Project: ‘Active Ageing Index (AAI)’


UNECE Grant No: ECE/GC/2012/003






Developing a Conceptual Framework


Part 2

Age Inflation and Lifetime Indexing: Redefining Age, Ageing, and Autonomy/Dependency


European Centre Vienna

April 2012



Paper prepared for a discussion during the 1st Meeting of the

Expert Group on Conceptualizing and Measuring Active Ageing


May 10 and 11, 2012

Hotel Bloom Brussels




Age Inflation and Lifetime Indexing: Redefining Age, Ageing,

and Autonomy/Dependency by Sanderson/Scherbov, Shoven et al.


Bernd Marin, 30 April 2012 (preliminary draft)



Can ageing societies simultaneously get younger?

In recent years, concepts of age inflation and lifetime indexing have emerged and may contribute to fundamentally rethinking age and ageing in the 21st Century. Within the still thin scientific literature on the subject, the work of Sanderson and Scherbov (2005, 2007, 2008, 2010, 2011, also Lutz, Sanderson, and Scherbov 2001, 2008) has been conceptually most innovative, intellectually stimulating, methodologically rigorous, and practically relevant.


It attempts at nothing less than redefining – and remeasuring - age, old age, individual and population ageing and also dependency an autonomy in rapidly ageing societies. For the first time, intuitively plausible insights and experiences such as “40 is the new 30” get scientifically sharply defined and precisely measured. Through pioneering new key concepts such as “standardized age” (2005) or later “prospective age” (since 2007), “prospective median age”, “proportional life cycle rescaling”, and life expectancy adjusted measures such as the “prospective old-age dependency ratio”(POADR),  the “adult disability dependency ratio”(ADDR) (2010), “prospective healthy age” (PHA) and “prospective cognitive age” (PCA),  or “hedonic age” (2012), they help to understand the core paradox that ageing societies may nevertheless grow younger at the same time.


The World Population after the war, for instance, has aged considerably ever since in terms of median age or in terms of the share of, say, 65plus years of chronological age, the conventional measures of population ageing. But thinking from the end of life in terms of remaining life years and taking into account the increases in life expectancy, the World is simultaneously becoming younger till about the year 2015 and then ageing at a much slower pace. Longevity gains may correspond to significantly increased or reduced proportions of “elderly” or “very old people”, depending on whether an “old-age” threshold is defined by a chronological age of 65+ or by a prospective age or remaining life expectancy of 15 years or less – the latter measure shows the World still rejuvenating (2005, 811 and 2008, Figure 5, panel A, 11).


Is “young” and “old” time-space neutral? Today many Europeans have a higher life expectancy at the age of 50 than Goethe had at birth. And how “young” died Mozart?

The basic ideas of “rethinking age and ageing” (2008) focus around queries such as how old is “old”, how young is “young” or “middle-aged”, and what age is a given age? Can age be determined exclusively by years spent since birth, irrespective of highly variable remaining life expectancy, survival probability, and age-specific mortality risk, and irrespective of highly diverse health status, cognitive capacity and (dis)ability? Had Mozart already outlived his life expectancy at birth when he died at age 36? Is dying with 36 dying “young” or dying “old”? Why would it considered to be “very young” today and still “mid-life” and not “old” in 1791, even if the life expectancy at birth of Mozart was below the years of chronological age reached when he died? Why were Kant or Goethe at their 50th birthday considered very old man (“Weimar grüßt den ehrwürdigen Greis”), whereas at the “same age” today many people have a higher further life expectancy at the age of 50 than Goethe had at birth?


Thus, quite obviously 50 is not 50, 65 is not 65, 40 is not 40 years of age, regardless of time and place, age structure and longevity developments of the overall population. Age cannot meaningfully be conceptualized and measured regardless of health, life expectancy, mortality risks and survival probabilities. By implication, “young” and “old” are highly different in ancient or post-war times and today, as they are highly different today in Moldowa and France, the Ukraine and Switzerland. They are also different for scientists and construction workers, and for men and women; but the gender longevity gap is highly different in Cyprus vs. Russia, Iceland vs. Belarus, United Kingdom vs. Lithuania, Israel vs. the Ukraine – and in Greece vs. France. Women of the “same” chronological age are significantly – but highly varyingly - “younger” then men of the “same age”, but the 5 -13 years gap at birth shrinks over the life course to 3-5 years at age 65 and to 1-2 years at chronological age 80 – as do cross-country and all other life expectancy differentials.


