higher care - no such incentive exists in SHI systems due to the competition between funds providing a range of choices for the patient. FTF also use spending caps, which are effectively a form of capitation, to curb moral hazard; in Canada each provinces Medicare system has a set budget for the year. OOP Expenditures as % of Total Healthcare Expenditures
199 8 15.4 7.1 11.0 7.6 8.4 32.9 13.7 9.9 199 9 15.2 7.2 10.9 7.4 9.0 33.4 13.9 10.0 200 0 15.3 7.1 11.1 7.0 9.0 33.0 13.8 9.9 200 1 16.0 7.2 11.2 6.5 8.7 31.8 13.6 9.6 200 2 16.2 7.0 11.4 6.9 8.0 31.6 13.5 9.5 200 3 16.5 6.8 11.7 6.9 7.3 31.6 13.5 9.7 200 4 16.0 6.7 13.2 6.6 7.2 31.9 13.6 9.8 200 5 15.7 6.8 13.2 6.5 7.1 30.6 13.3 9.3 2006 15.8 7.0 13.4 6.5 5.6 30.8 13.2 9.6

Year Country

Social Health Insurance Finance

Belgium France Germany Luxembourg Netherlands Switzerland Mean % S.D

UK Canada

14.1 16.2

13.6 16.2

13.4 15.9

13.4 15.2

13.2 15.2

12.8 14.5

12.3 14.6

11.8 14.6

11.4 14.9

51

Full Tax Finance

Spain NZ Denmark Mean % S.D

23.2 16.3 16.6 17.3 3.5

23.3 15.9 16.1 17.0 3.7

23.6 15.4 16.0 16.9 3.9

23.9 17.0 15.9 17.1 4.0

23.7 16.1 15.8 16.8 4.0

22.7 N/A 14.6 16.2 4.4

22.6 17.0 14.7 16.2 3.9

22.4 16.8 14.8 16.1 4.0

21.4 16.6 14.3 15.7 3.7

Table 1: OOP Spending as % of Total Healthcare Expenditures in Selected FTF and SHI Systems. Source: OECD (2010)

With regards to technical and allocative efficiency, Mossialos and Dixon (2002) note that there is no direct link between the financing method and attainment of either aspects of efficiency, although SHI is likely to be more responsive to consumer demand due to competition between multiple sickness funds. Regarding effects on the wider economy, Ballard and Goddeeris (1999) use a general equilibrium model and examine the effect of introducing a FTF system in America they find substantial ($27bn) efficiency losses and a 2.5% fall in the labour supply due to a 7.5% rise in tax rates in order to fund the system, however the model does not take into account efficiency gains from lower administration costs, greater risk pools and less underinsurance, so the losses are likely to be exaggerated. In addition, Mossialos and Dixon (2002) note that it is theoretically possible for a SHI system to lead to labour immobility due to contributions being deducted from wages however, job-lock is small even in the US40 and they find no substantive evidence of it in Germany. Cost-containment in FTF systems is, as noted by Nonneman and van Doorslaer (1994), primarily a political issue; the Government ultimately decides the portion of general tax revenues to be allocated to healthcare.
40

With a smaller safety net than Western European systems.

52

In addition, a single purchaser of private care has the advantages of economies of scale to reduce administration costs, and monopsony power in the market place that they can use to acquire lower cost care. However, the budget constraint in a FTF system is soft if expenses exceed the cap, more resources can be drawn. Additionally, the optimal allocation of funds to healthcare is highly contentious and subsequently, revenues may fluctuate noticeably due to what Nonneman and van Doorslaer (1994) call the whimsical nature of Governments however, whilst they find a larger standard deviation for average annual revenue from Government than Non-Government sources, their study is limited to Belgium. Cost containment in SHI systems parallels our discussion of cost containment in the US system, as the tools used are analogues. Hofmarcher and Durand-Zaleski (2004) note that sickness funds use selective contracting of private providers of care, with payment schemes being a mix of FFS, capitation and salary, with the latter two increasing as a proportion of total payment schemes. Wagstaff (2009) notes growing attempts at managing care, in a similar vein to the US MCOs. However, such efforts in a SHI system are ultimately diluted due to the lack of autonomy sickness funds have in specifying their contracts; Hofmarcher and Durand-Zaleski (2004) note that, whilst autonomy is increasing in some systems, the ultimate financing decision rests with the Government. Indeed, such limits in contract specification reveal themselves in expenditure data; Mossialos and Dixon (2002) find that average

healthcare expenditures are lower in FTF systems than in SHI systems.

