Valuing gains in life expectancy: Clarifying some ambiguities

It is well-understood that a given gain in life expectancy can, in principle, be generated by any one of an infinite number of different types of perturbation in an individual's survival fimction. Since it seems unlikely that the typical individual will be indifferent between these various types of perturbation, the idea that there exists a unique willingness to pay-based Value of a Statistical Life Year (VSLY), even for individuals within a given age-group, appears to be ill-founded. This paper examines the issue from a theoretical perspective. Within the context of a simple multi-period model it transpires that if gains in life expectancy are computed on an undiscounted basis then it will indeed be necessary to adjust the magnitude of the VSLY to accommodate the nature of the perturbation in the survival function, as well as the age of those affected. If, by contrast, gains in life expectancy are computed on an appropriately discounted basis then a unique VSLY will be applicable in all cases.