The economic implications of changing age structures

1. The economic implications of changing age structures Ronald Lee University of California at Berkeley Based on research supported by National Institute of Aging

2. The demographic transition • During the demographic transition, typically mortality begins to fall decades before fertility declines. • Initially, families find themselves with more surviving children to feed and educate than before. • Likewise, there are more surviving children in need of public education and health care.

3. Once fertility begins to decline, this process is reversed. Families and governments have fewer children to provide for, and consumption pressures ease. • This leads to the first demographic dividend, a period of declining numbers of consumers per producer. This alone tends to boost the growth of income per capita.

4. The transition continues… • At some point, the number of children per working age person stops declining, and the first dividend, in the sense of a boost to the rate of increase in per capita income, comes to an end. • Around this time, lower fertility also leads to slower growth of working age population while the number of elderly keeps growing rapidly, and speeds up due to falling mortality: Population aging.

5. Population aging raises the number of elderly dependents just as the number of child dependents has fallen. • This puts renewed pressure on consumption and slows the growth of per capita income. But it also may lead to increased accumulation of wealth per capita which generates non-labor income and raises labor productivity. This is the second demographic dividend.

6. The dividends • The first dividend occurs quite mechanically, regardless of institutional setting. Independent variations in productivity growth may reinforce or offset its effects. • The second dividend depends on policies and institutions. • If consumption in old age is provided through transfer programs, the second dividend is weakened or absent. • If old age consumption is funded through asset accumulation, the second dividend does occur.

7. Crucially, the economic effects of population aging depend on institutions and policies.

8. Questions about the dividends • Does consumption of an elderly person cost more or less than that of a child? • How long does the population share of dependent children continue to decline? • How much does the ratio of workers to consumers rise in total and per year? • Are increased resources per consumer used to raise consumption, invest in education of children, or saved for retirement?

9. To answer these questions • We will consider a typical demographic transition. • We will also examine the way consumption and labor income vary with age in different settings. • We will see how old age consumption is paid for in different countries.

10. THE CASE OF INDIA

11. Indian life expectancy began to rise around 1900, here simulated to go from 24 to 80 years. Life expectancy in years Earlier UN projections Simulations based on smooth mathematical trajectories for fertility and mortality Actual data (*) Actual data (*)

12. Indian fertility began to fall around 1960, here simulated to go from 6 to 2.1 births. Children per woman

13. Changes in the child dependency ratio Once fertility begins to decline, the child dependency ratio falls. Then declining fertility reduces the ratio. Increasing survival of children initially raises the ratio Child dependency ratio (<15/15-64)

14. Changes in the old-age dependency ratio • Serious population aging begins more than a century after the transition starts. • The old-age dependency ratio rises rapidly, by a factor of six. Onset of serious population aging is late in the transition Old-age dependency ratio (65+/15-64)

15. Variation in the total dependency ratio Rising dependency as mortality falls Rising dependency as mortality falls The “first dividend”: Dependency falls Population aging Dependency ends where it began. Transitory effect.

16. Variation in the total dependency ratio Rising dependency as mortality falls The “first dividend”: dependency falls At start: Many children and few elderly. At end: Many elderly and few children. This generates the “second dividend”. Population aging

17. How Dependency Ratios Change Over A Classic Demographic Transition:Actual and Projected for India and Simulated, 1900-2100 There is great variation in projected old age dependency ratios for 2050 • Ratio in Japan projected to be 6 times as high as in the least developed countries. • Differences are due to position in transition, baby booms and busts, and fertility below replacement. Japan Southern Europe Europe China USA Least developed countries

18. II. How labor income and consumption vary by age • To understand economic implications of age structures, we need to know how labor income and consumption vary with age. • The National Transfer Accounts project (NTA) is estimating these for many countries.

19. NTA measures labor income as average income at each age, whether the individual is working or not. • It includes fringe benefits and self employment income.

20. Source: Sang-Hyop Lee, report on NTA labor income data

21. Source: Sang-Hyop Lee, report on NTA labor income data

22. Source: Sang-Hyop Lee, report on NTA labor income data

23. Source: Sang-Hyop Lee, report on NTA labor income data

24. Source: Sang-Hyop Lee, report on NTA labor income data

25. Source: Sang-Hyop Lee, report on NTA labor income data

26. Source: Sang-Hyop Lee, report on NTA labor income data

27. Source: Sang-Hyop Lee, report on NTA labor income data

28. Source: Sang-Hyop Lee, report on NTA labor income data

29. Source: Sang-Hyop Lee, report on NTA labor income data

30. Consumption by age • NTA includes private household consumption and also publicly provided education, health care, and other items.

31. A typical LDC example: Thailand, 1998

32. Some MDCs have higher consumption in old age (US, 2003); Public health care is important Source: National Transfer Account data.

33. III. Economic Consequences of Age Distribution: Support ratios • The balance of workers and consumers for the whole population is summarized by the support ratio • Add up population times labor income at each age • Add up population times consumption at each age • Ratio of labor to consumers is the “support ratio”.

34. A high support ratio is favorable. • The increase in the support ratio in the middle of the demographic transition is the “demographic dividend”. • Here are examples for some LDCs at different stages of their transitions.

36. Niger first dividend: 2014-2090, 76 years, increase of 52% .55%/yr

37. First dividend for China and S. Korea are about finished. China: 1971-2013, 42 years, increase of 35% or .7%/yr 2007

38. India is in middle of first dividend phase. Brazil is near the end. 2007

40. For Japan, Spain, Italy, and Germany, the support ratios drop substantially by 2050. For US, less so. For Japan, Spain, Italy, and Germany, the support ratios drop substantially by 2050. For US, less so.

42. For MDCs other than the US, this is an annual decline of .5 to .8%. Compared to expected productivity growth of perhaps 1 to 2% per year, this is significant.

43. IV. Population change, saving, and capital • The first demographic dividend is transitory. • Given the right policies, age structure changes can produce a second demographic dividend which is permanent.

44. The second demographic dividend • Typically, adults accumulate assets over their life cycles. • Thus elderly hold more assets than others. • Population aging raises the population share of elderly, and therefore raises the average amount of wealth and asset income. • More capital per worker also raises labor productivity. • This is the second dividend.

45. The second dividend may be reinforced by demographic change • Longer life requires increased saving for retirement. • Lower fertility may mean higher saving by parents with fewer children. • For these reasons, elderly may accumulate even more wealth. • The second dividend may be larger.

46. This does not mean the aggregate saving rate will rise with aging • Actually, aggregate saving may well fall. • The second dividend still occurs, because with slower labor force growth, even lower saving can raise capital per worker.

47. However, the second dividend depends on institutions. • If elderly are supported by their adult children, then they save less and hold less wealth. • If elderly are supported by unfunded public pensions, then they save less and hold less wealth. • Then second dividend is reduced.

48. How old age consumption is financed in three countries 13% 41% 43% 67% 44% 35% Source: National Transfer Accounts.

49. Important differences across these four countries • Assets account for more than 40% in Thailand and the US, but only 13% in Japan. • With these arrangements, the second dividend will be substantial in Thailand and the US. • In Japan, population aging will just raise the transfer burden on the working age population.

50. Conclusions