Rationing of Demand


In the presence of health insurance and limited capacity, an excess demand for services remains a permanent feature of several publicly funded health systems. The demand for health care needs therefore to be rationed in one way or another. This article describes three different common types of demand rationing. It distinguishes between (1) direct rationing, (2) rationing by waiting time and quality, and (3) price rationing.

Direct rationing refers to allocation mechanisms which explicitly rule out the provision of certain types of care to patients within the public sector (either by the rules set by the public insurer and/or because the doctor explicitly tells the patient when they demand care).

Rationing by waiting or by quality refers to allocation mechanisms which are implicit: the patient is not explicitly refused care. It is instead the presence of waiting times or low quality of care (either clinical or nonclinical), which induces some patients not to seek care from the public sector and either to opt for no care or care delivered in the private sector.

Finally, the article discusses price rationing, which in publicly-funded health systems takes the form of copayments or coinsurance rates for specific types of care. The article discusses each of these three types of rationing in turn. Although each is discussed in isolation, in practice these coexist in many health systems. The focus is on publicly funded health systems. Demand rationing within private health insurance markets is not discussed.

Direct Rationing

In the presence of complete coverage under public health insurance, patients could potentially demand treatment up to the point where the marginal benefit is zero. In a system with no capacity constraints, this would induce excessive consumption, the well-known issue of ‘ex-post moral hazard.’ However, most countries do limit the supply of care to a level below the one required to satisfy all potential demand, which results in an excess demand. The demand for health care has to be rationed in one way or another.

The capacity constraint is set by policymakers who decide on the number of hospital beds and doctors working in the health sector. The supply of health care can vary significantly across countries (and even within countries) and so does the size of the excess demand. Lower supply levels will require more demand rationing. Organization for Economic Co-Operation and Development countries vary significantly in their health expenditure per capita, with National Health Service (NHS)-type systems spending less than public insurance ones. Rationing is therefore more prominent in NHS-type systems.

From an efficiency point of view, in the presence of a capacity constraint, a natural way to manage the excess demand is to allocate care according to the highest benefit–cost ratio. Policymakers could rank all possible health treatments by their benefit–cost ratio and assign care to patients in descending order until the capacity is exhausted. To some extent this is in line with how governments operate. Treatments that are perceived or shown to have low benefit–cost ratio are not available within the public-insurance package. This is the case for some type of dental and ophthalmological care, plastic surgery, physiotherapy, or alternative medicine. However, listing all possible treatments to which patients are entitled and to compare them on the basis of benefit–cost ratios would be very costly. Public agencies (like National Institute for Health and Care Excellence) do increasingly encourage an evidence-based approach to resource allocation (Drummond et al., 2005), but these still cover only a selection of treatments. One attempt to rank all possible treatments on the basis of cost-effectiveness criteria is the Oregon Experiment, which shows how drawing such comprehensive lists may generate surprising and counterintuitive results (Tengs, 1996).

Moreover, even if policymakers can exclude certain treatments (rationing ‘across’ treatments), it is optimal for governments to pursue rationing also ‘within’ a given treatment. Consider, for example, all patients who could benefit from hip replacement. For some patients costs will exceed benefits: these patients should therefore be optimally rationed. Governments could provide a detailed description for each treatment of the criteria to be used by the provider to ration care. These could be based on severity, patients’ characteristics, pain, and overall health status. Providing such detailed descriptions for each treatment would again be very costly. Although governments do provide guidelines for some treatments, these are unlikely to be comprehensive.

The difficulty for governments in designing detailed rationing rules both ‘across’ and ‘within’ treatments suggests that doctors will play a critical role not only in providing health care but also in rationing care to patients. It is they who ultimately decide who should receive treatment. The rationing function is indeed (implicitly or explicitly) delegated to doctors in most countries. Doctors therefore act as agents on behalf of governments in implementing optimal rationing rules (McGuire, 2000). Governments can (and indeed do) outline the basic principles to health care entitlement and let doctors implement them. Below, such basic principles are first discussed and, then, the conditions under which doctors act as (im)perfect agents and how different incentive schemes may affect rationing.