53

V.ii International Quality Comparisons Docteur and Berenson (2009) caution that quality comparisons are inherently difficult to make reliable and accurate due to factors such as fragmented data, heterogeneous collection methods and a large number of metrics and variables. That said, there have been numerous attempts at international comparisons amongst select developed countries, and we shall focus on two in particular: a contemporary study by The

Commonwealth Fund (TCF) (2010), and a comprehensive, but dated, study by the World Health Organisation (WHO) (2000).

Metric

Health Status Level (DALE)

Attainment Goals Responsive ness (Distributio Responsive ness (Level) Distributio n

of

Performance (Level of Health) Per Capita Health Expenditure Vertical Equity

Total Attainment

Total Performance

Country

54

USA

24 16 3 22 13 8

32 26 12 20 15 10

1 16 17 16 17 5 9 2 -

3 38 3 38 3 38 3 38 3 38 3 38

54 55

1 15 4 3

72 28 4 41 19 26

15 13 6 14 8 2

37 21 1 25 17 20

Full Tax Finance Social Health Insurance Finance

Belgium France Germany Netherlands Switzerland

3-5 26 29 6-7 20 22 38 40 -

9 2

UK Canada NZ Denmark

14 12 31 28

2 18 16 21

26 27 7-8 22 23 4

3 38 3 38

8 11 17 19 23 25

26 10 20 8

24 35 80 65

9 7 26 20

18 30 41 34

3 38 3 38

3-5

Table 2: Ranking of Selected Countries over Nine Metrics. Source: World Health Report, WHO (2000).

TCF (2010) examined seven countries41 that have a mix of financing and delivery systems, and ranked them according to: the quality of care provided, access to care, technical efficiency and horizontal equity. Surprisingly, for a system whose participants largely rank as

excellent/good42, the US is last place for the aggregate ranking and all metrics apart from overall quality, where it is second to last. It comes

41

Australia, Canada, Germany, New Zealand, UK, Netherlands, US.

42

A Gallup poll released on November 19th 2010 shows 62% of Americans rate the quality of the systems output as excellent/good. 55

fourth (its highest ranking) in effective care, which is based on preventative and chronic care for reasons we have discussed, this a relatively neglected sector of the US healthcare system and in patientcentred care. What is a surprise is that the US does not rank higher on the latter metric; one argument in favour of the system is that due to the competition of the free market, insurers will be more willing to adapt resources to match an individuals preferences. However, such a mediocre ranking would suggest that the distortive effects of market failures are sufficiently strong to mitigate that responsiveness. The effect of FTF on responsiveness is ambiguous New Zealand ranks first, however Canada and the UK rank sixth and seventh respectively. Whilst TCF (2010) provides an accessible ranking, an altogether more comprehensive study was undertaken by the WHO (2000) which used eight metrics over 191 countries. Critically, in using disability-adjusted life expectancy43 it provides a ranking of health status amongst the countries; as noted at the start of chapter II, the ultimate aim of a healthcare system is to maximise a nations health. Table 2 collects the findings for a selection of countries with a mix of FTF, SHI and provision mechanisms. The study reveals a number of interesting findings about the US system. In international dollars the US spends more on healthcare per capita than anywhere else; however, this expenditure does not translate into the highest quality of care. Indeed, the study differentiates between attainment and performance; the former is an absolute measure of quality, whilst the latter is a measure of cost-efficiency. What can be
Life expectancy net of years with disability (life expectancy in full health). 56
43