Unfortunately, traditional measures of chronological age, i.e. years lived since birth, do not take these crucial variations into account. Sanderson and Scherbov raise just these questions – and developed a sophisticated methodology to provide meaningful answers to them: when in history did a 43-year-old European man have the same residual life expectancy as a 60-year-old today? They show that a 62-year-old Australian man in (the “base year” or “reference year” or “standard year”) 2000 had exactly the same (19.63 more years) further life expectancy as 54-year-old men in (the “index year”) 1950 (2008, 7); or a French woman 40-years in 2005 the very same 44.7 years remaining life expectancy than a 30-years old woman in 1952 (2008, 5).



Mortality-risk-based age measures (Shoven et al.)

John B. Shoven (2007, Shoven and Goda 2008, Goda, Shoven and Slavov 2007, etc.) has developed age measures based on remaining life expectancy and, above all, on mortality-risk, the percentage chance of dying within the next year. Such mortality-risk-rated age measures allow, for instance, to state that for US men between 1970 and 2000 “65 is the new 59”, that the age of 65 in 1965 corresponds to the age of 69 for women and 72 for men in the year 2000, meaning that men are seven and women only four years “younger” around retirement age than 35 years ago; that men aged 68 in 1970 have the same age than men of 75 years in 2012 (defined by a 4% mortality risk), that men aged 58 in 1970 are 67 today (defined by a 2% mortality risk), so that men around pensionable age are eight to nine years “younger” than 40 years ago. The shorthand formula: “60 is the new 51”, “67 is the new 58”, “75 is the new 68” – since 1970.


In the European context, this can be illustrated both historically and with reference to the most important political implications. My generations’ grandmothers (born around 1890) had the same 46 years of life expectancy at birth than our mothers had as further life expectancy at the age of 30 in the 1950s/1960s and as a 38 year old women has today in countries like Switzerland or Austria. 14 years of life expectancy gained at birth over half a century since around 1960 translate into 10 years difference at midlife (the age of 40 today corresponds to 30 in the 1950s) and around eight years at retirement age (over a period about a decade shorter) - “73 is the new 65” of the 1970s.


The political implications are most tangible and sensitive. When the very same people celebrating the “trente glorieuses” or “golden times” up to the mid 1970s – a period one used to work till the age 62 or up to almost 67 – now simultaneously protest against the prospect of possibly having to work till 70 in 2060, they simply ignore that 62-67 in 1970 corresponds to 70-75 already today and to around 74,5 and 79 in the year 2060. Thus, European populations have already worked much longer in so called golden times than they fear they would have to under circumstances of much longer lives and lifetimes of leisure. Political resistance against adjustments in pensionable age, for instance, may be quite different if policy targets would always be formulated in terms of prospective age as well, taking age inflation automatically into account through “lifetime indexing” (Marin 2012).


In order to provide for a scientifically based determination of pensionable age adjustments to age inflation through lifetime indexing, changes in further life expectancy at legal retirement age are important but will not be sufficient. Equally important is the probability to survive up to that age and a standardization of age in terms of mortality risks requires the identification of mortality rates as a suitable life table indicator. This will allow to precisely stipulate how much “younger” people of the same chronological age have become over a specific time period or how much later a fixed chronological age threshold would have to be set in order to correspond to a fixed prospective age, remaining lifetime, mortality-risk or survival chance. 


Today older people tend to have fewer disabilities than people of the same age in earlier decades, and now there is some evidence that cognitive decline is being postponed as well, despite the newly discovered phenomena of “Mental Retirement” (Rohwedder and Willis 2010). The latter authors show that early retirement has a negative impact of cognitive decline about a decade later and triggers off more rapid mental ageing with early labour market exit (not easily compensated by other, non-occupational tasks) and a kind of  “on-the-job-pre-retirement” effect years before actual out-of-work transition. Thus, early retirees become “older” and less able cognitively than their continuously active counterparts of the same chronological age who did not exit the labour force.


Even the media have recognized some of these fundamental demographic changes. We often read that “40 is the new 30” but this is more than just a pop culture phrase. It is a challenge to demographers to rethink how they measure a population’s age and the pace of ageing.” (2008, 3). Sanderson & Scherbov developed several measures adjusted for considering differences in life expectancy.