Policymakers in several countries often state the general principle that access to care in publicly funded systems should based on need. The word ‘need’ can be subject to different interpretations. Need could refer to current health status of the patient, expected benefit from treatment, or patient’s severity (Wagstaff and van Dooerslaer, 2000). For some types of care, these criteria go hand in hand if patients with worst health (higher severity) also have high expected benefits. But this may not be case for other types of care where patients with worst health (higher severity) have low expected benefits (as for some types of cancer care). In the first instance, it is highseverity/benefit patients who are likely to get the treatment. In the second one, a tension between equity and efficiency arises (Hauck et al., 2002). On efficiency grounds (health maximization), patients with higher benefit should receive the treatment. On equity grounds, the priority may be reversed and patients with higher severity (and worst health) should receive the treatment. It is also worth emphasizing that the principle that access should be based on need offers little guidance on how costs should be taken into account. If patients with higher need are very costly, it does not necessarily follow that those patients should receive the treatment. In summary, general principles are useful but they are open to different interpretations by providers. A large body of the empirical literature suggests that the amount of care offered can differ to a great extent among doctors (Phelps, 2000).

Even if clear allocation rules could be established by policymakers, doctors would have an incentive to implement them only if they act as perfect agents on behalf of the government. Doctors’ behavior may be affected by their own preference and the way they are paid. For example, salaried hospital doctors may put considerable weight on a patient’s benefit as opposed to costs. In terms of hospital payment, fixed budgets give strongest incentive to ration because extra patients do not generate additional revenues. Prospective payment systems of the diagnosis-related groups (DRG) type give weaker incentives to ration because additional patients increase revenues. Providers however may still have an incentive to ration high-cost and unprofitable patients (Ellis, 1998). Generous fee-for-service (FFS) systems or cost reimbursement rules give weakest incentive to ration patients. The incentive to ration depends also critically on the degree of altruism of the doctors, highly altruistic doctors being more reluctant to ration patients. Altruism, therefore, plays a critical role in determining the design of incentive schemes (Ellis and McGuire, 1986; Chalkley and Malcomson, 1998): higher altruism typically requires lower powered incentive schemes because highly altruistic doctors paid with FFS arrangements are unlikely to exert much rationing.

In summary, direct rationing (explicit refusal of treatment to a patient) is a pervasive feature of many health systems. Rationing occurs both ‘within’ and ‘across’ treatments. Some rationing is implemented through government allocation rules, which define the ‘package’ covered by the public sector. However, most rationing is exerted by the doctors and will be based on their preference and their payment schemes.

A key issue in devolving the rationing role to doctors, is that ‘turning patients down’ can be an unpleasant activity. Without clear rules, doctors may feel reluctant to refuse treatment to patients with positive benefits. This will be exacerbated when the capacity constraint is tighter. Moreover, it is the doctors who will need to explain to patients why they are not offered treatment. They may also be liable for taking an unfair or unjust decision, and they may be at risk of breaking laws, which prohibit discrimination among patients.

If doctors are reluctant to explicitly ration patients, they may instead add the patients to a waiting list so that patients will have to wait before they receive treatment. This will generate a different type of rationing, which is described in Rationing by Waiting Times and Quality.