observed is that, whilst the US does fairly well in terms of total attainment44, it performs poorly on total performance, with a large differential between the two rankings. Where the system does excel, in contrast to TCFs findings, is in responsiveness it ranks highest in the World, with Switzerland (a system with substantial private involvement see OOP payments in table 1) coming second. Such a differential can be accounted for by the use of different metrics being examined via the studies respective surveys the WHO report focuses more on dignity and autonomy, whilst TCF focuses on the patientphysician relationship. Whilst a SHI or FTF system does not guarantee an improvement in either attainment or performance (although France provides an example of how a SHI system can excel in such metrics), theyre shown to have a smaller disparity between total attainment and total performance than the US, with France obtaining a higher ranking for performance than attainment; this suggest that whilst the complex organisational arrangements in the US are able to mitigate the effects of market failures in terms of outcomes, alternative structures are able to provide a comparable level of care at substantially lower cost by maximising the performance of their resources. V.iii Lessons for the US The US not only has much to learn from SHI and FTF systems, but already incorporates features of both systems in its own. Medicare part A broadly resembles a single-payer SHI system with a single sickness-fund, whilst Medicaid is essentially a FTF system. However, innovations such as risk44

Heavily penalised by a very poor ranking on vertical equity, as is to be expected.

57

adjusters and payment caps, although not without problems, can be adapted for use in the private sector as they involve only minimal Government intervention, which is congruous with the US systems libertarian ideology. While such measures may be weaker than in a pure SHI/FTF system, they would be combined with stronger MCOs and payment schemes than either alternative system is able to implement, due to insurance firms commanding greater autonomy than sickness funds. In part, this could foster a resurgence of MCOs as the risk adjustment transfers potentially allow more choice in the range of care provided. Importantly, both alternative systems can be reconciled with a level of care that is not only comparable with the US, but markedly more cost-efficient; however, as the evidence of New Zealand in table 2 shows, care must be taken to ensure that such cost-efficiency and attainment is realised. Further, a study by Gechert (2010) finds that administrative costs in a SHI system are lower than in a supplementary private health insurance system, suggesting SHI would curtail the substantial US administrative costs. The ACA appears to move the US system towards a Switzerland-style SHI system, with mandatory insurance for individuals, with a system of subsidies for low-income individuals, and insurance firms, who are unable to decline an applicant based on risk, offering a compulsory baseline package. However, the Swiss system goes further in partially divorcing premium and risk; an insurance firm cannot make a profit off the compulsory baseline package, and may only offer a community rating premium, not a premium based on individual risk. In addition, prices of
58

medical services are set once a year by a Government agency. Such crucial omissions will put more pressure on private sources of cost containment in the US system, although one may expect a convergence towards Switzerlands performance level.

VI. Conclusion
The US healthcare system is going through a period of extraordinary change. In applying a theoretical framework constructed in chapter II we were able to analyse the system, finding substantial problems in terms of imperfect information and cost-containment. We also analysed the responses to such market failures, such as MCOs and EPHI, concluding that such measures were left wanting in some regard. However, we also found a system that remained largely true to the nations libertarian aspirations for such a system and reigned in Government involvement. Still, it was clear that a major reform of the system was necessary to curtail a drastic rate of growth of expenditures, not comparable to the OECD average, and a high, structural level of uninsured any such reform would need to tackle the core problems of adverse selection and moral hazard to be fully successful. The ACA, ultimately, does not fully confront all these problems; it does, however, make considerable progress towards such an aim. The crucial elements of the reform package are the mandate and Medicaid expansion, which mitigates the worst effects of adverse selection and financially-related access to care problems. On moral hazard, the picture is mixed; under-consumption of preventative care is directly addressed, and there are attempts at confronting the third-party
59

payer problem, but they are largely trials and limited to operating in the public sector. The HIEs and EHB provide an innovative solution to imperfect knowledge and are an interesting attempt at cost containment through competition, although whether they will bend the cost curve to be sustainable is left to be seen. Is the ACA sustainable? It is unclear; although adverse selection is mitigated, it is left to be seen how insurance firms will cope with drastically more stringent rules and fees, in addition to the greater competition HIEs are expected to foster. A further move to a Switzerland SHI system would not be impossible, nor would a move to a Canadian FTF system, although the ACA implies the US is heading towards the former. Indeed, post-ACA, the US system occupies a middle-ground between pure private and SHI, reflecting a statement by Barr (2004: 285, emphasis in original): ... any strategy for addressing the problems of actuarial medical insurance leads inescapably to institutions with the major characteristics of social insurance. A move toward a fuller SHI system is ultimately a political, rather than economic decision, and that was reflected in an exchange between President Barack Obama and an audience member at a rally in Iowa after the legislation had passed. The member appears to ask why a public option45, which Cutler (2010) notes could be the basis for a full SHI system

45

A Government-sponsored plan.