Individual vs. population ageing

Conceptually, “population ageing differs from the aging of an individual. People who survive grow older with each year they live. Populations, on the other hand, can grow younger. Because a wide variety of matters such as the cost of medical care, retirement, bequests, consumption and the accumulation of human and tangible capital depend not only on age but also on time left to live, our understanding of population ageing must also reflect both of these factors.” (2005, 811) Two years later, Sanderson and Scherbov write: “The aging of populations and of people have different dynamics. Surviving people must grow one year older each year. Populations, on the other hand, do not necessarily grow one year older each year. Populations can grow more than one year older, less than one year older or even grow younger with the passage of time. When age is measured as a two dimensional variable our descriptions of populations aging grow more complex. With two ages to consider, populations can simultaneously grow younger according to one measure and older according to the other.” (2007, 29).


But even for individuals, chronological age as a static measure is a poor summary indicator of their age as a stage in their overall life cycle, ignoring crucial age-specific characteristics of persons such as their remaining lifetime, their health conditions, cognitive capacity, general (dis)ability, etc. The very plasticity of age and ageing requires such two- or even multi-dimensional indicators in order not to mis-measure them and to significantly and momentously mislead policy-makers and the public at large – today, and even more so in the future.


People have simultaneously two ages: chronological and  prospective age

Sanderson and Scherbov “think about people as simultaneously having two ages. One is chronological age – the number of birthdays a person already has had. The second is prospective age – based on the number of birthdays a person can expect to have. That future number is their remaining life expectancy. With two different age concepts, a person can be both 40 and 30 at the same time.” (2008, 4, emphasis added). Treating two 65-year-old persons in 1970 and in 2010 as if they would have the “same age” (they actually have the same chronological age, but highly different prospective ages) is grossly misleading.


The concept of prospective age is meant to compensate this deficiency by supplementing conventional measures of age such as chronological or retrospective age (2007, 28) and ageing. It builds on the economist Victor Fuchs (1984) who “suggested that people have two different ages. Borrowing from the common distinction in economics between values measured in current prices (nominal values) and those adjusted for inflation (real values), Fuchs suggested people have “nominal” and “real” ages. In 2005, we independently reinvented Fuchs’ proposed “real age” and provided examples of how it could be consistently measured over time and across countries.” (2008, 5, emphasis added).


Adjusting median age for life expectancy is like adjusting prices for inflation – or factoring-in “age inflation”: “If you were told that a pair of shoes would cost $500, 50 years from now, would you be able to tell whether those shoes were cheap or expensive? Certainly not. Adjusting for inflation, those shoes might cost $30 in today’s prices or perhaps $300. If you were told that a person was 65 years old 50 years from now, would you be able to tell whether that person was old or not? Certainly not. People at age 65, 50 years from now, could have remaining life expectancy of five or 35 years.” (2008, 6).


As comparing monetary values from one period to another needs an appropriate price index, allowing for adjusting for inflation, adjusting median age for longevity by prospective age just requires appropriate (period and cohort) life tables. Computing prospective age asks for matching remaining life expectancies in two life tables. It takes life expectancy differences between different historical times in the same country or between different countries at the same point in time into account. In contrast to the “backward-looking” chronological age, prospective age is “forward-looking” – but can only supplement and not substitute for the conventional age measure. “These measures are not just different metrics for measuring the same thing. They measure different aspects of aging, ones in which biological and behavioural factors play a larger role.” (2010, 288)


They have schematically outlined the difference between chronological or retrospective and prospective age as follows:

Sanderson & Scherbov Diagram Showing How Prospective Age is Determined

Source: Sanderson and Scherbov 2007, 33

See Appendix p 2

In an article a year later, the authors have illustrated the distinction between the backward-looking and the forward-looking conceptualization of age. They utilized the Human Mortality Database of the University of California, Berkeley and the Max-Planck-Institute for Demographic Research, with reference to French women during the last decades as a case in point.

Chronological vs. Prospective Age: “40 is the new 30” – French Women 1952 and 2005

See Appendix p 3


Limits to chronological age – and to prospective age as well ?

