Facing aging – how employment plays a critical role

After the identification of several aging patterns (see blog articles on Japan, Sweden, Canada, the United States and France), in this post, we first present a synthetic picture of these aging patterns. We then move one step further to compare across countries the resources generated by economic activities to cover the “dependent” population in society. Employment appears as a key factor to generate these resources, largely through earnings and taxation of incomes.

Aging causes a sharp rise in the dependency ratio

In order to measure the economic burden of a non-active population, demographers and economists have developed a simple, synthetic indicator, the dependency ratio. In its most simple – and most used – expression, it is the following ratio:

Population in the numerator consists of the population the least likely to be engaged in economic activity, i.e. children and older adults. Population in the denominator consists of the population most likely to be engaged in economic activity (the so-called working age population). The ratio is interpreted as: for every hundred persons with a capacity to generate economic resources, how many non-active persons are supported. In short, the lower this ratio, the better off a society to cover the needs of its members who “depend” on others’ support.

The gallery of charts below presents the evolution of the dependency ratios in each of the five countries for which we have characterized the aging process in previous blog posts. By clicking on the side arrows, you can review all five charts and compare them. Each chart presents three lines: the total dependency ratio which corresponds to the ratio above, and a disaggregation of the total dependency ratio into its two additive components, the child and old age contributions. In all five countries, as fertility decreases, the contribution of young people aged 0 to 14 years to the dependency ratio diminishes. And so does, in the early years presented on the charts, the dependency ratio. But as aging progresses steadily with rapidly increasing life expectancy, so does the contribution of old age to the dependency ratio, and its global rapid increase. In the five countries, the contribution of old age surpasses that of children before 2015, except in the United States where it is projected to happen shortly before 2025.

By 2015, with the last data reflecting the actual situation, the child dependency ratio has reached its low historical level: every hundred persons of working age “supported” 20 to 30 children aged 0-14 years – 21 in Japan, 24 in Canada, 27 in Sweden and 29 in the United States and France. As for this ratio, it should be considered as rather a sign of hope when it is higher: this is supporting a generation that will move to a supporting role as it grows up to working age. The projected child dependency ratio remains at about the same level through the projection period by design of the underlying assumption about the fertility rate by country for the “medium” estimates reflected in the presented projection

By the same date, 2015, in every country the old-age dependency ratio has started to move up under two sets of pressure: larger generations born after World War II and in the 1950s were reaching old age (65 years or older) and these generations were significantly living longer and longer (the rapid rise of life expectancy). As we said, by 2015, the old-age dependency ratio has risen above the child dependency ratio, except in the United States: every hundred working age persons “supported” 43 old-age persons in Japan, 31 in Sweden, 30 in France, 24 in Canada and 22 in the United States. The rapid growth of the old-age dependency ratio in the projection period is largely a result of the gains in life expectancy. Indeed, with rather stable and low projected growth of the populations aged 0 to 14 years, the old-age population is almost fully accounting for the major upward trend of the total dependency ratio.

From the dependency ratio to the “economic” dependency ratio

The total dependency ratio and its disaggregation into the child and old-age dependency ratios, examined above, provide a demographic perspective. To consider the ratio in economic analysis, it is important to come closer to the reality of how resources are generated to then be used to support the population that needs such support. Resources are generated through the productive capacity of the employed population. Employed persons – the actual working population – earn salaries and income to allow them to support privately those who need support in their private sphere (children, students, ill or aging family members). They also contribute to public programs that offer support through the taxes they pay. We then adapt the dependency ratio in the following way:

In the denominator we put the population that generate the resources, i.e. the total number of employed persons, whatever their age. On the numerator of the ratio we put the non-employed population as the sum of two sub-populations, the population aged less than 15 years (a population that cannot legally be gainfully employed) and the non-employed population aged 15 years or more. In the formula above, this second population is identified in red to remind us that it can be further disaggregated by age group, a feature we will use later in this article. This ratio is interpreted as follows: for every hundred persons gainfully employed, how many non-employed persons are supported.

