ABSTRACTS
How to Measure Value of a Statistical Life (VOSL) Under Prospect Theory
Han Bleichrodt (University of Alicante)
The value of a statistical life (VOSL) helps to make better policy decisions, but it is subject to biases. Carthy et al. (1998) proposed a method that chains contingent valuation and the standard gamble (SG) used in quality of life measurement. It avoids biases, but a direct and an indirect variant of the method led to different results. Sánchez et al. (2021) showed how these differences could be reduced by a distinct chaining. However, they find that the VOSL depends on the health state used in the SG. Both studies analyze survey responses using expected utility (EU), which people usually violate. I show how to derive the VOSL under the descriptively more realistic prospect theory (PT). A re-analysis of the data of Sánchez et al shows that their inconsistencies reduce under prospect theory. Mean VOSL drops to Euro 300K-600K, lower than what most governments use. A pilot experiment in which we measure prospect theory's parameters for each respondent finds few differences between EU and PT. Estimated VOSL lies between Euro1.5M and Euro 2.25M.
Temporal Separability and the Value of Longevity
John Broome (ANU and University of Oxford)
Economic development has lengthened people’s lives. We generally think this is one of its greatest benefits. But this idea turns out to be hard to reconcile with something else we generally think: that the value of the world can be divided into the separate values it has at each particular time. I shall explain the problem and explore some possible ways of overcoming it. One possible way is associated with the World Health Organization’s approach to measuring the burden of disease. I shall argue that this way fails. Another possible way is associated with neutral-level (or critical-level) utilitarianism. In principle this could succeed, but it requires a radical revision of our valuations. I shall argue that, in the end, we shall probably have to abandon temporal separability. Its implications are too demanding.
A 'Statistical' Approach to Valuing the Environment: Can We Go There?
Susan Chilton (Newcastle University)
Clear, economic theoretic guidance is available for policymakers to value our safety and longevity. This contrasts with that available to the wide variety of parties who exercise decision making over the environment. This presentation posits that environmental decision making in general (and benefit-cost studies in particular) are subject to a so-called ‘environmental valuation’ trap caused at least in part by the fact that non-market environmental valuation studies are too costly to be widely used and are too few to infer reliable generic values. Drawing on experience in estimating the public value of safety improvements, this presentation proposes an alternative approach, at least for the class of environmental goods that fall into the following category: the class of environmental assets that consist of many individual sites across a country and, ex ante, we expect on average one (for example) industrial accident e.g. factory explosions/leaks pollute rivers/waterbodies, major chemical spill pollutes surrounding air or soil in the forthcoming period but we do not know where in the country that will be. A safety analogue might be: we expect a rail fatality in the next year, but we do not know where on the network.
I would like to present a policy valuation measure/framework – Value of a Major Adverse Event to the Environment (VMAEE) – which values reductions in risk of major (moderate/minor) adverse environmental outcomes which also incorporates the principles of Total Economic value from environmental economics. This could in principle deliver the flexibility of the VSL approach to environmental valuation. The conceptual foundation has already been set out (Clough et al., 2018) but I would like to drill down with a VSL audience as to whether further refinements are necessary or if a fundamental problem exists in extending the VSL framework in this manner. [Ref.: Clough, P., Chilton, S., Jones-Lee, M., Metcalf, H. (2018). A New Approach to Environmental Valuation for New Zealand. Policy Quarterly, 14(2), 50-57]
What’s your health worth? A 30-year journey through valuing health in monetary terms
Cam Donaldson (Glasgow Caledonian University and ANU)
In this presentation, we will explore a body of work on the monetary valuation of health – especially the value of quality adjusted life years (QALYs). In particular, Cam will: outline the origins of the notion of the value of a QALY and how it then developed into two major empirical research projects, the Social Value of a QALY (SVQ) and European Value of a QALY (EuroVaQ); describe the methods and results of SVQ and EuroVaQ; and briefly discuss how this work has evolved through to the current day, both in terms of other applications and challenges going forward.
