Cost of Aging

Population aging and slower labor force growth affect the economy in many ways, reducing the growth rate of GDP, increasing costs to working age people of supporting the elderly, and stressing public budgets by raising program costs for the elderly. At the same time, population aging may raise capital intensity, boosting productivity and wages, and reducing interest rates. Secular stagnation is a risk if slowing economic growth discourages firms from investing, even as interest rates fall and loanable funds are abundant. Firms, pension funds, and others may turn to higher risk investments in an effort to maintain rates of returns.

Slowing growth of population and labor force

Because of low fertility, first population growth rates and then labor force growth rates have declined in the advanced economies, despite immigration. Basic growth models (Solow, 1956) tell us that slower growth of population and labor force leads to slower growth of GDP, and many empirical studies have confirmed this, roughly on a one-to-one basis. In the US, the growth rate of the population age 20–64 has slowed by almost 1% per year from 1975–2015 to 2015–2055 (United Nations, 2015), and we expect that will lead to a corresponding decline in the growth rate of GDP and aggregate consumption. A number of advanced countries already have declining population of working age, and in Europe as a whole it will decline at an average pace of .6% per year from 2015 to 2055, dropping more than 20%. This slowdown in aggregate growth will change the investment climate and the shares of the advanced economies in global output. Individual well-being depends on per capita, not aggregate, growth, however, and the same standard growth models tell us that slower labor force growth will lead to rising output and wages per worker. But will higher output per worker translate into higher per capita income? That will depend on how the rise in output per worker trades off against the decline in workers per dependent as the population ages. To get beyond the standard model and these basic ideas, we need to look more closely at how economic activity varies by age. This we can do by drawing on National Transfer Accounts which breakdown National Accounts by age, and additionally cover private transfers within and between households (see NTAccounts.org, Lee and Mason et al (2011), and United Nations (2013)).

Rising old age dependency and falling support ratios

As we all know, children consume more than they produce, and the same is true for the elderly, on average. In between, prime age adults produce a surplus over their consumption, as shown in Chart 1. As the population ages, the proportion of the population at the right of the chart increases, where consumption is high relative to younger adult ages and labor income is low. In some countries, such as the US, Sweden, and Japan, relative consumption by the elderly is much higher than in Chart 1, while in other countries such as Spain and Austria, the relative increase is much smaller. The increase in the share of elderly may be partially offset by a reduction in the proportion of children in the population. If fertility rates begin to recover from low levels, however, the proportion of children and elderly in the population may increase in the future.

Open in a separate window

Chart 1

The economic life cycle in high income countries: labor income and consumption by age

Source: Data are taken from National Transfer Accounts, as processed and extended in Mason, Lee, et al (2017 forthcoming). For a detailed discussion of methods see www.ntaccounts.org and United Nations Population Division (2013).

Notes: The chart shows simple averages of country-specific estimates for 24 high-income countries as classified by the World Bank in 2015. Labor income and consumption for each country are first standardized by dividing by average labor income at ages 30–49 for that country, and then averaged. Labor income is all pre-tax returns to labor composed of earnings including employee benefits; self-employment labor income; and an estimate of the value of labor supplied by unpaid family workers. Consumption by age is based on consumer expenditures of each household, imputed to individuals using methods described in United Nations (2013), plus in-kind public transfers received by people of each age.

The budgetary pressure of the age distribution can be summarized by the “support ratio”, the ratio of effective workers to effective consumers. Effective workers is the sum of the population age distribution weighted by country-specific age profile of labor income on which the average profile in Chart 1 is based. Effective consumers is defined similarly. The age profiles shown in Chart 1 are held fixed at their baseline level, but as the projected population age distribution changes the support ratio also changes. Holding other factors constant, the level of age-adjusted per capita consumption varies in proportion to the support ratio. Chart 2 shows how the support ratio is projected to change in the US, the other rich nations as a group, and in China. Between 2015 and 2050 it will drop by .26% per year in the US, by .40% per year in other rich nations, and by .82% per year in China. Others things equal, this means that adjusted consumption will drop by 25% in China, 9% in the US and 13% in other high income countries. This is a simple measure of the dependency side of the impact of population aging.

