For younger cohorts, the costs of a pay-as-you-go pension scheme rise and benefits fall as populations age
As income rises, families tend to have fewer children and people tend to live longer. Together, these forces have raised the old age dependency ratio in many countries.[1] This demographic transition is already well under way in OECD countries, and dependency ratios are expected to climb in much of Latin America, Central Asia, Eastern Europe, and China over the next two decades. By the year 2050, only Africa will still be "young."
As the resources of fewer workers are stretched to support a larger old population, pay-as-you-go schemes will inevitably yield rapidly diminishing payoffs to future generations, unless productivity rises fast enough to offset the effects of demography. Projected demographic profiles for Argentina, China, Hungary, Japan, and Kenya illustrate the increasing costs and decreasing returns to younger cohorts under pay-as-you-go pension plans, with populations aging between now and 2075.
Japan's population is younger than those in most OECD countries today, but it will age rapidly. Although its population is still growing slightly, growth will turn negative after 2005. Hungary, a typical Eastern European country, has a population that is already old and shrinking, though the decline will be more gradual than in Japan. The number of people in Argentina, as in many other Latin American countries, is expected to increase for the next century, though at a declining rate. Argentina's old age dependency rate will thus increase less dramatically than Hungary's or Japan's. China now has a younger population than the others, but the old age dependency ratio will increase precipitously when the large working age population retires and is replaced by much smaller cohorts of younger workers. Kenya starts out young, with a rapidly growing population, and remains young until after 2030, when today's children will be bearing children and when its demographic transition is projected to begin.
For each country, the projections show the contribution rate required to finance a 40% average benefit rate (that is, the average benefit divided by the average wage) and the lifetime transfer that this type of pension system implies for a program beginning in 1995 (see the figure). Everyone is assumed to start work at age 20 and to retire at age 60. Everyone over 60 receives the same pension---including, at the beginning, those who never contributed. The real average wage in the economy is assumed to rise 1% each year, and wages also grow 1% for every year of experience. Pensions are indexed to the economywide average wage.
In all five countries, the dependency ratio rises throughout the life span of the individuals born in 1995, when the plan is implemented. This effect is greatest in China, where the old age dependency rate nearly quadruples. In all five countries, the result is a sharp rise in contribution rates of later generations, if the benefit rate is held constant at 40%. To maintain a constant 40% benefit rate over this period, the contribution rate would have to double in most countries and more than triple in China. If evasion, administrative expenses, unemployment, disability, survivors' benefits, and early retirement are also taken into account, the required contribution rate would exceed 35% when today's children have retired. This high contribution rate, in addition to the income tax rate, increases the probability that people will evade and retire early, making the situation worse.
The first few generations---the grandparents and parents in 1995---experience large and positive gains (net present value of pension benefits received by each generation minus the net present value of its contribution stream divided by the net present value of its lifetime earnings, assuming a 2% discount rate). This happens because older cohorts contributed for only part of their working lives but receive full pensions, and because they have many young workers to support them. But in Hungary and Japan the net gain turns negative for cohorts who retire in 2035 and 2025, respectively (see the table). In China, children who are born in 1995, the year the program is put in place, and retire in 2055 lose lifetime income. If the discount rate is higher than 2%---which seems plausible---the negative transfer would be greater and would start sooner. These cohorts will never recoup in pensions the present value of the taxes they paid to support the pensions of their parents and grandparents. Evidence from OECD countries, where populations are already aging, is consistent with this pattern of rising tax rates, lower benefit rates, and intergenerational redistributions. Since their systems began many years ago, they are already well along in this process of high tax rates and low returns.
While variables such as retirement age, contribution rate, and expected longevity can change, the underlying pattern remains. When less predictable perturbations---natural disasters, famines, epidemics, immigration, and wars---are superimposed on this pattern, some cohorts fare better and others fare worse than the projections here show. But the trend toward higher old age dependency rates is inevitable, as are higher required contribution rates or lower pension rates and the redistribution from younger to older cohorts for the next 50 years or more, under pay-as-you-go.