The use of stated preference methods for valuing health outcomes: a couple of critical issues
Dorte Gyrd-Hansen (University of Southern Denmark)
The design of stated preference experiments, and in particular the framing of the experiment, determines which elements of utility are included in the valuation task when seeking to elicit preferences for health care services and the associated health gains, including the distribution of health gains across patient groups. Choice of payment vehicle, and whether one applies a ex ante or ex post valuation perspective will have implications for which sources of value are captured in our willingness-to-pay estimate. The aim of this talk is to raise some critical issues that should perhaps be more acknowledged within the stated preference research field. The presentation will focus on two studies, where each paper illustrates a central problem. One study focuses on the link between choice of payment vehicle and the role of altruism on valuations (Simonsen NF, Kjær T, Gyrd-Hansen D, 2021. Pure altruism and misjudgment. A bad combination? Journal of Health Economics). The paper questions the use of non-voluntary payments as payment vehicle if respondents are pure altruists. The other paper focuses on the potential limitation of eliciting distributional preferences from the commonly applied ex post decision maker perspective. Results suggest that the ex ante perspective (insurance perspective) is likely to favor more equal distributions of health outcomes, and that the standard approaches to eliciting equity preferences may therefore possibly lead to resource allocations that are not welfare optimizing.
Quandaries in valuing mortality risk
James Hammitt (Harvard University)
The conventional approach to valuing mortality risk has been developed over the last half century and much progress has been achieved. However, there remain a number of significant quandaries, from both the individual and social perspectives. From the individual perspective, the monetary value of a non-marginal change in risk implied by the conventional model seems implausibly large. Although the marginal willingness to pay for risk reduction decreases as more is bought, the decrease is not sharp enough to avoid near impoverishment for modest risk reductions, which seems implausible. From the social perspective, an individual’s VSL may not be a good guide to how much should be spent on reducing her risk; while an individual facing high mortality risk may have a low opportunity cost of spending and a large VSL, the social opportunity cost is much larger. Moreover, lack of an interpersonally comparable utility measure limits analysts’ ability to compare benefits and harms across individuals and to determine whether an intervention constitutes a social improvement.
Why weight? Benefit-cost analysis and welfare
Kirsten Mann (ANU) and Katie Steele (ANU)
Traditional benefit-cost analysis (BCA), which values the costs and benefits of policies in terms of willingness to pay/accept in money, is widely acknowledged to have serious drawbacks. For example, the benefits and costs of a policy that accrue to a richer person will be larger on this measure purely because of the diminishing marginal utility of consumption. A recent literature has proposed using healthy years of life as the common currency for BCA, either directly as the numeraire or indirectly as the basis for weighing the monetary inputs. This makes sense: while the connection between money and welfare is non-linear and mediated through other goods, ‘healthy life years’ are surely as close as we can get to capturing welfare itself. Critics of this approach, however, have argued that using healthy life years as the currency for BCA has similar problems to using money. In response to recent work in this vein by Fleurbaey and Hammitt (2024), we argue that whether or not healthy life years are a good guide to the welfare costs and benefits of a policy depends on assumptions about the form of the individual welfare function. There is a further question about how individual welfare contributes to social welfare, but we claim that, in practice, the two should be clearly distinguished. Along the way, we consider the implications of our analysis for an analogous debate about disability discrimination in healthcare allocation.
Setting optimal test thresholds: a consumer choice model and an experimental test
Morgan Benson (Newcastle University), Peter McMeekin (Northumbria University), Jytte Seested Nielsen (Newcastle University) and John Wildman (Newcastle University)
Millions of individuals every day, globally, undertake screening and diagnostic tests and millions more are invited to take part in such tests. Such tests are seen as crucial to improving population health, increasing the efficient use of health resources, and supporting a healthy working population. We explore preferences regarding the trade-off between sensitivity and specificity in screening and diagnostic tests required for the choice of test threshold. For a given test, increasing sensitivity, detecting more true positives, comes at the expense of decreasing specificity, giving more false positives. We present a theoretical model of consumer choice for sensitivity and specificity. We devise a novel experiment that mimics threshold selection for a given test, to test our model predictions. Results show that respondents' choices are equally determined by disease characteristics and individual preferences. On average, respondents choose higher sensitivity, lower specificity thresholds than the cost-effective threshold. Our findings have implications for the demand/uptake of tests and the importance of setting appropriate test thresholds.
