Get In-Stock Alert. Delivery not available. Pickup not available. About This Item We aim to show you accurate product information. Manufacturers, suppliers and others provide what you see here, and we have not verified it. See our disclaimer. Details Coming Soon Health Economics. Specifications Publisher Routledge, Taylor and Francis. Customer Reviews. Write a review. See any care plans, options and policies that may be associated with this product. Email address. Please enter a valid email address. Walmart Services.
Get to Know Us. Customer Service. In The Spotlight. Shop Our Brands. All Rights Reserved. Cancel Submit. In industry mechanisation, division of labour and new materials have allowed goods to be produced at much lower cost. To some extent health care can be seen as such an industry. The prices of handicraft goods generally rise over time relative those of other goods. Whereas the workers in industries with rapid technical change may obtain higher wages from sharing in the gains from higher productivity, in health care higher wages are likely to lead to higher costs of care.
The increase in the relative cost of health care may therefore be seen as an inevitable feature of an industry that is more handicraft than mass-production in its technology. It is never easy to answer this question. That much is simple. But what determines the price we have to pay for inputs, and what is the value of the best alternative use of the resources? Indeed, what is the best alternative use of the resources? Health care systems are normally subject to many rules and regulations — often these are important sources of safety and protection to the users.
But restrictions can also distort prices. In many cases the price of inputs wages and prices, capital and consumable inputs include an element of economic rent — that is, they are paid more than is necessary to make them available. More complicated issues arise with patented devices and drugs. Development costs are recovered through charging prices that are above the direct cost of producing the goods and maintaining the equipment.
But it does mean that the price charged is above cost. The practical question of how we assess what is the opportunity cost is explored further when we discuss economic evaluation in Chapters 11 and An obvious question is whether there are economies of scale, i.
However, it is important to be clear why this might be the case. There is some evidence that the costs of relatively high-technology procedures fall with volume e. Cronin et al. Although this trend may be now reversing, it is still common to hear planners and clinicians referring to such changes as rationalisation. In the past most developments of health care infrastructure have been Cost of delivering health services 39 carried out without good evidence about the relationships of what is produced, how it is produced and at what cost it can be produced. Estimating cost functions for health care provision can allow better understanding.
There is a growing literature on cost functions in health care provision, and it is becoming clear that economies of scale are quite limited, but can be important in certain types of service. Case mix is normally very important. In this case total cost rises at a decreasing rate at lower output, and at an increasing rate at higher output. This suggests that at lower levels of output there are economies of scale, but these disappear and are reversed at higher output. For some purposes we are interested in the total cost of provision — for instance, if we are setting a budget for the service.
However, for most purposes we are interested in two other measures — the cost of each unit of service the average cost and the cost of one more or less the marginal cost. From the total cost curve here we can calculate the average cost. We can see that at lower output the average cost is falling, since with each extra unit of service all units become cheaper. However, as output rises this is reversed. This is shown on Figure 5. From this pattern we can derive the average cost — it will be U-shaped in this case.
We can also calculate the marginal cost. The marginal cost is the additional cost of one more unit of service. This can be described as the change in total cost divided by the change in output — and of course this just describes the slope of the total cost curve. This slope is positive all the way along the total cost curve, but the slope is initially decreasing, and Figure 5.
Similarly the marginal cost curve therefore starts by falling, and then increases. The average and marginal cost curves are shown in Figure 5. The observant reader will notice that the MC curve crosses the AC curve at its lowest point. This is not an accident. If the average cost is falling, it follows that the cost of one more unit of service the MC must be less than the average. If the average is rising, the cost of one more must be above the average.
Thus the MC curve always intersects the AC curve at its lowest point. The normal interpretation of the cost of a treatment or operation is the average cost. As will be shown in Chapter 11, we need to be careful. We argued that there were good reasons, both theoretical and from the observation of behaviour, to expect downward-sloping demand curves and upward-sloping supply curves in health markets.
For example, where prices for health services are charged or are implicit for example in distance and waiting times , we expect less health services to be demanded the higher they are. We also normally expect suppliers of health services — such as suppliers of pharmaceuticals or of skills and labour — to supply more as the price increases, ceteris paribus although we will note important exceptions to that in the next three chapters.
Where we have downward-sloping demand curves and upward-sloping supply curves, the determination of price, output and the allocation of production and consumption should, in principle, be straightforward. Figure 6. Consider what would happen at any price away from this point. The result is a shortage of goods, and suppliers can increase prices and still sell what they want. Price can be stable only at the equilibrium point. At equilibrium, consumption is allocated to those willing to pay this price or more those whose demands are illustrated to the left of the demand curve and production to those willing to supply at this price or less whose willingness is illustrated to the left of the supply curve.
This is a very simple model and, as much of the rest of the book will explore, does not capture the complexities of many real markets including health. However, it is an idea with enormous power to generate initial expectations about behaviour in markets which often turn out to be realised, even in health markets.
