Mortality Risk Valuation in Policy Assessment (OECD)

• Calculating health costs two ways

The economic benefits of prevented fatalities can be expressed in terms of a “Value of Statistical Life” (VSL), which represents the value a given population places on avoiding the death of an unidentified individual, such as reducing the number of deaths from air pollution. Importantly, even if a change in mortality risk is not explicitly valued in monetary terms, it is implicitly valued through policy decisions with economic implications. However, such implicit values can vary considerably from case to case. While people may object to explicit valuations on ethical grounds, the use of implicit values is pervasive, even if these values are not always observable. Explicit values derived from carefully conducted valuation techniques improve the information available to decision makers by providing a consistent and transparent metric that can be applied across policy areas.

One important tool to promote consistency in policy assessments is cost-benefit analysis (CBA). CBA compares the total expected costs of a given policy against its total expected benefits, to determine net benefits for society. As policy decisions that impact human life are a major concern, the economic valuation of human health effects, including VSL, is therefore a cornerstone of CBA.

VSL estimates can be derived from surveys where individuals are asked about their willingness to pay for small improvements in mortality risk (stated preference methodologies) or from observing trade-offs that individuals make between income or wealth and the risk of dying (revealed preference methodologies), such as taking a higher-risk job in exchange for a higher salary. Given that VSL often constitutes a sizeable share of the total economic benefits of many public policies, it is important to use the best available estimates.

This report presents an update to the OECD’s 2012 VSL report, which included the largest VSL meta analysis (based on primary stated preference VSL studies published between 1970-2008) and quickly became an important international reference. In contrast, this report includes a meta-analysis of more than 4 000 individual estimates from 277 studies across 49 countries between 1970-2023, from which about 2 400 VSL estimates between 2009-2023 were used for the base VSL estimates. The current report is once again the largest VSL meta-analysis ever attempted and the first to develop VSL estimates based on both revealed and stated preference methodologies. While most of the primary VSL studies in the database are from Europe and North America, it also includes numerous studies from Latin America, Africa, Australia and Asia.

The report provides base VSL estimates (mean and median values, and confidence bands) for six country groups: OECD Member Countries, European Union, United States, Low-and middle-income countries, High-income countries and a Global estimate. The estimated mean base VSLs range from about USD 1 million in low- and middle-income countries to between USD 7.1 million and USD 8.5 million for OECD and high-income countries. Due to using newer data based on more recent methodological improvements, more comprehensive geographical coverage and the inclusion of both revealed and stated preference methods, the results presented in this report are not directly comparable to those in the OECD’s 2012 report.

The base VSL estimates should be considered as a starting point for policy analyses that involve mortality risk and should be adjusted to the specific country and policy context whenever possible. Perhaps the most important determinant of the VSL is income, measured as household income or as GDP per capita. These findings stem from both the meta-analysis of valuation studies from all country groups, as well as from an analysis of the relevant scientific literature on the drivers of VSL.

This report provides the following recommendations for the use and adjustment of base VSL estimates:

• Use base VSL estimates from the smallest relevant country grouping to derive country level estimates. Base VSL estimates can be used to compute country-level estimates for those that do not already have sufficient VSL evidence or guidelines. It is recommended to scale the base VSL estimates from the relevant country group to an individual country by using the ratio of GDP per capita between the target country and the country group of the base VSL estimate. Since countries may belong to more than one group, it is recommended that policymakers use the VSL estimates from the smallest grouping to which a country belongs as a starting point and further adjust VSL estimates to account for specific national and study contexts whenever possible (for example impacted population subsets or job sectors).
• Use country-specific VSL estimates for evaluation of national policies. For policies with multi country impacts, it is recommended to also use national VSL estimates for each respective country affected. However, in some situations it may also be justifiable to apply a common VSL estimate across multiple jurisdictions to ensure that similar mortality effects receive similar consideration in all impacted jurisdictions (for example within the European Union).
• Consider and value morbidity effects in addition to mortality. In most situations where mortality risk is impacted, there are also impacts on morbidity (pain and suffering). Morbidity impacts should be considered and valued separately from mortality costs, both for morbidity prior to premature death as well as for non-fatal health outcomes.
• It is not recommended to adjust VSL estimates for the type of mortality risk or the age of the population. Except for cancer, this report suggests that the type of mortality risk does not significantly impact VSL estimates. Cancer, when included in primary valuation studies, led to significantly higher VSL estimates compared to studies where it was not included. However, cancer is a complex disease that can take many forms and be associated with differing symptoms, durations and outcomes, the details of which are not possible to capture in the meta-data used for this report. Therefore, even if cancer could in principle be a valid basis for adjusting VSL estimates, the data underlying this report does not offer sufficient evidence to recommend a specific approach for adjusting the VSL for cancer risk. Similarly, it is widely recognised that the age of the studied population could have a significant impact on the VSL estimate. However, adjusting for age is controversial as it implies that public policy and cost-benefit analysis outcomes could differ solely based on differences in the age of the populations studied and it is not common practice to make such adjustments.
• Conduct uncertainty and sensitivity analyses using alternative income elasticities of VSL when transferring benefits between countries and/or regions, as well as when adjusting VSL estimates over time. As shown in the ranges presented in this report, considerable uncertainty surrounds the VSL estimates. It is therefore recommended to conduct sensitivity analyses using alternative (but plausible) assumptions to test the robustness of any CBA that uses VSL.

The results presented in this report are expected to be widely applicable to policy analysis in which there is a need to assess the economic benefits of policies that impact mortality risks.