This economic dependency ratio (EDR) gives a more useful picture to the economist. Moving from the “demographic” dependency ratio (DDR) it considers rightly that not everyone of “working age” actually works: students with no job, non working parent with family responsibilities and unemployed would be the largest groups, and actually need financial support from private or public sources. And, among the population aged 65 years or more, some continue to work, hence contribute to augment the pool of resources that can be made available to those who have to live on the support of others who work.

The chart above provides an easy characterization of moving from the demographic dependency ratio to the economic dependency ratio. The two blocks to the left are constituent to the calculation of the demographic dependency ratio and the two to the right constituent to the calculation of the economic dependency ratio. The two upper blocks form the numerator of the respective formulae, while the two lower blocks are their denominators. Vertically the totals of the two blocks are the same and represent the total population of the given country – here we show as example the situation for Canada in 2015. The total population is 35.95 millions. From the three age-based sub-populations considered in the calculation of the demographic dependency ratio, one is not prone to be in an employment situation – at least in the developed countries we analyze here –, the 0-14 year old population, hence remains in the numerator for both ratios (the blue sub-blocks in the chart). The other two sub-populations have a substantial proportion actually employed: some people 65 years or older are still in employment, hence contribute to the generation of support resources – the steady green sub-block moves from the numerator of the demographic dependency ratio to the denominator of the economic dependency ratio; the opposite happens in the third sub-population, the working age population: most of this group is employed, but a substantial part is not and draws on resources generated by employed (the checkered yellow sub-block), hence moves from the denominator of the demographic dependency ratio to the numerator of the economic dependency ratio. Considering the relative size of the sub-groups of the population that move, this explains that the economic dependency ratio is substantially higher than the demographic dependency ratio.

The chart below compares demographic dependency ratios and economic dependency ratios, and their evolution, for our five countries:

• What strikes first is the systematic – as just explained – and rather large gap between the two for every country. The economic dependency is typically two to three times higher than the demographic dependency ratio.
• The spread of economic dependency ratios across countries is much wider than it is for the demographic dependency ratios. During the period considered, the spread of the demographic dependency ratios amounts to 20 points, while a 70 point difference can be observed among the economic dependency ratios. The wider spread for the economic dependency ratio is in large measure explained by the situation in France where the EDR stands 20 to 60 points above that for other countries. EDRs for the other four countries, between 1995 and 2015 are contained within a 25 point range.
• Variations within countries are much wider for the EDRs than they are for the DDRs. The DDRs evolve uniquely on underlying demographic developments – of the types we identified for each country in previous articles. These developments are earth shattering over the projected long term, but progress at a pace and in a direction predictable. While fertility and aging also directly impact on the EDR, this ratio is mostly driven by national conditions of employment, which reflect both the demographic structure of employment and the employment response to variations in economic conditions. These factors have a huge influence on the EDR and explain much of the differences across countries. We will see later how the level of employment measured by the employment rate critically explains differences in countries’ EDR.

Age-sex contributions to the economic dependency ratio

Economic dependency ratios vary substantially across countries. The demographic composition (by age and sex) reflects both the underlying demographic structure of the population and the national specifics of labour market parameters such as the propensity to combine studies and work, the economic conditions reflected in the level of unemployment, the age of retirement.

The chart below presents the composition of the EDR for our five countries in the most recent year available, 2015.