These types of predictions often hold. It is best to think of a perfect market as you would think of a perfect square when studying geometry. It is not found in nature but it is an enormously useful tool to use to understand shapes which are. If you want to calculate the area of a complex shape it is useful to understand the principles behind calculating the area of a square. It is safe to say that no perfect market exists, although we may be able to point to some markets which come quite close.
Price then captures the value of the last unit of the good purchased the marginal value to the individual, in terms of whatever alternative good is considered. When all choices have been made, the value of the last cent spent must be equal for each good purchased — assuming that an independent decision can be made over each cent. Money therefore represents all other goods the individual values and knowledge of how demand changes as price changes tells us about how much individuals making up a market value each individual good at the margin.
Whatever the price is, the utility-maximising individual will purchase those units of the good for which his utility values are higher than the price and will stop purchasing at the point where his values fall below the price. At the margin, he equates his value with the price. At the market level, all individuals are making the same assessment. It follows that demand curves are powerful tools for telling us about how individuals value the goods they Basic market models 43 purchase, and indicate the aggregate marginal value or marginal utility of goods in that market see Box 6.
Box 6. It is measured by the area under the demand curve and above the price line, since this sums the additional value to the consumer of each unit purchased above the price, or marginal cost to the consumer of purchasing it. The demand curve reveals the marginal value to her of each additional packet of aspirin consumed per year. As the quantities considered get larger either because we consider smaller units like individual aspirin for Mrs Brown, or because we consider markets involving large numbers of people it becomes more accurate to draw a smooth demand curve, and more accurate to consider the area of consumer surplus as a triangle, underneath the demand curve and above the price line.
Total revenue is equal to the price multiplied by the quantity sold. Under price-taking circumstances, the given price indicates both the average revenue and the marginal revenue which is generated by selling one more unit of output. In other words, the supply curve can be traced along the part of the marginal cost curve which is upward-sloping and lies above the average cost curve.
As explained above, it will choose to produce where price equals marginal cost. Consumers have perfect knowledge, are purchasing a homogeneous product identical in every respect, including ease of access to each supplier and will purchase from any supplier, up to the point where they equate price and marginal utility. There is no opportunity for an individual supplier to sell goods at higher than the prevailing price — consumers would all desert and go to other suppliers.
These assumptions produce price-taking behaviour. The market can reach equilibrium only through the entry and exit of buyers and sellers to and from the market in response to price movements which are generated by the market, independently of individual buyers and sellers. This produces downward pressure on price, implying a new equilibrium at p2.
Productive activity has been allocated to those products which consumers value in excess of their cost marginal cost has been equated with marginal utility. All those remaining must be operating on isoquants not above. All those remaining must be equating the marginal rate of substitution of inputs in the production process with the price ratios of those inputs.
If all markets in the economy are assumed to work on the same basis we can move from a partial to a general equilibrium analysis. A general equilibrium analysis enables the analyst to focus on the linkages between markets. For example, each output uses inputs in its production, and these inputs are themselves traded in markets which for the purpose of this analysis are assumed also to be perfect. There is no reallocation of resources which could increase the welfare of one consumer without a cost to the welfare of another. In such a perfectly competitive economy, a number of marginal conditions are shown to hold, including: 1 2 3 Consumption.
Consumers will consume at the point where their marginal rate of substitution between two goods is equal to the price ratio between those goods. Firms will produce at the point where the marginal rate of technical substitution between the two factors is equal to the ratio of the prices of those factors. Production sectors plans are brought into line with the plans of consumers through the price signals of the market — which always indicate the marginal costs of production. However, conditions for perfect competition rarely prevail and in many cases are unlikely ever to prevail.
Consider again the market linkages discussed above. Suppose that one market which provides a substitute good to that under analysis is produced by a monopolist. Review the monopolistic market model in section 6. This implies that the price is above marginal cost. In choosing between the good produced in the perfectly competitive market, and that produced in the monopolistic market, consumers will equate marginal utility per rupee spent across the two goods. It would cost society less to achieve the same levels of utility for consumers by switching production towards the good produced by the monopolist where the marginal costs of production are lower.
Note that this conclusion is independent of any concern for equity in market allocations which, as discussed in Chapter 9, is not a component of the Pareto criterion.
Given the wide range and number of deviations from perfect competition existing in any economy, precise measurement of the degree and implications of individual deviations within their own or related markets is not a feasible objective of economic policy. Rather, a more general point might be carried from this discussion to the debate over laissez-faire versus interventionist approaches to health policy. We explore this point further in the context of regulation in Chapter Under monopoly there is no supply curve because the supplier dictates price and sets price and output together.
If all units of the good produced by the monopolist are sold at the same price an assumption of this model , price equals average revenue. Marginal revenue can be derived from the demand curve, but is always less than average revenue because in order to sell one more unit of the good the monopolist must reduce the price charged for all the units sold. Basic market models 49 Figure 6. The monopolist may operate at the minimum point of the short-run average cost curve, but only if this happens to coincide with the output level associated with equating marginal revenue and marginal cost.
