Re: Population ageing: the timebomb that isn’t?
This Rapid Response is in reply to several comments that were made to the BMJ article I co-authored with John MacInnes, criticising the Old Age Dependency Ratio (OADR) and suggesting an alternative Real Elderly Dependency Ratio (REDR) 1. The OADR takes the number of people who have reached the State Pension Age and divides it by the number of working age adults to estimate the ratio of older people to those who pay for them. It thus defines all people above the statutory pension age as ‘dependent’, taking little account of how rising life expectancy (RLE) may render them ‘younger’ and healthier than their peers in earlier cohorts. Our REDR takes the age at which remaining life expectancy (RLE) is less than or equal to 15 years rather than a fixed age as an indicator of ‘dependency’, and since many of ‘working age’ are not in paid employment, uses the numbers actually at work as its denominator.
Basten suggests we gave insufficient credit to work by Sanderson and Scherbov. We disagree. We clearly stated in the article that our work was a continuation of theirs and one quarter of our citations were to their work! A short ‘analysis’ article is not an adequate forum to discuss e.g. their Prospective Old Age Dependency Ratio (POADR) or Adult Disability Dependency Ratio (ADDR), and discuss their relative merits. We were also aware of Bongaarts, cited his work, but did not discuss his “Pension-Worker Ratio” because the issue of pension scheme sustainability lay well beyond our scope.
We agree with Basten that the meaning of dependency should be better defined and hope our article encourages a fuller debate about it. Economic ‘dependence’, for example, is a poor term to describe someone who may currently receive a state pension but has spent decades paying tax and national insurance. Many jobs counted in the denominator of our measure ‘depend’ directly or indirectly on the consumption (including that of health services) of those counted in the numerator. As the proportion of older people in all countries rises, so too will this number. Health, personal and social care services all imply the existence of ‘dependence’ rooted in permanent or temporary health status or disability, but so too does less visible similar work done by informal or family carers. Unlike mortality data, there are no good, consistent, sufficiently long time-series data on physical or mental health status, limiting long-term illness or disability by age, and much of what does get collected excludes the population resident in nursing or care homes. Unlike death, health status has a subjective component. There are also gender and international differences in disability that are only partly associated with the level of life expectancy (the so-called health-survival paradox 2). For these reasons there is a fierce debate over the actual and prospective compression or expansion of morbidity. Hence while we agree with Crawford’s observation about the greater demands ageing baby boomers will place on health services, we would qualify it in two ways. The first is that this demand will continue to be driven more by innovation in medical technologies than demographic shifts. The range of possible treatments (and therefore of the numbers affected) will continue to grow, and most beneficiaries will be older people. The second is that (contrary to the misleading impression created by population ‘pyramids’) although the baby boomers were born together, they will not die (and, by implication, age) together. Some are already dead, while the future centenarians still have a half century or more to live. We doubt the value in the longer term of a catch-all category such as ‘dependence’. We retained the word in our indicator only because there was no obvious preferable substitute that would have been easily recognizable to our readers.
Turning to the denominator of our measure, Jacobs suggests taking account of part time work. However give the range of hours of part (and full) time workers, this would have to be done by counting annual working hours. Over the last century and a quarter, these hours have halved, partly because of a shorter working week, but also owing to fewer working weeks (more and longer holidays) and fewer working years (later entry to, earlier exit from, employment). For this reason, and contrary to conventional wisdom, we see no demographic rationale for changing the State Pension Age of 65, introduced, let us remember, in 1928, when UK GDP per capita was less than one quarter of today’s level. What afforded the trend fall in hours per year can easily enable a gradual fall in the proportion of year across the lifecourse devoted to paid work. Raising the State Pension Age may simply increase inequality: only the poorest, without the assets to go early, will work on. Many will be divorced women. Measures to increase the employment rate (around 72%) of those of working age are better. The dramatic fall in working hours has been made possible by trend labour productivity growth of roughly 2% a year. Thus the logic of taking hours into account would be to use the value of economic production (Gross Domestic Product) as the denominator. This has much going for it, since it would foreground the issue of transfers of resources over the life course and between generations, and of the mechanisms that can achieve this (the family and the state). However that requires at least a book, not a short article.
1. Spijker J, MacInnes J. Population ageing: the timebomb that isn’t? BMJ: British Medical Journal, 2013;347:f6598.
2. Van Oyen H, Nusselder W, Jagger C, Kolip P, Cambois E, Robine JM. Gender differences in healthy life years within the EU: an exploration of the “health–survival” paradox. International Journal of Public Health, 2013;58(1):143-155.
3. Sanderson WC, Scherbov S. A new perspective on population aging. Demographic Res 2007;16:27-58.
Competing interests: No competing interests