Dread and latency impacts on a VSL for cancer risk reductions

J Risk Uncertain (2016) 52:137–161 DOI 10.1007/s11166-016-9235-x Dread and latency impacts on a VSL for cancer risk reductions Rebecca L. McDonald 1 & Susan M. Chilton 2 & Michael W. Jones-Lee 2 & Hugh R. T. Metcalf 2 Published online: 26 April 2016 # The Author(s) 2016. This article is published with open access at Springerlink.com Abstract We propose a structural relationship between the value of preventing a statistical cancer fatality and the value of statistical life (VSL) for risks of an instantaneous road accident fatality. This relationship incorporates a context effect reflecting both the illness or ‘morbidity’ associated with cancer fatality and the ‘dread’ or horror associated with the prospect of eventual death from cancer, as well as a latency effect that captures the discounting likely to arise because the onset of the symptoms of cancer typically occurs after some delay. We use a Risk-Risk trade-off study to validate this model by directly estimating the influence of context and latency effects upon the relative size of the VSL for cancer and for road accidents, confirming that both effects are significant and estimating their size using regression analysis. We show that morbidity accounts for the majority of the context premium. We use the elicited coefficients to reconstruct VSL estimates for a range of cancers characterised by their latency and morbidity periods. Keywords VSL . Cancer . Risk-Risk trade-off JEL Classification J17 This article and the work it describes were funded by the Health and Safety Executive (HSE). The article’s contents, including any opinions and/or conclusions expressed, are those of the authors alone and do not necessarily reflect HSE policy. The paper was also supported by the Economic and Social Research Council, UK [grant number ES/K002201/1] and the Leverhulme Trust [grant number RP2012-V-022]. * Hugh R. T. Metcalf 1 University of Warwick, WBS Scarman Road, Coventry CV4 7AL, UK 2 Newcastle University Business School, Newcastle University, 5 Barrack Rd., Newcastle NE1 4SE, UK 138 J Risk Uncertain (2016) 52:137–161 1 Introduction The Value of Statistical Life 1 (VSL) reflects the monetary value to the affected population of a small reduction in the risk of premature death (Jones-Lee 1989). In the UK, it is conventionally defined as the aggregate willingness to pay by a large, representative sample of individuals for small reductions in the risk of death which, taken across the group concerned, will reduce the expected number of fatalities during a forthcoming period by one and will hence save one Bstatistical life^. As such, the VSL will clearly be determined by the marginal rate of substitution (MRS) of wealth for risk of death for individuals within the affected group, and for a group enjoying equal marginal risk reductions, it will be equal to the arithmetic mean MRS for the group concerned—see, for example, Jones-Lee (1989) and Viscusi (1998). By contrast, in the USA, the term is typically applied directly to the MRS. Both definitions clearly share the same conceptual foundations. In the study reported in this paper, we analyse the data at the individual level which is akin to the US approach, although for UK policy extrapolation we aggregate the data to conform with the UK definition and application of the VSL. In the UK, the (aggregated) VSL was initially implemented in the transport sector (Jones-Lee et al. 1985) and elicited in this context. However, increasingly there has been a policy-driven trend to extend its use to other sectors, both in the UK and elsewhere, as has been noted by the Organisation for Economic Co-operation and Development (OECD 2012). Whilst intuitively we might expect contextual factors including the mode of death to affect the perceived value of risk reductions, governments generally apply the same VSL across contexts. The exception is when the change in risk applies to cancer fatalities, as exemplified by the UK Health and Safety Executive (HSE)’s 2001 current recommendation for a cancer VSL.2 ‘Where the benefit is the prevention of death, the current convention used by HSE, when conducting a CBA is to adopt a benchmark value of about £1,000,000 (2001 prices) for the value of preventing a fatality (VPF). This is the VPF adopted by the Department of Transport, Local Government and the Regions for the appraisal of road safety measures. It may well be the case that individuals’ willingness to pay for risk reduction—measured in aggregate by the VPF—will vary, depending on the particular hazardous situation. Thus, the particular hazard context will need to be borne in mind when a VPF figure is adopted. Currently, HSE takes the view that it is only in the case where death is caused by cancer that people are prepared to pay a premium for the benefit of preventing a fatality and has accordingly adopted a VPF twice that of the roads benchmark figure. Research is planned to assess the validity of this approach.’ (HSE 2001) Applying a cancer premium to the VSL is principally intended to reflect the fact that death from cancer is typically preceded by a protracted period of ill-health and dread caused by the prospect of imminent fatality. In addition, evidence from the first stated 1 Also termed the Value of Preventing a Statistical Fatality (VPF) in the UK. The Directorate-Generale of the EU takes a similar position, recommending a cancer premium of 1.5 (see European Commission 2001). 2 J Risk Uncertain (2016) 52:137–161 139 preference-based VSL study conducted in the UK indicated that as far as members of the public were concerned, willingness to pay (WTP) for a given reduction in the number of deaths from cancer was almost double the corresponding amount for reducing the number of road accident deaths by the same figure—see Jones-Lee et al. (1985). This position is supported by economic theoretic and psychological evidence which suggests that willingness to pay for risk reductions may naturally be expected to vary across causes, for example because of differing baseline exposure risks (Pratt and Zeckhauser 1996; Covey 2001) or personal risk perception (Slovic 1987; McDaniels et al. 1992). When well-informed and carefully considered survey responses generate a VSL that incorporates a premium for certain types of fatality, then a strong case can be made that the VSL figure used in policy making should reflect this premium. 2 Context, latency and cancer risks Nevertheless, two important reservations can be made with respect to the upward adjustment of the VSL for cancer (hereafter VSLCANCER). The first relates to the fact that death from cancer will not take place immediately even if exposure to this risk is in the current period. This is due to the latency period in the development of most cancers. Latency is defined as the delay between an individual’s exposure to the conditions that are the basic cause of cancer and the onset of symptoms, with death occurring after a period of il (...truncated)