print this

2. WHAT ARE THE DEMOGRAPHIC PROSPECTS AND THE POSSIBLE CORRECTING POLICES?

Should the projected demographic changes and emerging demographic disequilibria give rise to worries and for what reasons? And what are the potential correcting policies? This section sketches both the projected demographic developments – according to the most recent medium variant projection of the United Nations’ 2005 projection – and a classification of world regions – according to their projected demographic development. This is followed by a brief clarification of some conceptual issues regarding the implication of these projected shifts for the financing of public programs, in particular, pensions and health care. The section ends with a short presentation of the main corrective policies.

2.1. The medium variant of the UN 2005 projection

The most recent demographic projections by the United Nations confirm that a demographic transition is taking place worldwide (UN 2005). The expected increase in life expectancy in most countries (except those severely hit by HIV/AIDS), combined with falling fertility rates in countries with a total fertility rate above the replacement rate and continuing low rates in countries with a total fertility rate below the replacement rate will substantially shift the demographic structure among countries and regions by 2050.

The main common and distinct demographic changes are:

• All countries in the world are aging, and the change is sometimes most pronounced in middle-income countries (such as China). This applies to all definitions of aging, such as the level and average population age, old-age dependency ratio (typically defined as the ratio of persons ages 65+ to persons ages 15-64), and share of elderly (ages 65+) in the total population.
  
• The total population in much of the developing world will continue to expand, as the strongest youth cohort ever enters the labour market.
  
• The old-age dependency ratio will increase in low-to-middle-income regions of Latin America, East Asia and the Pacific, Middle East, North Africa, and Turkey, South and East Asia, and Sub-Saharan Africa from as low as 0.06 (Sub-Saharan Africa in 2005) to as high as 0.29 (Latin America and the Caribbean in 2050), the total dependency ratio in these regions will either remain unchanged or decline significantly (such as for Sub-Saharan Africa) – see table 1.
  
• In the developed world in the “North” i.e. Europe plus Russia, the high-income countries of East Asia and the Pacific (Australia, Hong Kong, Japan, New Zealand, Republic of Korea, and Singapore), North America (United States and Canada), but also in China, the demographic projections under the medium variant foresee low or even negative demographic growth until 2050 and a substantial deterioration in both the old-age and the total dependency ratio within the next 45 years (table 3). For high-income countries in East Asia and the Pacific, the old age dependency ratio is projected to reach 0.63 in 2050, and the total dependency ratio is projected to be 0.88.
  
• The effects of these demographic shifts are even more pronounced when considering the impact on the labour force. For the medium variant, the labour force is projected to decrease until 2025 in Europe plus Russia and in high-income East Asia and the Pacific by 38 million and 6 million people, respectively. The decrease is projected to treble by 2050, including China, with a decrease of 89 million. And this projected change includes (moderate) migration assumptions. In a zero-migration scenario, the fall in the labour force would be more pronounced for Europe and high-income East Asia and the Pacific (but not China, which is assumed to remain a net exporter of migrants); North America also would see its labour force decline by some 9 million people by 2050 (see table 3). Under the zero-migration variant, the labour force in the North is projected to lose some 244 million persons, while the labour force in the South is projected to add some 1.55 billion, predominantly in South and Central Asia and in Sub-Saharan Africa.

Table 3. Dependency ratios and labour force, by region, 2005-50

Sources: United Nations (2005), author’s calculations

2.2. Some Implications of Demographic Disequilibria

These projected demographic shifts—in particular, the dramatic aging of the population, together with low or even negative growth of the population and labour force, in many countries in the North – are giving rise to many speculations. They include issues of national and international security, shifts in economic power, and consequences for the financing of national social programs, pensions and health care.

Such public programs cater to the elderly but are financed by the contributions and non-consumed income of the working population (whether they are pay-as-you-go financed or pre-funded). They already consume a major share of the general budget and national output in the richer (and less rich) countries in the North. Some estimated 15 percent of GDP, on average, goes to public pensions and health care in the Organisation for Economic Co-operation and Development (OECD) countries, and an additional 5 or more percentage points of GDP are needed for privately financed pension income and health outlays. The projected demographic shifts (in North and South) are rightly expected to put further pressure on these already stressed social programs and public budgets.