Almost every assumption of the perfect market model is a rare or non-existent phenomenon. In some respects, the models themselves help to explain why that should be. Suppliers collude in order to avoid price taking. What are your expectations of the attitude of medical associations to the expansion of medical student numbers? In their ability to cast light on these phenomena the models are already revealing their ability to perform as analytical tools — rather than to describe reality.
Nevertheless, some markets look a lot more like perfect ones, and others a lot more monopolistic. Within the health sector, the market for surgery might be concluded to be less perfect than the market for over-the-counter pharmaceuticals. Must it be true that attempts by government to intervene in markets are misguided?
Health Economics | An International Perspective | Taylor & Francis Group
Such ideas seem to contradict common sense in health markets, where government intervention is popular and intuitively useful, and encouraging free entry to medical services — allowing anyone to practise medicine or nurse, regardless of their training — appears foolhardy.
Fortunately, such ideas contradict economic theory too. This leads to a great deal of confusion as to the purpose of the perfect market model. To repeat, it is not an attempt to describe reality, and it is not in itself capable of prescription. The theory of the second best provides a path through this morass. It follows that the conditions for perfect markets have no normative implications. Perfect government intervention is as rare as a perfect market.
It is as important to be aware of government failure as of market failure. The model is a tool of analysis which cannot produce policy prescription in itself. If economies of scale are exhausted at levels of output consistent with large enough numbers of sellers to ensure price-taking behaviour, the model continues to work. However, economies of scale are usually assumed to be exhausted at higher levels of production than this, and to be one of the factors which explain the existence of markets characterised by one or a few sellers.
If so, perfect competition can be assumed to imply the absence of economies of scale in the long run. The standard model assumes that consumers are sovereign, or are the best judges of their own interests. As soon as we enter health markets a range of objections to this perspective implicitly and explicitly arise. Hence, almost all countries require citizens to purchase insurance, or do so for them through the tax system see Part IV.
Similarly, a range of health-risking behaviours are usually proscribed by law — some drugs are almost universally outlawed for recreational use, and diving into the sea from Brighton pier is prohibited; and some health-promoting behaviours are required, such as immunisation in some countries.
These rules can be interpreted as requiring socially responsible behaviour protecting society from the selfinterested behaviour of individuals , a standard explanation of rules in any society, but they seem also to contain an element of policy makers knowing best. This knowing best may involve a belief in superiority of judgement in some cases you may well understand the risk of recreational use of heroin but we still will not allow you to damage yourself , an attitude termed paternalism. More commonly, it involves the belief that the information held by the rule maker is superior to that of the ruled you do not understand the risks of recreational use of heroin and must be protected from your ignorance.
This is one type of information problem encountered in health markets. More generally, there are many important areas of specialised knowledge involved in the seeking of health status improvement which are not involved in the seeking of a banana. Almost everywhere there is a professional there is an agent in disguise.
In some cases of agency, including much but not all of its occurrence in health markets, the agent is also the supplier of the service. It also gives rise to the potential for that much debated phenomenon in health economics: supplier-induced demand. What would the objectives of a doctor acting as a perfect agent in recommending and supplying health services be? Possible candidates include: 1 2 3 To maximise the health status of the patient. To maximise the utility of the patient. To maximise the health status or utility of the whole society. Lack of clarity in this is expressed in a wider debate about the role of doctors in society which pervades much more than health economics textbooks.
The second option, maximising the utility of the patient, might suggest that the primary role of doctors should be to provide information to patients and, as far as possible, leave patients to make their own decisions. This seems to accord with many current ideas. The third option, maximising the health status or utility of the whole society, responds to the growing awareness of the inevitability of rationing in health markets — awareness that to provide a service to one patient may entail denial of another patient. Attitudes to this objective are likely to be health system-dependent. Doctors employed privately by the patient have a much clearer duty to that individual patient, and any rationing implications for others are at least indirect and may not arise if it is simply a question of that patient deciding to spend less of their own resources on other goods.
Doctors employed in a public system are potentially the agent of both the state and the Supplier-induced demand and agency 53 individual patient. Some health systems impose this role on doctors much more than others. If drugs are in very short supply, the doctor cannot but consider those given to one patient as lost to another — perhaps an hour later in the same clinic.
This leaves even the best intentioned doctor in an unenviable position in attempting to clarify their own role. Inappropriate advice motivated by imperfect information, for example, is usually not considered SID. The explanation that demand is induced where supply exceeds the level it would otherwise attain requires a stronger test. SID implies the shifting of the demand curve. A doctor who aims to encourage more patients to visit her surgery by reducing her fees in other words is aiming to produce movement along the demand curve is acting only as any other supplier of goods may act.