• First, it is important to situate at what stage in the aging process each country stand in 2015. Let’s refer to the first set of charts in this post. In Japan, the rise of the dependency ratio is already well advanced – since 1995, the proportion of persons aged 65 years and more has been rising sharply while the proportion of young people less than 15 years old has remained steady. In the other four countries, the demographic dependency ratio is just starting to rise, essentially under the influence of aging, as the proportion of the young population has reached its steady (low) level.
• The total height of each bar represent the EDR in 2015 for each country respectively. The five EDRs appear substantially different: in France, it is the highest with 100 employed persons supporting 144 non-employed persons. This is far above the ratio for the United States (115) and those for the other three countries, standing at 100 or just above.
• The contribution of each age-sex group drawing on the support of the employed also varies a lot from one country to another:
  • First, the contribution of persons below minimum working age (less than 15 years) is a direct reflection of demographic factors, i.e. how high or low has been the fertility rate in the last couple of decades before 2015. We should express clearly that it is rather healthy for a country to support a large population of very young people as they will later contribute, through earnings-generating work, in commensurate numbers to the support of the then non-employed population. In 2015, the contribution of these young persons to the total EDR is the lowest with only 26 points in Japan (i.e., 100 employed persons support 26 young persons 0-14 year old). The corresponding number is highest in France (45), then in the United States (41).
  • At the other end of the age spectrum, we see that France again has the highest contribution of the age group 65 years and more, with 100 employed persons supporting 19 men and 26 women (a total of 45 persons) aged 65 and more. Japan, although much more advanced in the process of aging population, presents lower numbers: 16 for men and 25 for women. In the other three countries, the contribution of these groups is significantly lower. As we will see later, these differences across countries are all due to different intensity of employment.
  • Also the level of employment greatly influences the burden imposed on the employed population by the non-employed persons among the working age population. Here again, France cumulates the highest needs for support in these age groups: EDR gets 38 points in the 15-54 age group and 15 points in the 55-64 age group, while the figures are at a relative low in Sweden with 25 and 6 points, respectively.

Let’s turn to the relationship between employment and the economic dependency ratio.

Employment plays a critical role in societies facing aging

From the formula to calculate the economic dependency ratio, we see that the higher the denominator, i.e. the number of persons employed, the lower the resulting ratio. This is actually all the more true as the denominator increases through gains of populations transferring from being counted in the numerator to being counted in the denominator as they get a job – such transfers have been highlighted in a chart above.

The most relevant measure of employment is the employment rate (also named employment-population ratio). It is calculated as the number of persons employed divided by the total population. Particularly relevant here, it can be calculated for the whole population or for any age group. It is in direct relation with the economic dependency ratio – two sides of the same coin, when one just excludes the population less than 15 years old: the higher the employment rate for a given age group, the lower the contribution of this age group’s population to the population in need of support and the higher its contribution to the population generating support resources. For the remaining analysis, we will use OECD data [1]. This restricts the population by excluding the youngest population (less than 15 years) which cannot be found in an employment situation – technically, it just removes it from the numerator when one refers back to the EDR calculation noted above. But it allows us to use more recent data, up to 2017, which we will often use as reference, and more fine-grained data at older ages.

Let’s review the employment rates for the five countries by age for the period 2000-2017 and explain how it impacts on the economic dependency ratio. One can follow the review by clicking on the side arrows to see each age-specific chart in the set of charts below.