Rationing By Waiting Times And Quality

If direct rationing is difficult to implement, other forms of rationing are needed to allocate the limited capacity. If doctors simply refer for treatment all patients who could benefit (because they are reluctant to turn down patients with low benefit), a waiting list of patients authorized to receive care is likely to build up. If capacity is well below demand, patients’ wait could be long. Demand may exceed capacity in every period, implying that in the absence of some way of limiting waiting lists, the length of the list could grow indefinitely. Waiting times may have a rationing effect if they dissuade some patients to seek treatment. Instead of waiting, the patient may opt for the private sector. Longer wait times will discourage more patients. Another possibility is that while waiting the patient may become unfit for surgery or recover. The empirical evidence from the UK suggests that waiting times do act as a rationing device to equilibrate demand and supply. Most empirical studies find that demand for care is inelastic, and that the elasticity is approximately 0.1: a 10% increase in waiting times reduces demand by only 1% (Martin and Smith, 2003; Iversen and Siciliani, 2011 for a review). The result that demand is inelastic implies that waiting times exert only a moderate rationing effect and that low levels of supply will translate into long waiting times.

Although waiting times eliminate the need for doctors to explicitly refuse treatment to patients, rationing by waiting times can be an inefficient form of rationing compared to direct rationing. Long waiting times impose a cost on patients, which is not necessarily recovered by anyone else (Gravelle and Siciliani, 2008). Some (short) waiting times may generate efficiency savings if they reduce the chance of idle capacity (i.e., the probability that supply is not used; Iversen, 1997). However, these efficiencies are fully exploited with short average wait times (Siciliani et al., 2009). Waiting times are in the order of months for many procedures, well beyond those required for such efficiency savings to arise.

To mitigate the cost for the patients generated by waiting, several health systems prioritize patients on the list so that more severe patients wait less than less severe one. This is often done informally by doctors. Some governments have further reinforced this idea by developing policies (mostly in the form of guidelines), which encourage doctors to prioritize patients through the use of formal scoring systems (patients who score higher points in terms of severity, pain, and need wait less).

Other governments have discouraged the use of waiting time rationing through the development of targets or maximum waiting time guarantees, which introduce penalties for hospitals having many patients waiting for a long time. These policies may encourage increases in productivity but may also induce a switch from waiting time rationing to direct rationing.

Waiting times are not the only factor which induces patients to opt for the private sector. Other factors, like amenities and quality of care, also contribute to the choice between the public and the private sector. Besley and Coate (1991) provide a theory that suggests that a government (which has constraints on distributional tools) may find it optimal to distort quality downward to encourage the richer subset of the population to opt for the private sector. The shift of the rich to the public sector helps to bring the demand in the public sector in equilibrium with the limited supply as well as to redistribute income. This theory seems consistent with casual observation that public hospitals offer lower amenities compared to private one (patients may need to share rooms, have less privacy, and overall less comfort). This is not necessarily the case for clinical quality: whether it is higher or lower in public hospitals is less straightforward. On one hand, improving clinical quality is at the core of policy efforts in many publicly funded systems. On the other hand, public hospitals do face large demands, which induce hospitals to keep length of stay to a minimum, a dimension of lower quality (Barros and Siciliani, 2011).

As for direct (explicit) rationing, also under nonprice (waiting time and quality) rationing, the payment rule for providers may affect the incentive to vary quality and waiting times. For example, a DRG-type hospital payment system may induce creaming or skimping, i.e., the incentive to raise quality for profitable patients and reduce quality for unprofitable ones (Ellis, 1998). DRG systems also induce providers to treat additional patients, which should translate into lower waiting times.

In summary, rationing by waiting and quality is also a pervasive feature of many publicly funded health systems. Compared to direct rationing, these are (to some degree) inefficient because for a given capacity they reduce patients’ welfare. However, they release doctors from the responsibility of directly rationing patients.

Price Rationing

Demand for health care can also be rationed through prices. In many publicly funded systems, this takes usually the form of a copayment or a coinsurance rate: the patient is asked to pay a fee or a proportion of the medical expenses. The idea is to make the patient cost conscious, who in turn demand less health care and in this way contain moral hazard (excessive consumption).