Open in a separate window

Chart 2

The changing burden of dependency: Support ratio, China, United States and High income countries other than the United States, 1980–2050. Values are expressed as a percentage of the 2015 support ratio.

Source: Mason, Lee, et al (2017 forthcoming).

Notes: The “Other high income estimate” is a simple average of country-specific estimates for 24 high-income countries as classified by the World Bank in 2015, excluding the US. See text for discussion of methods. Comparisons of the levels of the three lines are not meaningful; comparisons over time are, as discussed in the text.

As seen in Chart 1, the elderly consume far more than they earn in labor income. How is their old age consumption paid for? Aside from labor income, a part is paid by income earned on their assets including the return to real assets, such as a farm or own business, the value of the housing received by home owners, as well as net interest, dividends received, and undistributed profits of corporation. Another part comes from the government in the form of cash such as pensions and in-kind public transfers such as health care, long term care, or the public education received by children. These public transfers are paid for by taxes, particularly those paid by the prime age adult population. Yet another part may come from net support (support received less support given) provided to the elderly by younger family members. In practice, net support is positive on average for the elderly mainly in East Asia, and even there this is no longer true in Japan and South Korea. In Europe and the Americas, and much of the remainder of Asia, the elderly on average give more to their younger family members than they receive.

Chart 3 shows how these patterns of funding the consumption of the elderly vary from region to region. In general, the higher the proportion of elderly consumption that is paid through their own labor or asset holdings, the less their cost falls on prime age adults through higher taxes. Europe stands out for heavy reliance on public sector transfers and little reliance on either asset income or continuing work, a recipe for heavy costs of population aging. The reverse is true for the US, where elders retire later and rely more on assets. Latin America is between the two, while Asia is more like the US. Regarding family transfers, in Asia these help to support the elderly, but in Europe and the US the elderly give net support to younger family members. Although public discourse often emphasizes the transfers received by the elderly, it is really only net transfers to the elderly, whether public or private, that make population aging socially costly.

Open in a separate window

Figure 3

Funding sources for consumption as a percentage of consumption, for adults 65 and older.

Source: National Transfer Accounts: www.ntaccounts.org

Notes: Values are simple averages of country-specific estimates for members of each region. North America consist of a single country: United States. Asia includes: Australia, Cambodia, China, india, Japan, Lao PDR, Philippines, South Korea, Taiwan, and Thailand; Europe: Austria, Finland, France, Germany, Hungary, Italy, Slovenia, Spain, Sweden and United Kingdom; and Latin America: Brazil, Chile, Colombia, Costa Rica, Ecuador, El Salvador, Jamaica, Mexico, and Uruguay. Labor income is all returns to labor composed of earnings including employee benefits; self-employment labor income; and an estimate of the value of labor supplied by unpaid family workers. Public and private transfers are net transfers defined as transfers received less transfers provided. Asset-based flows are defined as asset income less interest expense less saving (or plus dis-saving).

Public sector transfers to the elderly for pensions, health care and long term care are a particular problem as populations age, because these payments, even after subtracting the portion funded by tax payments from the elderly, already absorb a large portion of public budgets. Projections indicate that typically these programs will be unsustainable without adjustments to net public transfers, by adjusting taxes or benefits in the future. Fiscal support ratios are a simple way to see the problem. They are constructed like the support ratios shown in Chart 2, except the numerator is effective tax payers rather an effective workers, and the denominator is effective beneficiaries rather than effective consumers. In the US, the fiscal support ratio will drop by 11% between 2010 and 2050 due to population aging. That means that to balance the tax revenues and expenditures in the public budget (federal, state and local combined) in 2050, either tax revenues will have to be raised by 11% or expenditures will have to be cut by 11%, or some combination of the two, just to offset the increased costs due to population aging. For European countries, the corresponding number is between 14% and 28%, and for Japan 26%. These figures refer to the total government budget, not just the social welfare component. Some governments are attempting the politically difficult task of restructuring their public transfer programs to address these issues, by raising the retirement age, switching to “notional defined contribution” pensions which mimic defined contribution systems but are actually pay-as-you-go, and so on. While these reforms may avoid the problem of long run fiscal instability, they cannot avoid the rising costs of supporting the elderly. Reforms only reallocate these costs and benefits within and across the generations in different ways.