There are “limits to chronological age”, making it potentially misleading as a sole indicator: “when using indicators that assume fixed chronological ages, it is implicitly assumed that there will be no progress in important factors such as remaining life expectancies and in disability rates. But many age-specific characteristics have not remained fixed and are not expected to remain constant in the future” (2010, 128) – including further increases in residual life expectancy and the (somewhat decreasing – or not?) speed of life expectancy increases. Consequently, assumptions about unchanged life prospects, disability rates or behaviour will almost certainly turn out to be wrong. Without taking “age inflation” into account, many phenomena cannot be properly understood, in particular the relevance of anything future-oriented – from investments into learning, new skills-formation, the acquisition of qualifications and long-term goods, saving based on long-run rates of return, the costs for long-term medical and social care, or the debates about shifts in pension eligibility age or qualifying age for Medicare and other age-related programmes and benefits.


On the other hand, the new concept of prospective age may have problems of its own. For many persons, it may not be fully plausible intuitively that all people of the same prospective age, say 45 years, have the same remaining life expectancy as a 45-year-old person in the standard year, say 1970.  It may also not be easily understood that the prospective median age declines when the life expectancy at the median age is growing faster than the median age.  But some measures based on prospective age such as, for instance, the proportion of “old”//”young” people with remaining life expectancies RLE-10/15//40/50 may be much more easily understandable, because they use a constant prospective age without defining any standard year!


Combining chronological and prospective age

When defining old age, Sanderson and Scherbov build on their double-track, simultaneously backward- and forward-looking perspective of combining chronological and prospective age. The latter concept is population-based and provides for advances in health and longevity. “With the enormous variability in life expectancy at older ages across countries and over time, a fixed age threshold for classifying people as old has not reflected reality.” (2008, 7) One can fix either chronological or prospective ages, and both are relevant for different aspects of ageing. “An alternative to having a fixed age at which people are categorized as old is to define old age as beginning at some threshold level of remaining life expectancy. This theory was first offered by Norman Ryder in 1975; he recommended that old age be considered to begin when remaining life expectancy fell below 10 years. Fuchs followed with a more complete analysis in 1984. In 1993 Jacob Siegel suggested the possibility of using a remaining life expectancy of either 10 or 15 years to demarcate the boundary of old age…..To our knowledge, Wolfgang Lutz and we did the first computation of the proportions of the elderly in the world and in the populations of major regions, basing the onset of old age in remaining years of life expectancy.


Choosing to define old age as beginning at some remaining life expectancy threshold is equivalent to defining it as occurring at a fixed prospective age. Thus, there are two ways of defining old age: an old-age threshold based on chronological age and one based on prospective age. Here, we define old age as beginning when people are at ages when the remaining life expectancy is 15 or fewer years.” (ibid.)


The authors concede that this “may not be the best way to determine which individuals are old”, but they would not know how to do it in plausible alternative ways. Also, and surprisingly open, they concede that “choosing a remaining life expectancy of 15 years or less instead of, say, 10 years, males our results less sensitive to…inaccuracies” of incomplete data in highest ages, in particular. (2008, 16, Footnote 9) From a political economy of pensions policy point of view it may sound quite amazing that the authors took their 15-year-threshold choice instead of a 10-year one purely “for a pragmatic reason” of minimizing inaccuracies, not even discussing the economic and political implications. While everybody could agree to the basic concept of defining old-age rather from the remaining years of life then from the retrospective age already achieved, the choice of the threshold is obviously crucial – both for the financial and fiscal sustainability of the overall system, as for the benefit level generosity, social adequacy – and political acceptability of the pension scheme proposed.


If nowadays people got used to decades of retirement periods – French, for instance, can expect 27 years of pension duration on average – a 10 or 15-year remaining life expectancy definition of old-age might imply that old-age pensions will be restricted to old-age according to this somewhat arbitrary threshold as well. While fixing some prospective eligibility ages or pension duration and extension rules might be a meaningful and fair idea, it requires a broad dialogue and consensus building. The authors themselves, for instance, mention the example of eligibility age for the U.S. Medicare Program was set at 65 in 1965 and remained constant since (2008, 8, Shoven and Goda 2008). But if that age were adjusted for life expectancy change since, the eligibility age today would be 73 for women and 75 for men instead of fixed 65.