Valuing health in Norway: The policy process behind sensible criteria
Jan Abel Olsen (UiT – The Arctic University of Norway)
Norwegian guidelines for Health Technology Assessment (HTA) are based on priority setting criteria recommended by Government appointed commissions. A key characteristic of these commissions is the diverse background and strong expertise among their around 15 members including; professors, patient organisation representatives, health professionals, previous MPs. A commission report is typically followed by public hearings and White Paper to the Parliament.
After the report from the Norheim Commission in 2014 their suggested equity criterion was discussed by an expert group and later revised in a subsequent White Paper. Three key criteria: benefits; resources; severity, have now been operationalised into detailed guidelines for equity weighted cost-effectiveness analysis to help reimbursement decisions. Of particular interest is the application of specific equity weights, depending on the magnitude of the absolute QALY-shortfall associated with a disease.
In this presentation, Jan Abel will briefly outline the Norwegian priority setting debate from the first commission in 1987 up until the most recent expert groups in 2024. As a result of the increasing use of HTAs, a crucial topic has been: How best to account for severity and equity in a transparent and consistent way. By use of diagrams, Jan Abel will illustrate different views on the equalisandum (‘Equality of what?’) and contrast the ‘End-of-Life’ criterion with the Norwegian policy of reducing inequalities in prospective health. Based on his involvement as commission member and subsequent advisor on how to operationalize the three criteria into specific guidelines, Jan Abel will reflect on the policy process behind what he considers to be sensible official guidelines for valuing health.
How much is a baby worth?
John Quiggin (University of Queensland)
The global decline in fertility rates has produced considerable concern based on the largely unquestioned assumption that a declining population is economically harmful. In particular, concern has been expressed about the possibility that there will be too few young people to care for an aging population. This paper begins with the observation that the caring work required to raise a child to adulthood is far greater than the end-of-life care typically required by old people. However, most childcare work is voluntary, contributed by parents who have actively chosen to bear and raise children. We lack the economic tools needed to analyse this work, without which it is very difficult to evaluate policies that affect fertility and, particularly, pro-natalist policies aimed at raising fertility.
How to Measure Value of a Statistical Life (VOSL) Under Prospect Theory
Han Bleichrodt (University of Alicante)
The value of a statistical life (VOSL) helps to make better policy decisions, but it is subject to biases. Carthy et al. (1998) proposed a method that chains contingent valuation and the standard gamble (SG) used in quality of life measurement. It avoids biases, but a direct and an indirect variant of the method led to different results. Sánchez et al. (2021) showed how these differences could be reduced by a distinct chaining. However, they find that the VOSL depends on the health state used in the SG. Both studies analyze survey responses using expected utility (EU), which people usually violate. I show how to derive the VOSL under the descriptively more realistic prospect theory (PT). A re-analysis of the data of Sánchez et al shows that their inconsistencies reduce under prospect theory. Mean VOSL drops to Euro 300K-600K, lower than what most governments use. A pilot experiment in which we measure prospect theory's parameters for each respondent finds few differences between EU and PT. Estimated VOSL lies between Euro1.5M and Euro 2.25M.
Temporal Separability and the Value of Longevity
John Broome (ANU and University of Oxford)
Economic development has lengthened people’s lives. We generally think this is one of its greatest benefits. But this idea turns out to be hard to reconcile with something else we generally think: that the value of the world can be divided into the separate values it has at each particular time. I shall explain the problem and explore some possible ways of overcoming it. One possible way is associated with the World Health Organization’s approach to measuring the burden of disease. I shall argue that this way fails. Another possible way is associated with neutral-level (or critical-level) utilitarianism. In principle this could succeed, but it requires a radical revision of our valuations. I shall argue that, in the end, we shall probably have to abandon temporal separability. Its implications are too demanding.
A 'Statistical' Approach to Valuing the Environment: Can We Go There?