The problem with attempting to observe the shift of the curve is that demand curves themselves cannot be observed, only equilibrium points. Consider Figure 7. One way of looking for SID is to follow what happens in a health services market when the supply of doctors increases. In a standard market Chapter 6, section 6. However, if doctors are able to shift the demand curve, they will be able to protect themselves against lower fees. They may be able to do so, partially, by shifting demand curve D to D1; wholly, by shifting the demand curve to D2; or may even shift the demand curve as far as D3, in which case fees would increase Figure 7.
However, the observer can observe only point A, and alternative points B, C and D, and not whole demand curves. If points A and B are observed, the pattern is consistent with the 54 Introductory health economics Figure 7. Only if a point to the left of point C is observed, such as D, is it clear that the market is not normal: supply of doctors increased but so did fees per visit. Looking for situations in which a movement from A to D seems to have been caused by an increase in the supply of doctors is known as the fee test of inducement, attributed to Reinhardt It will not surprise you to learn that the evidence about inducement is mixed and that most reviewers of this literature conclude that SID is an unproven hypothesis McGuire et al.
Two attempts to identify presence of SID are discussed in Box 7. Supplier-induced demand and agency Box 7. As discussed in the text, there are many alternative explanations of association besides supplier-induced demand. If demand is induced, it would therefore be expected that the proportion of physician-initiated visits would be higher than if it is not. For the population in this study 14, households interviewed across the United States between and , 39 per cent of all ambulatory visits were physician-initiated and 52 per cent patient-initiated the remaining 9 per cent were referrals from elsewhere, or of unknown initiation.
Wilensky and Rossiter applied regression analysis to these data, and data relating to the supply of physicians the physicians or surgeons per , population in the county or small area in which the patient resides. However, evidence of inducement in relation to surgery, total physician-initiated expenditures and physician fees was absent.
Most studies of supplier-induced demand, like that of Wilensky and Rossiter have focused on the situation in the United States, where fee-for-service reimbursement is an important mechanism for paying for services, especially at ambulatory level. In most other countries the problem is less likely to arise. For example, in the UK medical care is not reimbursed on a fee-for-service basis, and the preconditions for supplier-induced demand are absent.
Roberts This can be represented as a shift in the demand curve. A shift in the demand curve caused by a falling shadow price cannot be distinguished from a shift caused by inducement. However, it is the shadow price of the visit which has fallen, and not of the services provided during the visit. If no inducement takes place, content per visit should decline as the shadow price falls, since marginal attenders would be likely to require a lower volume of treatment. In contrast, Grytten et al.
They argue that where there is a scarcity of physicians the choice to work longer or more intensively meets unmet demand, rather than induces demand. However, this relation failed to hold in areas of high physician density, suggesting that inducement does not occur. Much depends on the source of our concern.
In public health systems this focus on demand rather than on need may also seem peculiar. It is clear enough from much less complex studies that doctors are human, and fall short of this ideal. When fee-for-service reimbursement is used, more services are provided than when systems of reimbursement that reward on a per-case or per-patient covered basis are used see Box 7. All this implies that we have a situation of imperfect agency. Medical science is imprecise, but that is not the source of the problem, or at least not the whole source.
When advice responds to economic incentives, doctors cannot be applying their best guess about the best possible treatment all the time. It may be that uncertainty is exploited to play safe when incentives are set to reward higher levels of activity. The opposite may be decided and slight risks taken when incentives lean the other way. Fees were paid for all types of contacts with patients and for diagnostic and certain curative services.
The table shows dramatic responses in terms of the provision of additional services for which fees were paid, and reduction in referral which not only did not pay but would in some cases have substituted for paying services. Number of contacts and activities in a week and number of enlisted patients in March , March and November for seventy-one doctors in Copenhagen city Activity March March November Contacts Diagnostic servicesa Curative servicesa Referrals to specialist Referrals to hospital Number of enlisted patients 9, 99 1, , 11, 1, , 10, 1, , a Services for which an additional fee is paid.
Source: Mooney This evidence suggests that while fee-for-service is associated with higher levels of provision of services by those receiving fees, it also reduces their referral of patients elsewhere. Mooney draws attention to the fact that over 20 per cent of referrals to specialists and nearly 30 per cent of referrals to hospital made under capitation appear to have been unnecessary, at least according to the implicit criteria adopted after the change. Changing payment, reimbursement and patterns of care in Vietnamese hospitals Sepehri et al.
During the s, among other changes in the system, health insurance was encouraged, and user fees were introduced. Public hospitals were reimbursed on a fee-for-service basis, either by patients through the user fee system, or by insurers at higher fee rates. By This is not all bad. Few health policy makers now ignore this opportunity. To this extent, the material of this chapter is as relevant to bananas as to health care.
In this chapter that tool will be used further. Mostly, the analysis involves only demand and supply curves, but it is important to remember that in a perfect market, demand curves are marginal utility curves and supply curves are marginal cost curves review sections 6. It is important to remember that divergences from perfect markets are not necessarily bad things in themselves — positive externality is a good example see 8.