Intelligence and fairness considerations in design choices: comparative policy and technical advantages of combined chronological and prospective ages over retrospective age thresholds only

Sanderson and Scherbov are aware of fairness considerations in design choices as well. “A fixed chronological age for receiving a normal pension is unfair to younger generations. As life expectancies increase, generations pay into the pension system for a fixed number of years, but receive benefits over ever-lengthening periods of retirement. But a fixed prospective age for receiving a normal pension is unfair to older generations. As life expectancies increase, they would have to pay into the pension system for more and more years, only to receive benefits over a fixed average period. Averaging chronological and prospective ages can produce an inter-generationally fair normal pension age.” (2008, 15, emphasis added) In this spirit, they welcome the ongoing rise of retirement age in the U.S. from 65 to 67 by 2027: “As it happens, this increase is generally quite consistent with ages suggested by averaging chronological and prospective ages. However, current legislation calls for increases in the normal pension age to end with the cohort born in 1960. But with changes in life expectancy, there should be shifts in the age of eligibility for full pension receipt even among cohorts born after 1960, if the pension system is to remain fair to younger and older generations.” (ibid.)


Prospective age has the comparative policy advantage over retrospective age thresholds of flexibly adjusting for life expectancy changes in both directions, as longevity gains are uncertain (in their pace) and could even reverse, if, for instance,  “obesity or diabetes epidemics” would cause reduced remaining life expectancy. Prospective age also has a comparative technical advantage over retrospective age thresholds in that Sanderson and Scherbov showed that changes in prospective age based on remaining life expectancy are insensitive to which type of life tables (period or cohort life tables) will be used. This invariance as to the use of specific life tables gives the measure a much broader and robust applicability, whereas chronological age indicators vary significantly depending on which kind of life tables are used or available at all.


New measures of population ageing

Using the new concept of prospective age allows to develop new measures of population ageing. Instead of adopting the share or proportion of elderly by counting the people above a fixed chronological cutoff point (e.g. proportion 60+ or 65+) as a share of the total population, one can use the ratio of those in age groups with a remaining life expectancy of 15 years or less to the overall population (proportion RLE -15). Similarly, when calculating old-age dependency ratios, the traditional OADR, based on retrospective or chronological age, simply divides the number of people 65 or older by the number of people aged 20 to 64 (one can, of course, also use ages 15 or 60, or any other span). The prospective old-age dependency ratio POADR, in contrast, is the ratio of the number of people above the – variable - age threshold (e.g. the age at which remaining life expectancy falls below 15 years) to the number of people age 20 to the old-age threshold, whatever it is.


It should come as no surprise that empirical evidence finds POADRs evolving significantly (though widely varying) lower than traditional OADRs till the year 2045, of the countries chosen by the authors (2008, 12, Table 3) most so in Japan and least so in Russia. As of 2005, not Italy, Japan and Germany are the oldest countries by OADR, but the Ukraine, Bulgaria and Belarus, followed by other Central and Eastern European countries, making up for the total top-ten-oldest countries as measured by POADR (ibid., Box 3, 13). Similarly, the percentage of the population 65+ is somewhat lower than the percentage of the population at ages with remaining life expectancy 15 years or less already today, as the age at which remaining life expectancy is globally already 66.3 years and a few years higher in the more developed regions. But the difference between the share 65+ and REL -15 till the year 2045 will be widening very strongly, with almost 50% higher a level of 65+ (15.2%) over REL -15 (10.9%) worldwide. In particular, the age at which remaining life expectancy is 15 years or less will rise to around 72 or 73 years in Europe, North America, Latin America and the Caribbean.


Changes upward in median age are conventionally utilized to determine the pace of ageing of given countries or regions. If the distinction between retrospective and prospective age is applied to median age, prospective median age is to be found by determining the median age first and then finding the prospective age corresponding to that age. “The prospective median age of a country in a particular year is simply the prospective age of median-aged persons in the country in that year.” (2008, 16) But in contrast to calculating RLE 15- and POADRs, where figures can meaningfully be compared across countries without a common standard, computing prospective median ages for different countries requires a standard life table as a common reference, without which a comparison would not be possible.