Susan Chilton (Newcastle University)
Clear, economic theoretic guidance is available for policymakers to value our safety and longevity. This contrasts with that available to the wide variety of parties who exercise decision making over the environment. This presentation posits that environmental decision making in general (and benefit-cost studies in particular) are subject to a so-called ‘environmental valuation’ trap caused at least in part by the fact that non-market environmental valuation studies are too costly to be widely used and are too few to infer reliable generic values. Drawing on experience in estimating the public value of safety improvements, this presentation proposes an alternative approach, at least for the class of environmental goods that fall into the following category: the class of environmental assets that consist of many individual sites across a country and, ex ante, we expect on average one (for example) industrial accident e.g. factory explosions/leaks pollute rivers/waterbodies, major chemical spill pollutes surrounding air or soil in the forthcoming period but we do not know where in the country that will be. A safety analogue might be: we expect a rail fatality in the next year, but we do not know where on the network.
I would like to present a policy valuation measure/framework – Value of a Major Adverse Event to the Environment (VMAEE) – which values reductions in risk of major (moderate/minor) adverse environmental outcomes which also incorporates the principles of Total Economic value from environmental economics. This could in principle deliver the flexibility of the VSL approach to environmental valuation. The conceptual foundation has already been set out (Clough et al., 2018) but I would like to drill down with a VSL audience as to whether further refinements are necessary or if a fundamental problem exists in extending the VSL framework in this manner. [Ref.: Clough, P., Chilton, S., Jones-Lee, M., Metcalf, H. (2018). A New Approach to Environmental Valuation for New Zealand. Policy Quarterly, 14(2), 50-57]
What’s your health worth? A 30-year journey through valuing health in monetary terms
Cam Donaldson (Glasgow Caledonian University and ANU)
In this presentation, we will explore a body of work on the monetary valuation of health – especially the value of quality adjusted life years (QALYs). In particular, Cam will: outline the origins of the notion of the value of a QALY and how it then developed into two major empirical research projects, the Social Value of a QALY (SVQ) and European Value of a QALY (EuroVaQ); describe the methods and results of SVQ and EuroVaQ; and briefly discuss how this work has evolved through to the current day, both in terms of other applications and challenges going forward.
The use of stated preference methods for valuing health outcomes: a couple of critical issues
Dorte Gyrd-Hansen (University of Southern Denmark)
The design of stated preference experiments, and in particular the framing of the experiment, determines which elements of utility are included in the valuation task when seeking to elicit preferences for health care services and the associated health gains, including the distribution of health gains across patient groups. Choice of payment vehicle, and whether one applies a ex ante or ex post valuation perspective will have implications for which sources of value are captured in our willingness-to-pay estimate. The aim of this talk is to raise some critical issues that should perhaps be more acknowledged within the stated preference research field. The presentation will focus on two studies, where each paper illustrates a central problem. One study focuses on the link between choice of payment vehicle and the role of altruism on valuations (Simonsen NF, Kjær T, Gyrd-Hansen D, 2021. Pure altruism and misjudgment. A bad combination? Journal of Health Economics). The paper questions the use of non-voluntary payments as payment vehicle if respondents are pure altruists. The other paper focuses on the potential limitation of eliciting distributional preferences from the commonly applied ex post decision maker perspective. Results suggest that the ex ante perspective (insurance perspective) is likely to favor more equal distributions of health outcomes, and that the standard approaches to eliciting equity preferences may therefore possibly lead to resource allocations that are not welfare optimizing.
Quandaries in valuing mortality risk
James Hammitt (Harvard University)
The conventional approach to valuing mortality risk has been developed over the last half century and much progress has been achieved. However, there remain a number of significant quandaries, from both the individual and social perspectives. From the individual perspective, the monetary value of a non-marginal change in risk implied by the conventional model seems implausibly large. Although the marginal willingness to pay for risk reduction decreases as more is bought, the decrease is not sharp enough to avoid near impoverishment for modest risk reductions, which seems implausible. From the social perspective, an individual’s VSL may not be a good guide to how much should be spent on reducing her risk; while an individual facing high mortality risk may have a low opportunity cost of spending and a large VSL, the social opportunity cost is much larger. Moreover, lack of an interpersonally comparable utility measure limits analysts’ ability to compare benefits and harms across individuals and to determine whether an intervention constitutes a social improvement.