Countries vary in the use and design of copayments with some countries making more use than others. With few exceptions copayments remain low in publicly funded systems. Large increase in copayments would also imply a significant reduction in the benefit from being insured against the cost of illness. According to theory, the optimal copayment should be designed such that it efficiently trades off the risk spreading benefits of a lower price against the ex-post efficiency benefits of a higher price (closer to marginal cost; Zeckhauser, 1970). The theory implies that the optimal copayment is positively related to the elasticity of demand: copayments should be higher when the elasticity is higher. Copayments are indeed observed for dental care, ophthalmology care, and drugs where the elasticity is arguably larger. They are more rarely observed for inpatient or surgical care, which is free of charge in most (though not all) countries, due plausibly to its more inelastic demand.

Several empirical studies have estimated the elasticity of demand. Early studies found a wide range of elasticities’ estimates, which could vary between -0.1 and -2.1 (Cutler and Zeckhauser, 2000). These estimates are potentially affected by selection bias if sicker individuals choose insurance plans with lower copayments and also demand more care. The Rand experiment eliminates such potential bias by randomizing individuals in different plans (Manning, et al., 1987). It suggests that demand is inelastic with an overall elasticity of approximately -0.1 or -0.2. The article by Sinaiko (2012) updates this literature.

The idea that copayments make patients cost conscious relies on the belief that patients are able to influence the choice on the care. Given the asymmetry of information, which characterizes the patient–doctor relationship, the choice of care is (to say the least) mediated through the doctor. Copayments may have little impact on demand of care if doctors base their recommendation only on medical ground and ignore the financial implications for the patient. Copayments will instead play a role only if patients are well informed (which may be the case for some conditions) or if doctors internalize patients’ disutility from higher prices. Behavior of the doctors will also be influenced by their financial incentives.

Both waiting times and copayments ration demand. Compared to rationing by waiting, copayments have the advantage that the cost imposed on the patient is recovered by the provider or the insurer in the form of additional revenues. Furthermore, if consumers accurately appreciate the value of health care, a price rations out low-value uses. However, copayments raise equity issues (if they are not income tested) if poor patients are deterred from demanding care compared to richer patients, a criticism that is often raised against their excessive use within publicly funded systems. Finally, the role played by copayments in rationing demand is mitigated by the asymmetry of information, which characterizes the patient– doctor relationship.


Three different types of rationing have been discussed: (1) direct rationing; (2) rationing by waiting time and quality; and (3) price rationing. Direct rationing can in principle allocate care efficiently by giving care to patients with highest benefit–cost ratio compatibly with the capacity constraint. This could be implemented by listing explicitly the treatments covered by public insurance (as well as those not covered) and/or by asking doctors to ration according to a set of established optimality criteria (potential health gains, heath status, and costs).

In practice, drawing an explicit list of eligible treatments is a complex and costly exercise and there are limits to this approach. Delegating the rationing role to doctors is inevitable. However, doctors may vary in the application of a set of rationing rules due, for example, to different interpretation of such rules, generating variations in clinical practice. Moreover, doctors themselves may be reluctant to exercise to a great extent the rationing role because they may find difficult or unpleasant to turn patients down and they may also be held liable for mistakes. This generates the scope for other forms of rationing.

If all potential patients who demand care are referred for treatment, a waiting list will quickly build up with (implicit) waiting time rationing replacing explicit direct rationing. Waiting times may ration public patients by inducing some of them to seek care in the private sector. Similarly, offering limited amenities in the public sector may shift some patients to the private sector. Compared with direct rationing, these forms of implicit rationing will however come at the cost of lower welfare for those patients who seek care in the public sector due to the waiting time or other costs imposed to them.

An alternative to rationing by waiting is to ration by price through the introduction of copayments or coinsurance rates. Its use can be a useful complement to other forms of rationing but needs to be traded off with the lower insurance coverage against the cost of illness. Moreover, large copayments (if not income tested) raise equity issues because poor patients may be discouraged from utilizing care. This can contrast the principle that access to care should not be based on the ability to pay, which is at the core of many publicly funded health systems.