Slower labor force growth and population aging will raise capital available per worker (capital intensity)

As mentioned earlier, standard growth models imply that if the aggregate saving rate remains the same while the rate of growth of the labor force declines, then capital per worker will rise, raising productivity and wages, and reducing interest rates. In the US, the annual growth rate for effective workers (as defined above) declined by about 1 percent for the 2015–2055 period as compared to the 1975–2015 period. This change alone would imply a substantial increase in capital per worker if savings rates were to remain the same as the population ages. It is tricky to analyze the effect of population aging on saving, because data attribute all household saving to the head, whereas other household members at different ages may in fact be partly responsible. After adjusting for this problem (see Mason et al, 2015), a further difficulty is that future saving will depend on future public policies regarding pensions, health care, long term care and public education. In Mason et al (2015) we have simulated private and public saving rates for Japan under different policy scenarios and found that in general, the combined aggregate saving rate rises somewhat in the medium term and then fluctuates within a range that remains a bit higher than earlier years. These results imply a small further increase in capital per worker due to higher saving. Different kinds of calculations based on the NTA data suggest similar or larger increases in capital per worker. Theoretical analyses based on more complicated models likewise suggest that population aging will lead to more capital per worker. To the extent that this occurs, it will tend to offset the adverse effect on per capita income and consumption of the falling support ratios and rising old age dependency.

One reason that capital per worker may not increase is that population aging may drive up government debt, crowding out funds for investment in capital. Another possibility is that if capital per worker does begin to rise, pushing down interest rates, then adults will choose to save less, reducing the increase in capital. A third possibility is that those with money to invest will seek higher returns in foreign capital markets, particularly in developing regions and emerging nations where populations are younger and rates of return may be higher. In this case, domestic workers will not benefit through rising wages and higher productivity, although returns on the foreign investments will still raise national income, although perhaps by assuming greater risk.

The threat of secular stagnation

Firms may choose to substantially cut investment in the domestic economy, even as interest rates fall, given expectations that output and consumption will grow more slowly in the future, anticipating slower growth of population, labor force, consumption, and perhaps total factor productivity as well. In this case, even if central banks drive interest rates below zero the economy could remain stagnant with high unemployment, a condition called “secular stagnation”, which could persist for many years. Some economists interpret the Japanese experience of recent decades and the failed recovery of Europe from the Great Recession in these terms (Teulings and Baldwin, 2014).

Summing Up

The key point is that population aging and slower population growth will almost certainly bring a slowdown in the growth of GDP and national income, but the effect on per capita income and consumption, the experience at the individual level, may well be quite different. Population aging will mean increased old age dependency, to the extent that the elderly are not self-supporting out of asset income or their own increased labor. But population aging may well bring more capital per worker, and rising productivity and wages, particularly if rising government debt does not crowd out investment in capital (Lee and Mason, 2011 for ; Lee, 2016). An analysis in Lee, Mason et al (2014) concludes that moderately low fertility, around 1.7 births per woman, is favorable for standards of living. Only very low fertility, at levels found in East Asia, Germany, and Eastern and Southern Europe, will produce aging that reduces standards of living.

Acknowledgments

Contributor Information

Ronald Lee, Professor of the Graduate School, University of California at Berkeley, ude.yelekreb.gomed@eelr, 510 540 8986.

Andrew Mason, Professor of Economics, University of Hawaii at Manoa and Senior Fellow, East-West Center, ude.iiawah@nosama.

References