Why weight? Benefit-cost analysis and welfare
Kirsten Mann (ANU) and Katie Steele (ANU)
Traditional benefit-cost analysis (BCA), which values the costs and benefits of policies in terms of willingness to pay/accept in money, is widely acknowledged to have serious drawbacks. For example, the benefits and costs of a policy that accrue to a richer person will be larger on this measure purely because of the diminishing marginal utility of consumption. A recent literature has proposed using healthy years of life as the common currency for BCA, either directly as the numeraire or indirectly as the basis for weighing the monetary inputs. This makes sense: while the connection between money and welfare is non-linear and mediated through other goods, ‘healthy life years’ are surely as close as we can get to capturing welfare itself. Critics of this approach, however, have argued that using healthy life years as the currency for BCA has similar problems to using money. In response to recent work in this vein by Fleurbaey and Hammitt (2024), we argue that whether or not healthy life years are a good guide to the welfare costs and benefits of a policy depends on assumptions about the form of the individual welfare function. There is a further question about how individual welfare contributes to social welfare, but we claim that, in practice, the two should be clearly distinguished. Along the way, we consider the implications of our analysis for an analogous debate about disability discrimination in healthcare allocation.
Setting optimal test thresholds: a consumer choice model and an experimental test
Morgan Benson (Newcastle University), Peter McMeekin (Northumbria University), Jytte Seested Nielsen (Newcastle University) and John Wildman (Newcastle University)
Millions of individuals every day, globally, undertake screening and diagnostic tests and millions more are invited to take part in such tests. Such tests are seen as crucial to improving population health, increasing the efficient use of health resources, and supporting a healthy working population. We explore preferences regarding the trade-off between sensitivity and specificity in screening and diagnostic tests required for the choice of test threshold. For a given test, increasing sensitivity, detecting more true positives, comes at the expense of decreasing specificity, giving more false positives. We present a theoretical model of consumer choice for sensitivity and specificity. We devise a novel experiment that mimics threshold selection for a given test, to test our model predictions. Results show that respondents' choices are equally determined by disease characteristics and individual preferences. On average, respondents choose higher sensitivity, lower specificity thresholds than the cost-effective threshold. Our findings have implications for the demand/uptake of tests and the importance of setting appropriate test thresholds.
Valuing health in Norway: The policy process behind sensible criteria
Jan Abel Olsen (UiT – The Arctic University of Norway)
Norwegian guidelines for Health Technology Assessment (HTA) are based on priority setting criteria recommended by Government appointed commissions. A key characteristic of these commissions is the diverse background and strong expertise among their around 15 members including; professors, patient organisation representatives, health professionals, previous MPs. A commission report is typically followed by public hearings and White Paper to the Parliament.
After the report from the Norheim Commission in 2014 their suggested equity criterion was discussed by an expert group and later revised in a subsequent White Paper. Three key criteria: benefits; resources; severity, have now been operationalised into detailed guidelines for equity weighted cost-effectiveness analysis to help reimbursement decisions. Of particular interest is the application of specific equity weights, depending on the magnitude of the absolute QALY-shortfall associated with a disease.
In this presentation, Jan Abel will briefly outline the Norwegian priority setting debate from the first commission in 1987 up until the most recent expert groups in 2024. As a result of the increasing use of HTAs, a crucial topic has been: How best to account for severity and equity in a transparent and consistent way. By use of diagrams, Jan Abel will illustrate different views on the equalisandum (‘Equality of what?’) and contrast the ‘End-of-Life’ criterion with the Norwegian policy of reducing inequalities in prospective health. Based on his involvement as commission member and subsequent advisor on how to operationalize the three criteria into specific guidelines, Jan Abel will reflect on the policy process behind what he considers to be sensible official guidelines for valuing health.
How much is a baby worth?
John Quiggin (University of Queensland)
The global decline in fertility rates has produced considerable concern based on the largely unquestioned assumption that a declining population is economically harmful. In particular, concern has been expressed about the possibility that there will be too few young people to care for an aging population. This paper begins with the observation that the caring work required to raise a child to adulthood is far greater than the end-of-life care typically required by old people. However, most childcare work is voluntary, contributed by parents who have actively chosen to bear and raise children. We lack the economic tools needed to analyse this work, without which it is very difficult to evaluate policies that affect fertility and, particularly, pro-natalist policies aimed at raising fertility.