Welfare Criterion

Welfare Criterion in Health Economics

A welfare criterion in health economics is a decision rule taken from welfare economics that states when one allocation of resources or health should count as socially better than another. It turns evidence about who gains and who loses from a vaccination programme, a sugar tax or a new NHS service into a social verdict. Every criterion carries a value judgement about whose gains count and how gains to some people are set against losses to others. The choice between the potential Pareto test behind cost-benefit analysis and the health-maximising rule behind cost-effectiveness analysis for NICE sits at the centre of the welfarist and extra-welfarist debate in health economics.

What a welfare criterion has to settle

Welfare economics aims to make defensible statements about whether some states of the world are socially preferable to others. Brouwer, Culyer, van Exel and Rutten (2008) quote Boadway and Bruce's account that the welfare economist judges a policy by explicitly stated ethical criteria and not by the analyst's own values. A welfare criterion is the formal version of those stated criteria.

Any criterion fixes two things. The first is the information it uses, which Brouwer and colleagues call the evaluative space. In welfarist economics that space holds individual utility only, while extra-welfarist frameworks can put health or capabilities in its place. The second is the rule for combining gains and losses across people, and Brouwer and colleagues note that even the Pareto rule could be applied to health instead of utility. Fleurbaey's entry on normative economics in the Stanford Encyclopedia of Philosophy adds that some criteria are partial, failing to rank many pairs of alternatives, while a well-defined social welfare function gives complete and consistent rankings.

The Pareto criterion and the limits of unanimity

Hausman's entry on the philosophy of economics explains that if some people prefer outcome S to outcome R and nobody prefers R to S, then S is a Pareto improvement over R, and the conclusion needs no comparison of one person's welfare with another's. A state with no possible Pareto improvement is Pareto efficient. Hausman stresses that a distribution in which millions starve can still be Pareto efficient.

The weakness is that the criterion is usually silent. Fleurbaey notes that most public policy changes hurt some groups for the benefit of others, so the Pareto principle generally gives no verdict, and Brouwer and colleagues add that choosing among the many Pareto optima is impossible with the Pareto criterion alone. Health care makes the problem acute. When a new drug is funded from a fixed NHS budget, the opportunity cost falls on patients whose care is displaced, so almost no funding decision is a Pareto improvement.

Compensation tests and the potential Pareto improvement

Fleurbaey describes how Nicholas Kaldor and John Hicks, both writing in 1939, extended the Pareto principle through compensation tests. Under Kaldor's criterion a change is an improvement if, after it, the gainers could compensate the losers. Under Hicks's criterion the losers could not compensate the gainers for forgoing the change.

Hausman sets out the version behind cost-benefit analysis. A change is a potential Pareto improvement when the money the winners would pay for it exceeds the money the losers would need in compensation, and the excess is the net benefit.

$$NB = \sum_{i=1}^{n} CV_i > 0$$

where $NB$ is the net benefit of the change, $n$ is the number of people affected and $CV_i$ is the compensating variation of person $i$, positive for a gainer's willingness to pay and negative for the compensation a loser would require.

Hausman points out that the compensation is entirely hypothetical and that the justice of the resulting gains and losses is irrelevant to the test. Brouwer and colleagues call the potential Pareto criterion a serious compromise of the Pareto principle, because it gives an equal unit weight to each money gain or loss whoever receives it. Willingness to pay also depends on wealth, so Hausman concludes that cost-benefit analysis weights the preferences of the rich more than those of the poor. An unweighted cost-benefit analysis of moving a clinic therefore makes a distributional judgement that nobody has stated.

The tests also have a logical flaw. Fleurbaey reports that the same criterion can declare A better than B and B better than A, and that Tibor Scitovsky's 1941 proposal to combine the two tests does not prevent intransitive social judgements. Hausman traces the same reversal to Scitovsky (1941) and Samuelson (1950).

Social welfare functions as complete criteria

A social welfare function gives a complete criterion by stating how individual welfare levels combine. Fleurbaey writes the Bergson and Samuelson form in general terms.

$$W = W(U_1(x), \ldots, U_n(x))$$

where $W$ is social welfare, $x$ is the social state being judged and $U_i(x)$ is the utility of individual $i$ in that state.