This subsection attempts to clarify some critical conceptual issues linked with population shifts and pension and health care programs in order to inform the following discussion of corrective demographic policy actions and general policy requirements. The issues addressed concern:

1. the irrelevance of the type of funding when faced with demographic shifts;
2. the dependence of benefits on the explicit and implicit rates of return;
3. the dependence of the implicit (and explicit) rate of return on the demographic structure and dynamics of a country; and
4. the importance of the causes of aging (reduced fertility or longer life expectancy) for policy reaction.

For the financing of pensions and health care, the form of financing (pay-as-you-go or pre-funded) matters much less than often assumed (see, for example, Holzmann, Hinz, and Bank team 2005). In the end, all of these outlays need to be financed out of current GNP and each generation of retirees and consumers of health care needs the next generation to pay contributions or to buy the accumulated assets. The form of financing matters with regard to the quality of collateral: Do the contributions and insurance premiums create property rights? Do they contribute to enhanced national savings?

As a result, policies that affect the number of individuals as payers of contributions or buyers of assets matter for the financial sustainability of these schemes.
While the notional (pay-as-you-go) or actual pre-funding of pension benefits is easily understood, the pre-funding of health care benefits is often not. Health care benefits – whether public or private – typically do not work as a spot insurance market in which the premium is determined by the current-period risk profile of the insured, which is typically related to age. This would make health insurance unaffordable for most elderly. As a result, the typical health insurance premium is above the current-period actuarial expenditure level at a younger age and includes a component of savings to finance future expected health care expenditures above current contribution revenues. This makes both health care and pension benefits dependent on the explicit or implicit rate of return of these social programs. And these rates of return are closely linked to the demographic structure.

It is increasingly understood that the explicit or financial rates of return are also dependent on the demographic structure and dynamics of a country. The savers of assets for retirement income or health care benefits need the buyers of these very assets once they need to sell them. International diversification of these assets in view of asymmetric aging across countries helps at the margin, but it does not constitute a solution (Holzmann 2002). And the interconnectedness of capital markets makes national interest rates also dependent on global demographic trends (McKibbin 2005). The implicit rate of return of unfunded systems is closely linked to the growth of the base of contributions (which depends on productivity growth/the growth of real wages per capita, and the labour force/the growth in the number of contributors).

Hence projected demographic shifts will influence the internal (and external) rate of return and thus the capacity to deliver pension and health care benefits through three channels. First is the change in labour force. For some regions, the impact of the projected fall in the labour force during the next 45 years is quite substantial. In Europe plus Russia, the impact would amount to a reduction in the (implicit) rate of return in the range of 0.7 to 0.9 percent annually. For North America, the assumed migration will add almost 0.6 percent annually to the rate of return during the next 45 years. Second is the impact of aging on productivity growth. There are a number of reasons why an aging labour force may exhibit lower productivity growth per worker, taking into consideration knowledge creation and entrepreneurial spirit. Econometric evidence for 115 countries suggests that the share of the elderly population has a statistically significant impact on growth of real GDP per capita (IMF 2004). This research suggests a reduction in the annual real growth rate of GDP per capita of 0.5 percent, on average, by 2050 (Martins et al. 2005 and IMF, 2004). Third is the impact of population aging on the ratio of assets to liabilities of unfunded pensions. While aging increases the liability position, it also increases the asset position by increasing the average number of years between contribution payments and benefit disbursement.

Population aging can occur as the result of a reduction in the total fertility rate or an increase in life expectancy. In the real world, both effects occur at the same time and have been of roughly similar magnitude in many countries in recent years. But the effects on social programs and hence the policy conclusions are not identical. With a total fertility rate at the replacement level (which in highly developed economies is in the range of 2.05 to 2.1 children per woman, but can be as high as 2.64 in countries such as Namibia if mortality rates are high throughout the fertility period), population aging occurs through a fall in the age-specific mortality rate, which raises life expectancy at all ages. In developing countries, recent gains in life expectancy have occurred at young ages; in highly developed economies, the gains have occurred at older ages (60 and beyond), as the mortality rates at younger ages are already low. A gain in life expectancy of some 10 years at age 60 – from 80 to 90 – and deterioration in the (demographic) old-age dependency ratio from, say, 2:1 to 3:1 can be easily addressed by raising the retirement age by 6.6 years. This would reestablish financial balance in the old-age income system and leave the internal rate of returned largely unchanged.