There is one other innovative application of this new thinking about age, ageing, autonomy and dependency, the creation of an adult disability dependency ratio (ADDR) (Sanderson and Scherbov 2010), a new dynamic measure of age-specific disability rates for 17 European countries. It is defined as the number of adult persons (at least 20 years of age) with disabilities, divided by the number of persons (at least 20 years of age) without disabilities. Please note that while there is a lower age-limit defining adulthood, there is no upper age limit restricting adulthood to, say, working age, or the age of 65, RES -15, RES -10, or any other age-threshold. What counts and what is counted in this measure are adult persons with and without disabilities and the dependency ratio based on these capacities, incapacities and respective burden and burden sharing within a population.


Technically, it takes an “estimation of the relation between disability-free life expectancy and unconditional life expectancy” (2010, Supporting Online Material, 1), using data calculated by the European Health Expectancy Monitoring Unit (EHEMU) from European Union Income, Social Inclusion and Living Conditions survey (EU-SILC) data on 17 EU countries and UN life table data to forecast life expectancies. Then “the prevalence of disabilities in each 5-year group” is calculated, “working sequentially from the oldest group 85+ to the youngest 30-34” (ibid.2). It takes into account only people who are (according to their self-assessment) “strongly limited” in their daily life activities, or so called activity limitations. They must also be persistently limited, i.e. for over half a year or permanently due to health problems, without distinguishing between physical and mental health.


The authors “make forecasts of disability rates for high-income OECD countries up to 2048, and the trends in Old Age Dependency Ratios and Adult Disability Ratios are dramatically different” (ibid.4), so the strong summarizing conclusions of scholars usually quite prudent and differentiated. Looking at the findings, these “dramatically different” dependency ratios cannot but be confirmed (Sanderson and Scherbov 2010 (Science 329, 1287, Supporting Online Material, p 7-9

See also the Appendix, p 12

The chart confirms what could have been expected from the reasoning above, and what the authors stated as a general finding: “Not only does the ADDR increase less rapidly than the OADR, it also increases less rapidly than the POADR, so that adjusting for the likely future path of disability rates does not simply replicate the results of adjusting aging measures for changes in longevity.” (2010, 288).


It also demonstrates quite impressively that getting older and getting healthier at the same time and at the same age are offsetting each other – age-related and age-specific dependencies have not been fixed in the past and will most probably also not remain unchanged in the future. This will become even more relevant as “the pace of increase in life expectancies in more developed countries has not slowed over the last half-century. Although somewhat controversial, there is an emerging consensus among demographers that there is little reason to expect a generalized slowdown in the near future.” (2008, 4, see also Christensen, Doblhammer, and Vaupel 2009, Bongaarts 2002 and 2006, Bongaarts and Feeney 2002, Lee 2006, Ediev 2011, Lee and Carter 1992, Oepen and Vaupel 2002, Olshansky, Carnes, and Brody 2002, Olshansky et al. 2009, Vaupel 2010, 2011).


In the end, Sanderson and Scherbov also express some hopes about the practical and political implications of their new measures: “Such new measures of aging can help educate the public about likely consequences of improvements in health and longevity. Slow and predictable changes in pension age, for example, justified by an increased number of years of healthy life at older ages may be more politically acceptable than large, abrupt changes justified on the basis of budgetary stringency….A change in…legislation, for example, that would increase the normal pension age by one-half year for each year of additional life expectancy at age 65 would go a long way of ensuring the sustainability of Social Security payouts, even without further reforms.” (2010, 288).


What retirement age is required to keep dependency ratios stable?

We cannot realistically expect increased employment fully – or only mainly – compensating for steeply rising old-age dependency ratios, whether conventional or new ones. Hence, another, comparative look at them seems advisable as to when and how much the statutory – traditional chronological - retirement age needs to be raised in order to maintain a given balance between age groups working and age groups above the – shifting or not – so-called working age. The following table provides these calculations for the EU-25.

Old-Age Dependency Ratios for 3 Different Statutory Retirement Age Thresholds (60, 65, 70) in EU-25, 1960 – 2050


See Appendix, p 51

As pay-as-you-go social security and pension systems are based on an poise between working and non-working, respectively retired populations, their relatively stable equilibrium is required to keep the system in balance – and sustainable over time. Here, the differences and similarities between European Union country averages and single country cases in point (here: Austria as an example) are quite telling: EU-25 countries ageing somewhat later, but a bit more rapidly, both during the last and into the next half of a century, but with basically the same equilibration requirements. In order to keep the existing balance between active and dependent population groups, the actual age threshold between them must be shifted for five years (from 60 to 65) between the first decade into the new Millennium and 2025, and for another five years (from 65 to 70) between 2025 and the year 2050 – on purely demographic terms.