Government intervention is recommended only if likely government failure to intervene perfectly is less than the market failure under assessment. This results in tangible costs. However, the neighbour on the other side may have enjoyed the programme — a positive externality on the consumption side! The ones which are likely to be most important are positive ones. There are clear externalities in the treatment and prevention of infectious disease.
If vaccination is received by the majority, the unvaccinated minority may also be protected. The second type of positive externality which may be important in health is known as the caring externality. It would appear that we are far more likely to care whether others receive health care or not than we are to care whether others drive prestigious cars or eat out at restaurants regularly. The good Samaritan provided health care to a stranger. Both these types of externality occur in consumption, or on the demand side.
In Figure 8. We assume that there is an upward-sloping supply curve equal to marginal cost. The Figure 8. He will therefore not purchase it. From this perspective the visit is a good buy, but the purchase will not be made in a free market. The problem is the presence of values which are not translated into willingness to pay in the market. We might use charitable institutions to solicit from the public, money to subsidise doctor visits. We might impose such subsidy through a tax system.
Both involve administrative transaction costs. Charity encounters free-rider problems — although the Samaritan cared about the stranger, he might have waited to see if someone else would pick him up. Tax systems may involve compulsorily extracting payment from those who do not care — subtracting more from their utility functions than they gain.
MSc International Health Policy (Health Economics)
If we are going to provide free services, we should try to estimate how many visits, or how much health education, is appropriate in the sense that social values are higher than cost. In the case of the doctor visits, we would have to ration Mr Pink to that number of visits and set up rationing systems which try to ensure that the visits made are the most appropriate ones. In many circumstances it is impossible to apply such methods and a more intuitive way of tackling the problem is needed.
Equivalently, negative externalities might be addressed by banning or taxing the relevant activity. Policing bans may also be expensive, and poorly policed bans may be the most expensive of all — potentially giving rise to criminal modes of behaviour and inability to exert any controls whatsoever.
A tax which tries precisely to internalise a negative externality i. Pigou — With respect to all these strategies market failure has to be balanced against government failure and an assessment of whether or not the best intervention improves on the market outcome has to be made. It is rival.
Bananas are excludable. Public goods are characterised by non-rivalness and non-excludability. A theatrical performance is non-rival to the extent that, once provided for one person, it costs nothing to provide it to others up to the seating capacity of the theatre. Non-excludability means that it is not possible to exclude nonpayers from consuming.
Minimum entry requirements for MSc International Health Policy (Health Economics)
Non-excludable goods are also usually non-rival, but non-rival goods are quite often excludable. It is quite easy, and indeed usual, to exclude non-payers from a theatrical production. Environmental health services and public health education campaigns are examples of goods which are both non-rival and non-excludable and are therefore close to being pure public goods.
Positive externality in consumption implies a degree of nonrivalness and although the two types of market failure are usually separated, it may be more useful to identify degrees of public goods characteristics. Figure 8. Cancer treatment is both rival and excludable and therefore a private good. Immunisation is also relatively easily excludable although there are advantages in providing the service to groups of people at once moving it slightly to the left in comparison with cancer treatment on the ease of exclusion axis. Consider Figure 8.
Source: Bennett , p Market failure and government 63 Figure 8. The private demand curve for tickets Dp , as usual, can be used to estimate the value of tickets to individuals. If the price is greater than zero for example, p1 , there will always be some people who would value a ticket would pay more than zero price but would not pay the price charged. Providing free tickets to these people would create more value in society than it would cost.
There is therefore an opportunity to increase social utility which is not seized. In many cases, the total cost of providing the good will exceed the willingness to pay even of the individual who sets the highest value on it. Under these circumstances the good will not be provided at all by the market. Since non-rivalness is an extreme form of positive externality on the demand side, health policy strategies responding to a recognition of public goods characteristics echo those discussed to deal with positive externality in section 8. However, if marginal costs are truly zero, there is a stronger likelihood that free provision of the good in question, as opposed to a limited subsidy, is the appropriate strategy, and if non-excludability also applies, this is the only feasible course.
It should be remembered that public provision implies taxation and uncompensated losses to those who do not value the public good. Public provision might answer both non-rival and non-excludable characteristics of public goods. The conclusion that it is impossible to do so might be read as meaning that the opportunity costs of doing so would be unreasonably high. A number of factors militate against this.
In contrast, if variable costs are relatively high, a business can start small and expand as its business does. These arguments are especially relevant to more sophisticated and high-technology services which require large capital investment, often of a sunk nature — for example, diagnostic services such as x-ray and CT scans, and operating theatres and equipment. More generally, the provision of hospital services as a whole encounters this constraint.
Rather, strategies are likely to focus on public provision and regulation. The latter subject is discussed in Chapter Market failure and government 65 8. In addition to failure resulting from agency itself, Chapter 7 suggested that imperfect knowledge causes the relationship between the demand curve and marginal value to be questioned.