Sum-utilitarianism ranks states by total welfare. The maximin criterion, which Fleurbaey links to Rawls's difference principle, puts absolute priority on the worst-off, and social choice theory has shown how to build a continuum of social welfare functions between the two.

Fleurbaey also notes that more sophisticated cost-benefit analysis avoids the flaws of compensation tests by using weighted sums of willingness to pay. HM Treasury's Green Book (2026) explains that an extra GBP 1,000 is worth more to a low-income household than to a high-income one, and its distributional weights give a higher social value to benefits for lower-income households. The weights rest on an estimated elasticity of marginal utility of consumption of 1.25, and weighted estimates must be presented beside unweighted ones. Equity-weighted analysis applies explicit weights to health gains in the same spirit.

Health maximisation as an extra-welfarist criterion

UK health technology assessment uses a different criterion. Brouwer and colleagues explain that under extra-welfarism uncompensated changes, even ones that would fail the potential Pareto test, may be judged social improvements by appeal to further ethical criteria. In health care the evaluative space has focused on health, often measured in QALYs, because QALYs measure health and not because they capture utility. They link the role of policy makers as a source of values to the decision-making approach of Sugden and Williams (1978).

Claxton, Walker, Palmer and Sculpher (2010) describe the view that in a democracy with universal health care, such as the UK NHS, the health care budget can be seen as a legitimate expression of social values. Cost-effectiveness analysis then supports decision makers trying to maximise health subject to that budget. The cost-effectiveness threshold expresses the shadow price of the budget constraint, and a technology is accepted when the health it gains exceeds the health it displaces.

$$NHB = \Delta h - \frac{\Delta c}{k} > 0$$

where $NHB$ is the net health benefit in QALYs, $\Delta h$ is the health gained, $\Delta c$ is the additional cost to the health care budget and $k$ is the threshold in GBP per QALY.

Claxton and colleagues show that this rule is equivalent to an ICER below $k$ and to a positive net monetary benefit. The rules are complete and appropriate if the social objective is to maximise health, all costs fall on the health care sector and all benefits of social value take the form of health. For wider effects they suggest a compensation test, which a technology passes if the wider economy could compensate the health service for the lost net health benefit and still gain. The NICE manual (PMG36) states in section 6.3.1 that no precise maximum acceptable ICER is used. Below GBP 25,000 per QALY the decision normally rests on the cost-effectiveness estimate, and above GBP 35,000 the committee needs an increasingly stronger case (sections 6.3.4 to 6.3.8). The Green Book uses a separate value of GBP 70,000 per QALY in 2020-21 prices, so UK health technology appraisal and wider public appraisal value a QALY in different ways.

Worked example with one clinic move judged by several criteria

All numbers here are illustrative. A health board proposes moving a specialist clinic from Town A, where incomes are lower, to Town B, at an extra NHS cost of GBP 1.0 million.

1. Money measures. Residents of Town A would need GBP 2.0 million in compensation to accept the move, so their compensating variation is minus GBP 2.0 million. Residents of Town B would pay GBP 3.0 million for it. Both figures include each town's tax share of the extra cost, so the net benefit is 3.0 minus 2.0, or GBP 1.0 million.

2. Pareto criterion. Town A is worse off, so the move is not a Pareto improvement and the criterion gives no ranking.

3. Kaldor and Hicks tests. Town B could pay Town A GBP 2.0 million and still be GBP 1.0 million better off, so the move passes Kaldor's test. Assuming no income effects, Town A would pay at most GBP 2.0 million to stop it, less than the GBP 3.0 million Town B places on it, so it passes Hicks's test too.

4. Weighted criterion. With illustrative weights of 1.4 for Town A and 0.7 for Town B, the weighted sum is minus 2.0 × 1.4 plus 3.0 × 0.7, which is minus 2.8 plus 2.1, or minus GBP 0.7 million. The move fails.