With a given life expectancy, population aging is driven by a fall in the total fertility rate. This also occurs when total fertility rates are falling and rates are above the replacement rate. If left constant, this situation leads to a constant population structure and a permanently growing population. At constant total fertility rates below the replacement level the population is permanently shrinking. While the deterioration in the population structure and the resulting increase in the old-age dependency ratio can be corrected by an increase in the retirement age (as under rising life expectancy), the fall in both population and size of the labour force needs an additional correction to deal with the otherwise lower benefit level due to lower implicit rates of return.

The above estimates suggest that some countries and regions have an aging-induced deterioration in the internal rate of return for pension income of 1.3 percent a year (0.8 percent due to the smaller labour force and 0.5 percent due to lower productivity growth). Such calculations do not include the price effects for retirees of higher personal health care costs, which are likely to reduce income by a similar magnitude. Nor do they include the effects of lower rates of return on pre-saving in health care insurance. Yet such a reduction in the annual rates of return over a whole life cycle has important implications for benefit levels. In a life-cycle setting, a 1 percentage point (100 basis point) lower rate of return translates broadly into a 20 percent lower pension benefit. The actual magnitude of a deteriorated rate of return may require compensation in a reduced replacement rate which exceeds 30 percent or more.

2.3. The set of corrective policy options

This subsection sketches the set of corrective demographic policy options, which includes a higher total fertility rate (that is, more children per family), higher labour force participation of the existing population, and increased immigration. There is, of course, a much larger set of potential non-demographic corrective policy options, such as enhanced productivity growth per employee or international diversification of investments. Although such policies may prove important, they are beyond the scope of this paper and are conjectured not to be full substitutes for the demographic policy options addressed here.

Higher total fertility rate

Fertility rates at the beginning of the new millennium reached a new low in many countries and regions: Russia, 1.2; Europe and Japan, around 1.4; China, 1.7; and North America – almost 2.0. The medium demographic variant assumes some recovery in the total fertility rate for Russia, Europe, Japan, and China toward 1.85 and a slight fall for North America by the end of the projection period. In all cases, the replacement level would not be reached by 2050.

One potential policy option is to attempt to reach a replacement-level total fertility rate as soon as possible and keep it there. This would stabilize the labour force in the long term and, eventually, create long-term population growth, but solely through an increase in life expectancy (or immigration). The short- and medium-term effects, however, would be limited. The past decline in the total fertility rate has already reduced the number of women of child-bearing age, so the increase in crude birth rates would be limited. And any increases in the number of births would need some time before they have an effect on the labour market – that is, some 15 years or more.

Furthermore, this option would contribute to an immediate deterioration in the total dependency ratio – that is, the sum of the youth dependency ratio (the ratio of ages 0–14 to ages 15– 64) and the old-age dependency ratio (the ratio of ages 65+ to ages 15–64).
Section 3 investigates the quantitative effects of this policy option, and section 4 sketches potential policy measures to make it happen.

Increased labour force participation

A second option consists of increasing labour force participation – that is increase the share of population participating in the labour market, including the elderly. While such an increase in labour force participation can expand the total labour force and hence may compensate for a fall in population in potentially active age brackets, it requires more: If the active population continues to decrease, even a partial compensation will require a continuous increase in the labour force participation rate. And increased labour force participation will not translate 1:1 into increased social benefits as such an approach would postpone, but not solve, the underlying financing needs of such programs.

Increased immigration

A simple mechanism in quantitative terms to compensate for low or negative population growth is to import population from other countries. Such an approach promises a number of direct advantages for the migrant-receiving countries. First, most of the migration typically takes place among person’s ages 25 to 35. Hence immigration immediately enhances the labour force, while only gradually contributing to a higher dependency ratio. Second, given an assumed elastic supply of migration-willing individuals in developing countries, this policy may, in principle, compensate extremely well for any population gap in quantitative terms and, with an appropriate filtering mechanism for a gap in skills and other characteristics. Last but not least, with appropriate policies and incentives, part of the migrant population may be induced to return to the sending country. The experience with such gap-filling immigration approaches, however, has met its limits, to which we return in section 4.