For example, in the early s there was a high demand for the drug thalidomide in pregnancy. Later information as to its link with certain congenital malformations reduced that demand to zero. Health service costs can be large and timeconcentrated. Health insurance ensures that contribution levels will prove adequate, even where costs for health service needs prove to be above average. However, health insurance also causes the market to fail. The two main ways in which it does so are through moral hazard and adverse selection.
Moral hazard arises when consumers face zero prices at the time of use: their utilisation decision does not take account of marginal cost. Adverse selection arises where low-risk individuals are able to opt out of the risk pool. As a result, average risk rises, causing the cost of insurance to rise and further lower-risk individuals to opt out. Ultimately it may be impossible to organise a market in health insurance, or at least one which does not exclude a large proportion of those needing to access health services.
Sensible strategies therefore do not try to imitate perfect markets by trying to remove insurance arrangements but are more likely to regulate the insurance market. Compulsory insurance or social insurance is a regulated arrangement which addresses adverse selection, for example see Chapter For example, the no claims bonus is an attempt to mitigate moral hazard see Chapter 14 on contracts. If 66 Introductory health economics health services are merit because they relieve pain or even extend life, perhaps those health services which do neither some types of plastic surgery might provide an example should be excluded.
Recognition that even the most prosperous societies could allocate their entire GNP to health interventions with life-extending and painrelieving potential however minute or unlikely suggests that the scale and likeliness of such potential have to be taken into account. These considerations tend to suggest that more relativist rationales for health policy are required. Hence markets fail in that even perfect markets make no special claims with respect to equity. This is a very persuasive argument, but it warrants some further examination.
Production of health services has to take place, or there is nothing to share round equitably. Ceteris paribus, even the most avid egalitarian would prefer more to share round than less. Concern that productive resources are not just wasted, and in fact are used to produce the most health care possible, for any given distribution must be important in any schema.
Second, a set of misunderstandings can arise from the normative rather than analytical use of the perfect market model. According to this list, equity may imply equality of: 1 expenditure per capita; 2 inputs per capita; 3 inputs for equal need; 4 access for equal need; 5 utilisation for equal need; 6 marginal met need; 7 health. Notes 1 This implies that Pareto optimality cannot be inferred — see Chapter 9. Part II Economic evaluation 9 The theoretical bases of economic evaluation 9.
To do this we check that there is no alternative mix of goods we could consume that would increase our welfare. Once this is achieved we cannot increase our welfare simply by changing the mix of goods and services we consume. By analogy, when a decision is made on behalf of society to provide some services collectively, the aim may be to maximise the welfare of the whole of society, and in a similar way to ensure that no change in the mix and pattern of consumption of goods and services would increase overall welfare.
But what works easily for the individual does not so obviously work for collective decisions. They will do this until the marginal rate at which they are willing to substitute the goods the marginal rate of substitution, MRS is equal to the ratio of the prices of the goods. What this is saying is that, for all pairs of goods, and all pairs of individuals, the marginal rate of substitution is the same. Keeping output constant, we can vary the proportions of equipment and labour.
The lowest cost of production will be when no further substitution leads to a reduction in cost, that is, when the MRTS is equal to the ratio of input prices. This principle applies to all inputs.
For production at minimum cost, the MRTS for all pairs of inputs should be equal to the ratio of the input prices. Box 9. Should we give priority to life-extending treatments over those that enhance quality of life? An individual can, in principle, answer these questions for herself, and decide on priorities for her treatment. However, it is important that some principles are applied to the issue of how we make decisions on the use of resources. One obvious starting point is the body of theory on the economics of welfare. The theoretical bases of economic evaluation 71 9.
Using market mechanisms may remain the best available option even when the conditions for perfect markets are violated Chapters 7 and 8. The great strength of markets is the way they use information. There is no need to gather and analyse the data on who wants what, and who wants to provide what services. Buyers and sellers bring the necessary information with them to the market, so it is available when and where it is needed.
What happens in a well functioning market can give us some guide to how this should be done. For example, if we employ ten people to do a job that can easily be done by six, it is just wasteful. If we use expensively equipped facilities for work requiring only basic equipment, that is also likely to be wasteful.
First we need to identify technically feasible ways of producing the desired outcomes. The process of searching for the best combination of inputs should continue until no further changes can be found that would lower the production cost. Formally this is when the marginal rate of technical substitution between any pair of inputs equals the ratio of the costs of the inputs see Box 9. There are options for primary prevention, either personal health promotion activities or by reducing the risk of transmission by the treatment of other sexually transmitted diseases.
There are potential interventions, which can improve the health and extend life for people who are 72 Economic evaluation already seropositive, such as treatment of TB. There are treatments that aim to control the virus directly using combinations of drugs. We could then choose the combination of these interventions that would maximise this for any given overall expenditure. At that point, no further changes in the mix of interventions will increase output. More formally, the marginal rate of transformation between all pairs of interventions is the same see Box 9.
But we are not interested only in interventions that extend life for people with or at risk of HIV. We need to choose the best ways of achieving not only health gain for this disease, but also the maximum health gain overall. Again a useful starting point is the likely choices we each make. An individual chooses to consume combinations of goods and services that maximise her welfare.
