Technical Appendix: Quantifying the Impact of Economic Outcomes

Construction of the IEO will involve some methodological and practical challenges, but this has been true also with GDP, and statistical agencies have routinely solved the issues that have arisen in that context. The inclusion criteria and weights for the IEO are based on estimates of the magnitude of the causal contribution of each candidate item to health and/or subjective wellbeing. Ideally, these would be derived from a comprehensive evidence base, but this will require development work. Initially, reliance will need to be placed on expert opinion where the evidence base is inadequate for this purpose. Calculation of the magnitude of the causal effect is based on the population attributable fraction, a standard epidemiological measure. This requires a value for the causal relative risk of each variable, adjusted for the effects of the others. The formula for this calculation is given, and explained, for a single exposure and for multiple exposures, for new cases (counts) or alternatively for the timing of deaths—the interval by which a death is brought forward, using survival analysis. The latter formula for years of life lost (YLL) can readily be extended to cover a reduction in the health-related quality, rather than the length, of life. This draws on the well-established literature on Disability-Adjusted Life Years (DALYs), which contains values for the severity of the loss of functional health due to disability or disease in terms equivalent to loss of life duration—e.g. diabetic foot is counted as 0.20, meaning that the value of five years living with that condition is considered equivalent to four years without it. The same calculation can be performed for subjective wellbeing using Wellbeing-Adjusted Life Years, and in principle these can be combined with health loss (DALYs) resulting in what I call D/WALYs. Each IEO component, such as illiteracy or insecure livelihood, is likely to play a causal role in various different types of impairments of health and/or subjective wellbeing. These would be combined as the sum of the D/WALYs lost that are attributable to that IEO component. Aggregation of the different IEO components to produce the overall index would start with the calculation of the complements of the scores of each item—for example, if the value is 0.05, its complement is 0.95. The IEO would then be calculated as their geometric mean (as has been used in the Human Development Index). The consequence is that the ideal IEO is 1 (probably best expressed as 100%), with real-life values being less than 1, and with the largest values indicating the most successful economies.