60

in a further reform, hadnt been included in the bill to which the President replied: Because, we couldnt get it through Congress.

Bibliography

Aaron, H. J. and Ginsburg, P. B. (2009) Is Health Spending Excessive? If so, What Can We Do About It?, Health Affairs, vol. 28, pp. 1260 1275. Akerlof, G. A. (1970) The Market for Lemons: Quality Uncertainty and the Market Mechanism, The Quarterly Journal of Economics, vol. 84, pp. 488 500. Anderson, G. F. and Poullier, J-P. (1999) Health Spending, Access, And Outcomes: Trends In Industrialized Countries, Health Affairs, vol. 18, pp. 178 192. Arrow, K. J. (1963) Uncertainty and the Welfare Economics of Medical Care, The American Economic Review, vol. 53, pp. 941 973. Baicker, K. and Skinner, J.S. (2011) Health Care Spending Growth and the Future of US Tax Rates, NBER Working Paper Series, vol. w16772, pp. . Accessed from: http://papers.ssrn.com/sol3/papers.cfm? abstract_id=1759844

61

Ballard, C. L. and Goddeeris, J. H. (1999) Financing Universal Health Care in the United States: A General Equilibrium Analysis of Efficiency and Distributional Effects, National Tax Journal, vol. 52, pp. 31 51. Barr, N. A. (2004) The Economics of the Welfare State, 4th ed, Oxford: Oxford University Press. Barros, P. P., Sanz-de-Galdeano, A. and Machado, M. P. (2008) Moral Hazard and the Demand for Health Services: A Matching Estimator Approach, Journal of Health Economics, vol. 27, pp. 1006 1025. Boulier, B. L., Tejwant, S. D. and Goldfarb, R. S. (2007) Vaccination Externalities, The B. E. Journal of Economic Analysis & Policy, vol. 7. Brito, D. L., Sheshinsky, E. and Intriligator, M. D. (1991) Externalities and Compulsory Vaccinations, Journal of Public Economics, vol. 45, pp. 69 90. Browne, M. J. (1992) Evidence of Adverse Selection in the Individual Health Insurance Market, The Journal of Risk and Insurance, vol. 59, pp. 13 33. Centers for Medicare & Medicaid Services, National Health Expenditure Web Tables. Accessed From: http://www.cms.gov/NationalHealthExpendData/downloads/tables.pdf Chiappori, P-A., Durand, F. and Geoffard, P-Y. (1998) Moral Hazard and the Demand for Physician Services: First Lessons from a French Natural Experiment, European Economic Review, vol. 42, pp. 499 511. Claxton, G., Gabel, J., DiJulio, B., Pickreign, J., Whitmore, H., Finder, B., Jacobs, P. and Hawkins, S. (2007) Health Benefits in 2007: Premium Increases Fall to an Eight-Year Low, While Offer Rates and Enrollment Remain Stable, Health Affairs, vol. 26, pp. 1407 1416. Collins, S. R. and Nicholson, J. L. (2010) Rite of Passage: Young Adults and the Affordable Care Act of 2010, The Commonwealth Fund. Accessed
62