Imagine she has a range of chronic conditions such as arthritis, coronary heart disease, diabetes and asthma and has access to only limited funds to buy services. She would choose more of a service if the extra would have a bigger impact on her health than any other use of the resources. Formally this means that her marginal rate of substitution between any pair of goods is equal to the ratio of the prices of the goods see Box 9.
Finally, we have to choose the extent to which we allocate our limited resources to health compared with other goods. Although it is tempting to believe that health care is always the most important good, we can observe that people do not act as if this were so. People take risks with their health by smoking; mountain climbing; driving faster than necessary and even crossing roads they could avoid.
At this level, too, the rational utility-maximising individual would choose to equate the marginal rate of substitution between health goods and other goods. When people interact in a market they trade until no further trades increase their welfare. In a perfect market, with all individuals aiming to maximise their welfare, we see the best use of resources for any given initial distribution of wealth and incomes.
At prevailing prices all individuals have their chosen combination of goods and services. Formally we have a situation where the marginal rates of substitution between all pairs of goods for all pairs of individuals are equal and are equal to the marginal rate of transformation between all pairs of goods and the ratios of prices. In production the marginal rate of technical substitution between all pairs of inputs for all goods and services is equal and is the ratio of the prices of inputs see Box 9.
If we accept the initial income distribution, then we have a situation in which all individuals have maximum welfare. Simply by carrying out voluntary trades people convert their initial endowments, which may include many goods they do not like, into their desired consumption. Those who like holidays spend more on holidays; those who want transport buy bicycles or cars; those who want more housing or food choose those. But before being too dismissive it is interesting to think about situations that are not Pareto-optimal. If further voluntary trade would give people more preferred combinations of goods and services, then their welfare will be improved when they trade.
This is known as a Pareto improvement. Perfectly functioning markets move us automatically towards a Pareto optimum. Collective decisions are normally taken in circumstances where markets have for some reason failed. When making these collective decisions a possible choice rule is to seek out Pareto improvements. This will increase welfare, in the Paretian sense, for any starting point in terms of income distribution.
When we develop a vaccination programme there are many who are saved from disease but some that are harmed and a few who may be killed by the vaccine. When we develop computing skills in teenagers some will help solve problems of world poverty and some will become hackers or write viruses. Almost no activity can be guaranteed to be without losers. When taxes are used to pay for services, then there are likely to be losers, since few people enjoy paying taxes.
We need some basis for judging the relative importance of the gains to the gainers and the losses to the losers. One approach is to ask those who gain to compensate those who lose. Sadly there is a problem with this argument. Since there are losers, then some sections of the population now have lower incomes. It was demonstrated by Scitovsky that there can be cases where the changes in incomes from doing a project lead to a paradox — the potential Pareto improvement can be used to justify a change, and also to return to the starting point see Box 9.
This is most likely Box 9. What Scitovsky was able to show was that there are circumstances in which the potential Pareto improvement criterion could be used to argue both for a change and conversely for a return to the original position. This is clearly worrying if the potential Pareto improvement is used as a rule for resource allocation. A simple example illustrates how this can happen.
For a fuller account of why this occurs see Quirk and Saposnik Imagine the hypothetical case of two people who have hypertension, but the drugs to control it are very scarce. Given the severity of their problems, they are both recommended to take one standard dose of each, and failing that to take double the standard dose of one. Suppose an intervention enabled a change in the world so that scenario 2 applied rather than scenario 1. On this basis the compensation criterion would approve the intervention. However, suppose the intervention instead enabled a change in the world so that scenario 1 applied rather than scenario 2.
It is not clear how likely it is that this will happen, but it is a warning that the apparently modest development of the idea of a Pareto improvement to a potential Pareto improvement may undermine the welfare economics basis of the analysis. Most of the time the Scitovsky paradox is treated as a theoretical nicety. However, many of us are willing to do this. In preparations for nuclear war, priority is to be given to senior politicians and leaders. It is interesting to think about our feelings about these types of priority. The potential Pareto improvement criterion is based on individuals and their welfare.
Other perspectives are possible, involving notions of a common good that are more than simply the sum of welfare of individuals see Box 9. If the utility of any individual rises with no fall for anyone else it increases social welfare in the Paretian sense, but this framework does not allow us to quantify the change. Social welfare increases with increases in the welfare of individuals, but otherwise the shape of the function is not determined.
This has two limitations. In thinking about health care priorities decisions on the use of resources may change the distribution of income and wealth. Many economists argue that if we do not like the current distribution of income we can take measures to change it, such as taxing the relatively rich or subsidising the poor. In the case of health interventions there are still further problems. A particular issue is that in general ill health is concentrated in parts of the population that are relatively poor.