• For the youngest age group of the working age population (15-24 year-olds), differences in employment rate are large between countries, highest in Canada (56% in 2017), then in the United States (50%), lowest in France (29%). For young people in this age group, not to be active participants in the labour market may not be a “bad” situation if they are gaining skills for future productive use, with society at large supporting such an investment. But why such differences between comparable countries? Differences in participation in education exist among these countries, but their magnitude is quite small – and explain only marginally the differences in employment rates noted: in France 67% attend education while this is the case for 61% in the United States and 62% in Canada, respectively 64% and 66% in Japan and Sweden [2]. In Canada and the United States, a significant proportion of students are also employed, searching to cover the higher (than in most of Europe) cost of post-secondary studies – 39% work in Canada, 30% in the United States, only 13% in France (most of them in apprenticeship-like programmes). And among those not in education, fewer in France are employed than in any other countries. The NEET ratio, for “neither employed nor in education or training”, is highest in France at 13.6% (with more than half being unemployed, thus actively looking for work) while lower in Canada (9.8%) and the United States (10.8%) with fewer unemployed. In brief, availability of jobs is an issue for this age group in France.
• For the next three ten-year age groups (25-34, 35-44 and 45-54), patterns and levels of employment rates are brought much closer to one another among the five countries, with a span of less than eight percentage points between the highest in Sweden (86.3% in 2017) and the lowest in the United States (78.6%), all above the average for the OECD countries.
• Employment patterns start again to diverge in older ages starting with the 55-59 age group (for a finer analysis of employment patterns at older ages, we present employment data in five-year age groups). For this group, employment rates remain very high (above 80% in 2017) and have been growing steadily in Sweden and Japan. In the United States, the employment rate just fluctuates a little below 70% between 2000 and 2017. In France, from a situation (just around 55%) much below the OECD average, the employment rate started to grow regularly from 2007, reaching 72% ten years later. In Canada, growth of the employment rate was also steady over the whole period, but starting from a low of 60% in 2000 and it remains nine percentage points below that of Japan in 2017. These gaps across countries are consequential as they translate into significant differences in share of population in need of financial support, be it unemployment insurance or welfare supports, rather than being net contributors to that support. With the employment rate of Sweden in 2017 for the 55-59 year-olds, there would be close to 2.3 millions additional jobs in the United States, 500,000 in France and 336,000 in Canada – and holders of these jobs would be contributors to the social net, not beneficiaries. These additional jobs would allow, in these three countries, a reduction of about three points in the overall economic dependency ratio!
• The story line which first appears with the 55-59 age group just gets aggravated in the older age groups, as revealed in the charts above. The aggravation is extreme for France. It is around age 60 that retirement policies and practices (including flexible retirement arrangements and early retirement opportunity offerings) get into play [3]. In recent years, with rising consciousness of the impact of aging, policies to encourage longer working lives make a deliberate attempt to somewhat counterweight, even reverse, the long trend of declining effective retirement age [3]. Indeed, in all countries, employment rates for 60-64, 65-69 and 70-74 year old age groups have been on the rise in the period 2000-2017. If it was also the case in France, it only maintains a gap with all countries – and the OECD average – that appears structural. Among our five countries, Sweden leads with highest employment rates up to age 64 with 68.4% in the 60-64 age group in 2017 – France only records 29.2%. After that age, Japan takes a solid lead with employment rates of 44.3% for the 65-69 years age group and 27.1% for the 70-74 years age group – with the lowest marks, France records 6.6% and 2.8%, respectively.

In 2017, in Sweden, every 100 workers generate resources to support 47 non-employed persons aged 15 years or more. Hundred workers support 97 non-employed persons, in France, 62 in Canada, 70 in Japan and 60 in the United States (76 on average in OECD countries). Just to put some perspective on this with tangible employment numbers, in order for France to attain Sweden’s level of support that its employed population provides to cover its non-employed population above the age of 15, France should count 8.8 millions more jobs – exactly a third more than it has presently! As rapid aging is a certainty in all these countries for the coming decades, actually starting just now, if not already well engaged, the challenge appears of an aggravated magnitude in countries where there are fewer resources generated through employment. Furthermore in France, we have seen that it happens in a unique demographic context in which the working age population (aged 15-64 years) does not grow in number at all and is not projected to grow in coming decades.

On an individual basis, one gets a high five and congratulations for a well deserved retirement and is looked upon with envy, the earlier they retire, of course essentially if this is on their own willing. But the situation of societies facing acute aging progress calls for looking at the phenomenon with a different public policy lens. It must consider that the level of employment in the country, as we have argued here, is critical to generate the financial resources needed to allow a comfortable situation for the ever increasing population reaching old age – whatever this age be, largely socially determined – and living ever longer. Here we focus on resources to face aging, as this is the population whose growth puts additional demands on the supporting society. Obviously, we do not forget that the resources generated by employment also has to cover the needs, in particular education ones, for the younger cohorts… and, by the way, France again has the highest contribution of the youngest group to its EDR!

References

[1] OECD.Stat (2018), LFS by age and sex, Data extracted on 13 November 2018.

[2] OECD (2018), Education at a Glance 2018: OECD Indicators, OECD Publishing, Paris. http://dx.doi.org/10.1787/eag-2018-en

[3] OECD (2017), Pensions at a Glance 2017: OECD and G20 Indicators, OECD Publishing, Paris. http://dx.doi.org/10.1787/pension_glance-2017-en

 

---

Non solum data – Data sine monito oculo nihil sunt.

---

Through this blog, we invite constructive comments and constructive article contributions – review our blog policy and have your say!