From: http://www.commonwealthfund.org/~/media/Files/Publications/Issue %20Brief/2010/May/1404_Collins_rite_of_passage_2010_v3.pdf Congressional Budget Office (2010) Cost estimate for the amendment in the nature of a substitute for H.R. 4872. Accessed From: http://www.cbo.gov/ftpdocs/113xx/doc11379/AmendReconProp.pdf Coulson, N. E., Terza, J. V., Neslusan, C. A. and Stuart, B. C. (1995) Estimating the Moral Hazard Effect of Supplemental Medical Insurance in the Market for Prescription Drugs by the Elderly, The American Economic Review, vol. 85, pp. 122 126. Culyer, A. J. and Wagstaff, A. (1993) Equity and Equality in Health and Health Care, Journal of Health Economics, vol. 12, pp. 431 457. Cutler, D. M. (2010) The Simple Economics of Health Reform, The Economists Voice, vol. 7. Cutler, D. M. and Zeckhauser, R. J. (1998) Adverse Selection in Health Insurance, Forum for Health Economics & Policy, vol. 1. Davis, K., Schoen, C. and Stremikis, K. (2010) How the Performance of the US Health Care System Compares Internationally 2010 Update, The Commonwealth Fund. Accessed From: http://www.commonwealthfund.org/~/media/Files/Publications/Fund %20Report/2010/Jun/1400_Davis_Mirror_Mirror_on_the_wall_2010.pdf Docteur, E. and Berenson, R. A. (2009) How Does the Quality of Health Care Compare Internationally?, Timely Analysis of Immediate Health Policy issues, a Series of Briefs Produced by the Urban Institute and the Robert Wood Johnson Foundation. Accessed From: http://ssrn.com/abstract=1482677 Docteur, E. and Oxley, H. (2003) Health Care Systems: Lessons from the Reform Experience. Accessed From: http://www.oecd.org/dataoecd/5/53/22364122.pdf
63

Doherty, R. B. (2010) The Certitudes and Uncertainties of Health Care Reform, Annals of Internal Medicine, vol. 152, pp. 679 682. Draper, D. A., Hurley, R. E. and Short, A. C. (2004) Medicaid Managed Care: The Last Bastion of the HMO?, Health Affairs, vol. 23, pp. 155 167. Executive Office of the President Council of Economic Advisers (2009) The Economic Case for Health Care Reform. Accessed from: http://www.whitehouse.gov/assets/documents/CEA_Health_Care_Report.pd f Fronstin, P. (2010) Sources of Health Insurance and Characteristics of the Uninsured: Analysis of the March 2010 Current Population Survey, EBRI Issue Brief, no. 347. http://www.gallup.com/poll/126929/slim-margin-americans-supporthealthcare-bill-passage.aspx http://www.gallup.com/poll/144848/ratings-healthcare-quality-coveragebest-years.aspx Gechert, S. (2010) Supplementary Private Health Insurance in Selected Countries: Lessons for EU Governments, CESifo Economic Studies, vol. 56, pp. 444 464. Glied, S. A., Hartz, J. and Giorgi, G. (2007) Consider it Done? The Likely Efficacy of Mandates for Health Insurance, Health Affairs, vol. 26, pp. 1612 1621. Grossman, M. (1972) The Demand for Health: A Theoretical and Empirical Investigation, New York: NBER. Harrington, S. E. (2010) U.S. Health-Care Reform: The Patient Protection and Affordable Care Act, The Journal of Risk and Insurance, vol. 77, pp. 703 708. Hickson, G. B., Altemeier, W. A. and Perrin, J. M. (1987) Physician Reimbursement by Salary or Fee-for-Service: Effect on Physician Practice
64

Behaviour in a Randomized Prospective Study, Pediatrics, vol. 80, pp. 344 350. Hofmarcher, M. M. and Durand-Zaleski, I. (2004) Contracting and Paying Providers in Social Health Insurance Systems. In: Saltman, R. B., Busse, R. and Figueras, J., Editors (2004) Social Health Insurance Systems in Western Europe, Glasgow: Bell and Bain Ltd. Iglehart, J. K. (1999) The American Health Care System Expenditures, The New England Journal of Medicine, vol. 340, pp. 70 76. Jansen, V. A. A., Stollenwerk, N., Jensen, H. J., Ramsay, M. E., Edmunds, W. J., and Rhodes, C. J. (2003) Measles Outbreaks in a Population with Declining Vaccine Uptake, Science, vol. 301, p. 804. Jensen, G. A., Morrisey, M. A., Gaffney, S. and Liston, D. K. (1997) The New Dominance of Managed Care: Insurance Trends in the 1990s, Health Affairs, vol. 16, pp. 125 136. The Joint Committee on Taxation (2010) JCX 17 10. Accessed From: http://www.jct.gov/publications.html?func=startdown&id=3672 The Kaiser Family Foundation, Implementation Timeline. Accessed From: http://healthreform.kff.org/timeline.aspx The Kaiser Family Foundation (2010) State High-Risk Pools: An Overview. Accessed From: http://www.kff.org/uninsured/upload/8041.pdf The Kaiser Family Foundation and Health Research & Educational Trust (2010) Employer Health Benefits 2010 Annual Survey, Accessed from: http://ehbs.kff.org/pdf/2010/8085.pdf Kenkel, D. (1990) Consumer Health Information and the Demand for Medical Care, The Review of Economics and Statistics, vol. 72, pp. 587 595. Kerr, E. A., Mittman, B. S., Hays, R. D., Siu, A. L., Leake, B. and Brook, R. H. (1995) Managed Care and Capitation in California: How do Physicians at
65