Also, the willingness of richer people to subsidise poorer ones may depend on what is being provided. They may be willing to pay taxes to help people restore their health but not to allow them to increase spending on gambling. Rich people might be willing to subsidise access for poor people to decent housing but not to allow more smoking. Thus the availability of tax funds to provide subsidies may depend on the use to which they are to be put. See Chapter 12 for discussion of how issues of income and equity are sometimes taken into account in economic evaluation.
However, aggregation presents complex issues which are avoided by the Paretian SWF. If this is the case, aggregation is impossible. This approach argues that, in making choices between goods, individuals reveal their marginal valuations of each. It is tempting to conclude that consumer surpluses can be aggregated to produce a measure of total utility — or a SWF.
Some argue that the attempt to derive the SWF from the aggregation of individual welfares is misguided. Social welfare is more than the sum of its parts. This notion provides a standard by which to order alternative states of the world. In contrast a Marxist approach rejects the existence of a unique SWF.
For example, we may wish to impose the condition that particular gains to any individual will be considered of equal value. In this case we are saying that for the purposes of analysis we assume that the utility of an additional year of life, or of a year of good eyesight, or of a year without pain from angina, are given the same value, regardless of who is the individual. In this case, for any category of treatment or care, we would maximise welfare by maximising the health gain measured as years gained, years of good sight, etc.
In principle there is no reason to choose any particular social welfare function.
- The Philosophy of Hilary Putnam (Philosophical Topics, Volume 20, Number 1, Spring 1992).
- L2 Interactional Competence and Development (Second Language Acquisition);
- Health Economics!
- Health economics: an international perspective.
Some people argue for equal access to care regardless of, for example, age. There is an obvious disadvantage in specifying social welfare in this way. A procedure to remove a cataract would produce ten more years of good eyesight in an average seventy-year-old than in an average eighty-year-old.
Most people would choose twenty years of extended life in preference to twenty days, or ten 78 Economic evaluation years of better eyesight over ten days. This also implies the ability to make inter-personal comparisons. In many countries there is a clear consensus in favour of policies and changes that have winners and losers, and it may be more fruitful to spend time eliciting values in the population that can then set the rules for economic evaluation.
Dissatisfaction with conventional welfare economics has led some leading researchers to reject Paretian ideas as a basis for economic evaluation. The main drive to develop alternative frameworks is normally associated with Amartya Sen Sen Culyer and others e. Some important implications follow from a move away from a welfarist basis for economic evaluation. Culyer argues that the objective to be maximised should be health rather than utility. Instead of attempting to devise measures of changes in utility, we have the slightly simpler task of measuring changes in health.
He proposed the use of quality-adjusted life years as the measure of health. In practice, many of the problems associated with utility and revealed preference see Box 9. The debate around the question of age see above has many dimensions in common with the debate around income weights in relation to revealed preference-based measurements. Nevertheless, without the constraints of attempts to remain within a welfare economics framework a wider range of information might be considered to be legitimate.
A case can be made for favouring workers over nonworkers or, as happens in many countries, those who have been injured in the service of The theoretical bases of economic evaluation 79 their country. It is perhaps poignant that both the United States and Vietnam give special access to health care to former soldiers. There is nothing necessarily right or wrong in such judgments, but it is important to know what they are if we are to make sense of the results of any analysis. For example, in assessing antenatal screening programmes it is important to know whether termination of pregnancy is thought of as a legitimate clinical intervention or is considered to be murder.
Economists often assert that the two important dimensions of health care outcomes are length of life and quality of life. So long as the judgements are clear and explicit, they can be included in the analysis, and any recommendations can be shown to depend on these values.
However, the interpretation of the results, and the acceptability of some criteria, may depend on the approach taken. Theoretical debates are not best resolved by voting or numbers. It can be helpful to understand how much and how little we can say on the basis of Paretian principles, since it helps in interpretation of the results of any evaluation. Another important dimension is time. Try an experiment with your colleagues. Ask one of them to buy you a bicycle now. It is unlikely that anyone will agree to this apparently fair deal, and it is interesting to think why.
Several factors may come into play. She may just prefer to have things now to getting them later. When borrowing money we have to pay back the original sum, and also pay interest to 80 Economic evaluation compensate the lender for having to wait for the use of the money. Interest rates may vary depending also on other factors, such as risk. Looking at the level of interest rates over time, there is tendency for rates in real terms i. This is just the working of compound interest. We can easily reverse the process. Tables of values of discount rates are available, and it is easy to set up programmes on a computer to calculate present values.
It is sometimes also convenient to calculate the cost of a project in terms of annual equivalent cost. A present-value sum can be converted into an annual equivalent cost by dividing by the annuity factor. There has been some controversy as to whether it is appropriate to discount all costs The theoretical bases of economic evaluation 81 Box 9.
In all these examples the interest rate is taken to be 10 per cent. The second option is to save a certain amount each year for ten years, allowing interest to accumulate on the saved amount and the van to be bought at the end. The third option is to pay in instalments over ten years after the purchase.
What is important to note is that in an important sense all these sums of money have the same value — in all cases they describe the cost of a new van to be delivered today.