  1. Barros, P. P. and Siciliani, L. (2011). Public-private interface. In Pauly, M., McGuire, T. and Barros, P. P. (eds.) Handbook of health economics, vol. 2, ch. 15, pp. 927–1002. Oxford, UK: Elsevier.
  2. Besley, T. and Coate, S. (1991). Public provision of private goods and the redistribution of income. American Economic Review 81, 979–984.
  3. Chalkley, M. and Malcomson, J. (1998). Contracting for health services when patient demand does not reflect quality. Journal of Health Economics 17, 1–19.
  4. Cutler, D. and Zeckhauser, R. (2000). The anatomy of health insurance. In Culyer, A. J. and Newhouse, J. P. (eds.) Handbook of health economics, ch. 11, pp. 563–643. Amsterdam, The Netherlands: Elsevier.
  5. Drummond, M. F., O’Brien, B. J., Schulper, M., Stoddart, G. L. and Torrance, G. W. (2005). Methods for the economic evaluation of health care programmes. Oxford, UK: Oxford University Press.
  6. Ellis, R. P. (1998). Creaming, skimping, and dumping: Provider competition on the intensive and extensive margins. Journal of Health Economics 17(5), 537–555.
  7. Ellis, R. P. and McGuire, T. G. (1986). Provider behavior under prospective reimbursement. Journal of Health Economics 5, 129–151.
  8. Gravelle, H. and Siciliani, L. (2008). Optimal quality, waits, and charges in health insurance. Journal of Health Economics 27(3), 663–674.
  9. Hauck, K., Shaw, R. and Smith, P. C. (2002). Reducing avoidable inequalities in health: A new criterion for setting health care capitation payments. Health Economics 11(8), 667–677.
  10. Iversen, T. (1997). The effect of private sector on the waiting time in a National Health Service. Journal of Health Economics 16, 381–396.
  11. Iversen, T. and Siciliani, L. (2011). Non-price rationing and waiting times. In Gleid, S. and Smith, P. C. (eds.) Oxford handbook of health economics, ch. 27, pp. 649–670. Oxford, UK: Oxford University Press.
  12. Manning, W. G., Newhouse, J., Duan, N., et al. (1987). Health insurance and the demand for medical care: Evidence from a randomized experiment. American Economic Review 77(3), 251–277.
  13. Martin, S. and Smith, P. C. (2003). Using panel methods to model waiting times for National Health Service surgery. Journal of the Royal Statistical Society 166(Part 2), 1–19.
  14. McGuire, T. (2000). Physician agency. In Culyer, A. J. and Newhouse, J. P. (eds.) Handbook of health economics, vol. 1, ch. 9, pp. 461–536. Amsterdam, The Netherlands: Elsevier
  15. Phelps, C. E. (2000). Information diffusion and best practice adoption. In Culyer, A. J. and Newhouse, J. P. (eds.) Handbook of health economics, ch. 5, pp. 223–264. Amsterdam, The Netherlands: Elsevier.
  16. Siciliani, L., Stanciole, A. and Jacobs, R. (2009). Do waiting times reduce costs? Journal of Health Economics 28(4), 771–780.
  17. Sinaiko, A. (2012). Studies of Demand Response Since the HIE, this Encyclopedia.
  18. Tengs, T. (1996). An evaluation of Oregon’s Medicaid rationing algorithms. Health Economics 5(3), 171–181.
  19. Wagstaff, A. and van Doorslaer, E. (2000). Equity in health care finance and delivery. In Culyer, A. J. and Newhouse, J. P. (eds.) Handbook of health economics, vol. 1, ch. 34, pp. 1803–1862. Amsterdam, The Netherlands: Elsevier.
  20. Zeckhauser, R. (1970). Medical insurance: A case study of the tradeoff between risk spreading and appropriate incentives. Journal of Economic Theory 2, 10–26.
Quality Reporting and Demand