Financial Risk Control their own Utilization?, Annals of Internal Medicine, vol. 123, pp. 500 504. Klees, B. S., Wolfe, C. J. and Curtis, C. A. (2009) Brief Summaries of Medicare and Medicaid Title XVIII and XIX of the Social Security Act as of November 1, 2009. Accessed from: https://www.cms.gov/MedicareProgramRatesStats/downloads/MedicareMe dicaidSummaries2009.pdf Lesser, C. S., Ginsburg, P. B. and Devers, K. J. (2003) The End of an Era: What Became of the Managed Care Revolution in 2001?, Health Services Research, vol. 38, pp. 337 355. McClellan, M. (2000) Medicare Reform: Fundamental Problems, Incremental Steps, The Journal of Economic Perspectives, vol. 14, pp. 21 44. Mechanic, D. (1989) Consumer Choice Among Health Insurance Options, Health Affairs, vol. 8, pp. 138 148. Miller, R. H. and Luft, H. S. (1994) Managed Care Plan Performance since 1980 A Literature Analysis, The Journal of the American Medical Association, vol. 271, pp. 1512 1519. Monheit, A. C. and Cooper, P. F. (1994) Health Insurance and Job Mobility: Theory and Evidence, Industrial and Labour Relations Review, vol. 48, pp. 68 85. Mooney, G. H. (1983) Equity in Health Care: Confronting the Confusion, Effective Health Care, vol. 1, pp. 179 185. Morris, S., Devlin, N. J. and Parkin, D. (2007) Economic Analysis in Health Care, Chichester: J. Wiley & Sons. Mossialos, E. and Dixon, A. (2002) Funding Health Care in Europe: Weighing up the Options. In: Mossialos, E., Dixon, A., Figueras, J. and Kutzin, J., Editors (2002) Funding Health Care: Options for Europe, Guildford ; Kings Lynn: Biddles Ltd.
66

Nonneman, W. and van Doorslaer, E. (1994) The Role of the Sickness Funds in the Belgian Health Care Market, Social Science & Medicine, vol. 39, pp. 1483 1495. OECD, OECD Health Data 2010 Frequently Requested Data. Accessed From: http://www.oecd.org/document/16/0,3343,en_2649_34631_2085200_1_1_ 1_1,00.html Olson, C. A. (2002) Do Workers Accept Lower Wages in Exchange for Health Benefits?, Journal of Labour Economics, vol. 20, pp. 91 114. Orenstein, W. A. and Hinman, A. R. (1999) The Immunization System in the United States the Role of School Immunization Laws, Vaccine, vol. 17, pp. 19 24. Pauly, M. V. (1974) Overinsurance and Public Provision of Insurance: The Roles of Moral Hazard and Adverse Selection, The Quarterly Journal of Economics, vol. 88, pp. 44 62. Phelps, C. E. (2003) Health Economics, 3rd ed, Boston: Addison Wesley. Rice, T., Desmond, K. and Gabel, J. (1990) The Medicare Catastrophic Coverage Act: A Post-Mortem, Health Affairs, vol. 9, pp. 75 87. Roemer, M. I. and Shain, M. (1959) Hospital Utilization under Insurance. Hospital Monograph Series, No. 6, Chicago: American Hospital Association. Rothschild, M. and Stiglitz, J. (1976) Equilibrium in Competitive Insurance Markets: An Essay on the Economics of Imperfect Information, The Quarterly Journal of Economics, vol. 90, pp. 629 649. US Census Bureau (2009) Income, Poverty, and Health Insurance Coverage in the United States: 2009. Accessed from: http://www.census.gov/prod/2010pubs/p60-238.pdf