To the extent that all EU countries (with the general exception of France and several exceptions for female eligibility age) have a statutory retirement age of 65 already, the necessity for the next 15 years is, above all, to shift the effective retirement age upwards to the legal one. Only after the year 2025, the legal retirement age itself will have to be raised upwards as well. The necessary shift upwards would be another five years for purely demographic reasons and could be softened and reduced to the extent that the labour slack potential could be mobilized (to about 68 or 69 years of chronological age).

Will the Statutory Retirement Age Have to be Raised Every Quarter of a Century for About Five Years? What Eligibility Age is Required to Keep the Old-Age Dependency Ratio Stable? EU-Europe 1966 – 1978 – 2003 – 2028 – 2059


See Appendix, p 52

When we ask for how much and when the statutory retirement age will have to be raised, we see the sharp difference between past and future: while there has been absolutely no demographic necessity for any rise in retirement age during the last 40 years (1966, 1978 and 2003 had the very same OADRs, with minor fluctuations up and down in between), the next quarters of a century would each require a rise of the actual retirement age (till 2028) and of the legal retirement age (till 2059) by about five years in order to keep the old-age dependency ratio and therefore the PAYG system stable.


Yet, both the core concepts of “age”, “autonomy” and “dependency” themselves could be re-defined according to the new concepts of either “prospective age” (and the “prospective old-age dependency ratio “/POADR) or the “adult disability dependency ratio” (ADDR) instead of the traditional old-age dependency ratio (OADR). Then, an increase in required mandatory retirement age may still be as many years as required under the OADR, but it would carry such a different label and different meaning that it could possibly be more easily acceptable to broad population groups. Unfortunately, nothing less than a mental revolution is required to accept these requirements of an ageing and simultaneously rejuvenating society: redefining age and old age by prospective age, mortality-risk and survival probabilities, making them and the corresponding health status, cognitive capacity, (dis)ability and autonomy/dependency more relevant than chronological age as such. 


The Time-Space Context of Active Ageing Indicators (Annex)



Age Inflation, Lifetime Indexing, Historical Timing of Ageing

Dynamic Age Definitions

(„Young“, „Old“, „Midlife“), Adjusting „Nominal Age“ for Unwarranted Fixed Thresholds to „Real Age“ – Population Ageing Through Space and Time

* Age and Proportion of People with RLE-15 vs. Share 65+ (1960 – 2012)
* Age and Proportion RLE-10 vs. Share 65+ (1900 – 2050)
* Age and Proportion of Persons with Mortality Risk > 1%, 2%,3%, 4%, 5%, 10% p.a. vs. Share of People 50+ to 80+

* Age and Proportion of Persons with Survival Rates > 50%, 66%, 75%, 80%, 90%
Pace of Ageing
* Prospective Median Age vs. Median Age 1950 – 2010 – 2050


Historical Timing of Population Ageing

* Time and Age at Which People Had / Will Have Remaining 40 (20, 15, 10) Years
   to Live
* Year When Certain Median Age Thresholds Were/Will Be Passed („Year When
   Half the Population Is Above/Below 20,30,40,50“)

* Year When OADR >=YADR (e.g. It 1980, Turk 2050)
* Year of Ageing Peak
* Years when Ageing of the Aged (Share of the 80+ in the population 65+, ratio >
   15%, 25%, 33%, 40%) or Longevity Thresholds Were/Will Be Passed

* Work, Education and Retirement over the Life Cycle 1960 – 2012
* Extension of Effective Retirement Duration 1960 - 2012
* Age-Inflation-Proof Measures of Working Age and Retirement Duration
   1960 – 2012


Pensions Literacy and Mental Fitness for Active Ageing

* Measuring „Retirement Illusion“ or „Pension Illiteracy“:
Misperceived Retirement Years and Underestimated Lifetime Pension Wealth Per Capita (real over assumed retirement years – and their pension wealth value)

* Incongruent Perceptions of „Young“, „Old“ and „Middle-Aged“ (how people define and how realistically they assess different ages and stages in the life cycle)