5. Compensation actually paid. If Town B paid Town A GBP 2.0 million, Town A would be no worse off and Town B would keep GBP 1.0 million, so the move would become a Pareto improvement.

6. Health maximisation. Suppose the move costs Town A 20 QALYs and gains Town B 70 QALYs, a net 50 QALYs. The ICER is 1,000,000 / 50 = GBP 20,000 per QALY. Health displaced at GBP 25,000 per QALY is 1,000,000 / 25,000 = 40 QALYs, so the net health benefit is 50 minus 40, or 10 QALYs. At GBP 35,000 per QALY the health displaced is 1,000,000 / 35,000 = 28.57 QALYs and the net health benefit is 21.43 QALYs. The move passes.

7. Maximin. If Town A already has the poorer health, the move lowers the worst-off group's position and maximin rejects it, whatever the total gain.

The evidence is identical in every row and only the criterion changes, so a verdict reported without its criterion hides the most contestable step.

How a welfare criterion differs from neighbouring ideas

A welfare criterion ranks states, while Pareto efficiency describes a property of a state under one particular rule. A decision rule in cost-effectiveness analysis, such as accepting a positive net health benefit, is an operational form of the health-maximising criterion, and Claxton and colleagues tie it explicitly to that objective. A social welfare function is one way of writing a complete criterion, and a cost-effectiveness threshold is a parameter inside a criterion.

Brouwer and colleagues note that since Arrow's work on the impossibility of a non-dictatorial social welfare function, economists' thinking on gains in social welfare has stayed dominated by the Pareto improvement.

Limitations and common misreadings of welfare criteria

The commonest error is to read a positive net benefit as proof that everyone gains. Under the potential Pareto test the losers stay losers unless compensation is paid, and Fleurbaey calls the bare possibility of compensation a meagre consolation to them, adding that such criteria are typically biased in favour of the rich.

A second error is to treat the choice of criterion as technical. Picking the Pareto test, a weighted sum, maximin or health maximisation is itself a value judgement, which is why Boadway and Bruce, as quoted by Brouwer and colleagues, ask for the criteria to be explicitly stated. A third is to assume compensation tests always rank consistently, when Fleurbaey shows they can contradict themselves.

Health maximisation has its own limits. Claxton and colleagues show that its decision rules need generalising once important effects fall outside the health sector, and unweighted QALY maximisation is neutral about who gains. When decision-making modifiers are applied, PMG36 section 6.3.2 asks for net health benefits at GBP 25,000 and GBP 35,000 per QALY with and without the QALY weighting, so that the effect on opportunity costs can be seen.

Sources

• Brouwer WBF, Culyer AJ, van Exel NJA, Rutten FFH. 2008. Welfarism vs. extra-welfarism. Journal of Health Economics 27(2): 325-338. https://doi.org/10.1016/j.jhealeco.2007.07.003 (author copy: https://eprints.whiterose.ac.uk/id/eprint/3789/)
• Claxton K, Walker S, Palmer S, Sculpher M. 2010. Appropriate perspectives for health care decisions. CHE Research Paper 54. York: Centre for Health Economics, University of York. https://www.york.ac.uk/media/che/documents/papers/researchpapers/rp54_appropriate_perspectives_for_health_care_decisions.pdf
• Fleurbaey M. 2026. Normative economics and economic justice. Stanford Encyclopedia of Philosophy, first published 28 May 2004, substantive revision 6 May 2026. https://plato.stanford.edu/entries/economic-justice/
• Hausman DM. 2018. Philosophy of economics. Stanford Encyclopedia of Philosophy, first published 12 September 2003, substantive revision 4 September 2018. https://plato.stanford.edu/entries/economics/
• HM Treasury. 2026. The Green Book (2026). Updated 28 October 2026. https://www.gov.uk/government/publications/the-green-book--2/the-green-book-2026
• National Institute for Health and Care Excellence. 2022. NICE technology appraisal and highly specialised technologies guidance: the manual (PMG36). Published 31 January 2022, last updated 13 October 2026. https://www.nice.org.uk/process/pmg36