67

US Department of Labour Bureau of Labour Statistics, Urban Consumer Price Index. Accessed From: ftp://ftp.bls.gov/pub/special.requests/cpi/cpiai.txt US Government, Timeline: Whats Changing and When. Accessed From: http://www.healthcare.gov/law/timeline/index.html van de Ven, W. P. M. M., Beck, K., Van de Voorde, C., Wasem, J. and Zmora, I. (2007) Risk Adjustment and Risk Selection in Europe: 6 Years Later, Health Policy, vol. 83, pp. 162 179. Wagstaff, A. (1986) The Demand for Health Some New Empirical Evidence, Journal of Health Economics, vol. 5, pp. 195 233. Wagstaff, A. (1993) The Demand for Health: An Empirical Reformulation of the Grossman Model, Health Economics, vol. 2, pp. 189 198. Wagstaff, A. (2009) Social Health Insurance Reexamined, Health Economics, vol. 19, pp. 503 517. Wagstaff, A. and van Doorslaer, E. (1992a) Equity in the Finance of Health Care: Some International Comparisons, Journal of Health Economics, vol. 11, pp. 361 387. Wagstaff, A. and van Doorslaer, E. (1992b) Equity in the Delivery of Health Care: Some International Comparisons, Journal of Health Economics, vol. 11, pp. 389 411. Wagstaff, A., van Doorslaer, E., van der Burg, H., Christiansen, T., De Graeve, D., Duchesne, I., Gerdtham, U-G., Gerfin, M., Geurts, J., Gross, L., Hkkinen, U., John, J., Klavus, J., Leu, R. E., Nolan, B., ODonnell, O., Propper, C., Puffer, F., Schellhorn, M., Sundberg, G. and Winkelhake, O. (2000) Equity in the Delivery of Health Care in Europe and the US, Journal of Health Economics, vol. 19, pp. 553 583. Wilensky, G. R. and Rossiter, L.F. (1983) The Relative Importance of Physician-Induced Demand in the Demand for Medical Care, The Milibank Memorial Fund Quarterly. Health and Society, vol. 61, pp. 252 277.
68

Williams, A. and Cookson, R. (2000) Equity in Health. In: Culyer, A. and Newhouse, J., Editors (2000) Handbook of Health Economics vol. 1B, Amsterdam; Oxford: Elsevier. Woolhandler, S., Campbell, T. and Himmelstein, D. U. (2003) Costs of Health Care Administration in the United States and Canada, The New England Journal of Medicine, vol. 349, pp. 768 775. World Health Organisation (2000) The World Health Report 2000 Health Systems: Improving Performance, France: Sadag.

69

Appendix A. Insurance Markets

Insurance markets, fundamentally, transform uncertainty into risk and offer protection against that risk. On the supply side, an insurer will offer protection against the risk of a loss of wealth (L) that will occur with an estimated probability (p), and also charge an amount to cover

administration costs (); therefore, an actuarially fair premium () for any individual can be estimated by:
i = (1 + ) pi L
Equation 3: An Actuarially Fair Premium

Figure 11: Risk Aversion in Wealth

70

The premium that an individual is charged is increasing in the administration costs faced by the firm, the probability the loss-triggering event occurs and the size of the loss induced by that contingency46. The demand for insurance arises as a result from risk aversion; specifically, we are concerned with a persons risk aversion in wealth. Let us assume an individual has a utility function of the following kind:
U i = u (..., W ,...)

Equation 4: Utility Function of Representative Individual

Where the individuals utility (Ui) is, inter alia, influenced by the level of wealth (W) in their possession. We can denote this individual as risk averse in wealth if the following two conditions hold:

U i 2U i > 0; <0 W W 2
Equation 5: Conditions for Risk Aversion

In essence, the individual must have a positive, but diminishing, marginal utility of wealth. The implications of this risk aversion become clear when shown graphically in figure 3. Let us assume that the individual only derives utility from wealth and that there are two states of the World; in the first state, the individual has wealth W1 and a utility level of U(W1) in the second, they suffer a medical emergency and have a lower level of wealth, W0 with a corresponding utility of U(W0). If we, for simplicity,
46

This is a simplification, of course, although the general principle of all actuarial insurance is based upon this notion of a supplier being able to link a premium to the degree of risk involved.

71

assume a value of p to be 0.5, then their expected level of wealth can be given by W, where W W0 = W1 W; under certainty, the level of utility associated with W would be U(W)47. However, due to the presence of uncertainty, the individual considers the expected utility instead - which is given by U(W*) and equivalent to the level of utility gained from holding W* with certainty. To link the supply and demand sides, we can observe the maximum premium, *, that the insurer can charge. Noting the fact that the insurer offers certainty, and that that individual is indifferent between W with uncertainty and W* with certainty:

* = W1 W * F = W 1 W ' R = W 'W *
Equation 6: Maximum Premium, Fair Premium, Risk Premium

As shown above, the maximum premium can be decomposed into both a fair premium, which will leave the individual as well off as they would be with their mean wealth, W, and a risk premium which increases the more the individual abhors uncertainty. It should be noted that the risk premium component is not purchased by risk neutral48 or risk loving49 individuals as U(W*) U(W); the risk neutral individual will purchase the
47

Due to the concavity of U(W), U(W) U(W0) U(W1) U(W) Risk neutral individuals have a utility function linear in wealth, such that:

48

49

U i 2U i > 0; =0 W W 2

Risk loving individuals have a utility function convex in wealth, such that:

U i 2U i > 0; >0 W W 2 72

fair premium, whilst the risk loving individual will purchase only a fraction of the fair premium. However, as a study conducted by Friedman (1974) shows, there is little doubt that consumers of health care tend to be, to varying degrees as the study notes, risk averse, so we shall not pursue the risk neutral or loving cases any further.

73

Appendix B. The Adverse Selection Death Spiral

Adverse selection is a problem of hidden information. In healthcare and arises when a patient has a superior knowledge of their likely demand for healthcare than an insurer; this means that an insurer is unable to differentiate between a good risk and a bad risk that will be more expensive to insure due to a higher p. It is in Akerlofs (1970) seminal article that the full extent of the problem is made clear; he notes that in the used car market, due to the information asymmetry regarding the quality of a used car, the presence of poor quality cars drives the price up until the market eventually the market breaks down. In health insurance a similar phenomenon arises; let us begin with the concept of community rating, which divorces a premium from the individuals level of risk, instead focusing on the average risk of a population.

C1 = ( 1 + ) ( p H + (1 ) p L ) L
Equation 7: Community Rating with High and Low Risk Groups

A premium arising from such a policy will resemble equation 3, where is the proportion of high risk individuals in the population and pH and pL are the respective probabilities of a high risk and low risk individual needing healthcare. If we note that pH > pL then we can rank the community rating premium against the premiums that each group would occur based on their specific risk:

74

 L < C1 < H
Equation 8: Ranking of Premiums

The high risks benefit from this community rating, as the low risks drive the price of the premium below what they would pay otherwise; however the low risks lose as they effectively subsidise the high risks. It therefore follows that the low risks would either drop out of the market, or substitute to a cheaper plan. The insurance company would therefore be forced to reconsider the premium offered:

C 2 = (1 + ) p H L = H
Equation 9: Community Rating Post Low Risk Exit

Hence, the price of the premium is inflated due to the higher average risk of the insureds. If we assume that there is a spectrum of high risk individuals50 then the above scenario repeats itself, with the lower risk group dropping out and the premium price rising once again. This is what Cutler and Zeckhauser (1998) term the adverse selection death spiral; because higher risks will always be better off mixing with lower risks, any plan that caters to the higher risks will become continually more expensive until it is no longer feasible and either everyone is uninsured or they end up on a single plan that caters specifically to low risks.

50

For example, one may decompose High risk individuals into Low-High Risk and High-High Risk.

75