A global overview of population aging and
intergenerational transfers, and some
implications for economic growth
Ronald Lee (UC Berkeley)
Santiago, October 20, 2009
NIA R37-AG025488 and MEXT.ACADEMIC
FRONTIER to NUPRI in Japan
Deep thanks to NTA country teams, Gretchen
Donehower, and Andy Mason.
Plan of talk
• Consider some economic consequences
of the demographic transition, including
population aging.
• Draw on results of the National Transfer
Accounts project (NTA) for 23 countries.
• Focus on Latin America.
Ronald Lee, UC Berkeley, 2009
Latin American NTA Country Teams — many
members are at this meeting.
•
Brazil
–
–
–
–
•
Lanza, Queiroz, Bernardo
Marri, Izabel
Renteria, Elisanda Perez
Turra, Cassio
Chile
– Bravo, Jorge
– Holz, Mauricio
•
Costa Rica
– Collado, Andrea
– Rosero-Bixby, Luis
– Zuniga, Paola
•
Mexico
– Ivan Mejia
– Félix Vélez Fernández-Varela
– Juan Enrique
•
Uruguay
– Bucheli, Marisa
– Gonzalez, Cecilia
Ronald Lee, UC Berkeley, 2009
1. Changing population age structure
across the demographic transition: the
case of Mexico, 1900-2100
• Demographic transition is process of
moving from initial high fertility and high
mortality to low fertility and low mortality,
and the accompanying changes in
population size and age composition.
• Most transitions follow a classic pattern,
but Latin America has many exceptions.
Ronald Lee, UC Berkeley, 2009
• In most Third World countries fertility decline
started around 1965 or later.
• Fertility decline started much earlier in some
Latin American countries
– E.g. in 1900 or earlier in Argentina, Uruguay, Cuba,
Chile.
– These declines stalled at mid century, contrary to
classic pattern.
• Mortality decline more typical, starting
sometimes before 1900, some times later.
• Diversity, no single pattern in Latin America.
Ronald Lee, UC Berkeley, 2009
Mexico, 1900-2100
• In some respects classic
– Mortality decline preceded fertility decline by many
decades
– Fertility declined steadily to near replacement
(TFR=2.3) and life expectancy is high (75 years)
• In other respects, not typical
– Mexican Revolution perturbed fertility and mortality in
early 20th century.
– High net outmigration (not so unusual).
• Transition data based on Perez-Brignoli (2009)
reconstruction and on UN data.
Ronald Lee, UC Berkeley, 2009
The demographic transition in Mexico
Total Fertility Rate
Life Expectancy
7
90
5
4
3
2
80
Average Annual Growth Rate
Life Expectancy from Birth
6
Total Fertility Rate
Growth Rate
4
70
60
50
40
30
20
10
1
1900
1950
2000
2050
0
1900
2100
Population Size
2050
2100
2
1
0
-1
1900
1950
2000
2050
2100
Simulated
1
120
Total Population (millions)
Total Population (millions)
2000
Dependency Ratios
140
100
80
60
40
20
0
1900
1950
3
1950
2000
2050
2100
Perez-Brignoli
Total
0.8
UN Historical
0.6
Old (65+)
UN Projected
Young
(<15)
Note that simulation includes
assumption of net outmigration by
ages 15-64 of approximately 0.6%
per year from 1980. (Data driven to
match observed population sizes.)
0.4
0.2
Oldest Old (85+)
0
1900
1950
2000
2050
2100
Ronald Lee, UC Berkeley, 2009
Dependency Ratios
Total Population (millions)
1
Total
0.8
0.6
Old (65+)
0.4
Young
(<15)
0.2
Oldest Old (85+)
0
1900
1950
2000
2050
Ronald Lee, UC Berkeley, 2009
2100
2. Economic behavior varies across the
life cycle – dependency and
surplus production
• The economic life cycle starts with child
dependency and ends with old age
dependency.
– Children consume much more than they
produce, particularly if enrolled in school.
– Elderly consume much more than their labor
earnings.
• NTA estimates patterns of labor income
and consumption over the life cycle.
Ronald Lee, UC Berkeley, 2009
Cross-sectional NTA estimates of
labor income by age.
• Averaged across all people of given age
– male and female
– Employed, self-employed, fringe benefits
– Formal or informal sectors.
– in labor force or not.
• Divided by average labor income for age
30-49 in each country to standardize.
Ronald Lee, UC Berkeley, 2009
NTA Age Profiles of Labor Income for Four Rich and Four
Poor Countries (unweighted averages)
1.2
Children have
more labor income
in poor countries.
Labor Income/Av labor inc 30-49
1
Rich (JP,US,SE,FI)
Poor (ID,IN,PH,KY)
0.8
Income drops
suddenly around
age 60 in rich
countries: Pension
structure.
Income peaks
earlier in poor
countries. .
0.6
0.4
In old age, people
in poor countries
work more.
0.2
0
0
10
20
30
40
50
60
Age
Ronald Lee, UC Berkeley, 2009
70
80
90
100
NTA Age Profiles of Labor Income, for Four Rich, Four Poor,
and Five Latin Ameictran Countries (unweighted averages)
1.2
Latin America (BR,CL,CR,MX,UY)
Latin America
looks similar to
poor countries.
Why?
Labor Income/Av labor inc 30-49
1
Rich (JP,US,SE,FI)
Poor (ID,IN,PH,KY)
0.8
0.6
0.4
0.2
0
0
10
20
30
40
50
60
Age
Ronald Lee, UC Berkeley, 2009
70
80
90
100
Age Profiles of Labor Income
Five Latin American Countries
1.2
BR
CL
CR
MX
UY
1
Ratio to av yl(30-49)
Early decline in Brazil
after 50.
0.8
Related to pension
structure?
0.6
0.4
Lower in Uruguay
above 65..
0.2
Related to pension
structure?
0
0
10
20
30
40
50
Age
Ronald Lee, UC Berkeley, 2009
60
70
80
90
Consumption (cross-sectional NTA
estimates)
• Average for all people of given age
• Includes
– Private expenditures by households, allocated
by age within the household (private health
and education estimated separately).
– Public in-kind transfers (e.g. education, health
care, long-term care).
• Standardized by dividing by average labor
income of each country.
Ronald Lee, UC Berkeley, 2009
NTA Age Profiles of Labor Income and Consumption for
Four Rich and Four Poor Countries
1.4
Rich (US,JP,SE,FI)
1.2
Consumption rises with
age in rich countries,
partly health spending,
mainly public.
Ratio to av yl(30-49)
Poor (ID,IN,PH,KE)
1
0.8
0.6
0.4
Flat consumption profile
in poor from 20 to end
of life.
More human capital
investment in rich
countries.
0.2
Co-resident elders,
household sharing?
0
1
11
21
31
41
51
Age
Ronald Lee, UC Berkeley, 2009
61
71
81
91
Age Profiles of Labor Income and Consumption for
Four Rich, Four Poor, and Five Latin American Countries
1.4
Latin America (BR,CH,CR,MX,UY)
Rich (US,JP,SE,FI)
Ratio to av yl(30-49)
1.2
Level is very high relative to
labor income. Why?
Poor (ID,IN,PH,KE)
1
0.8
0.6
0.4
Shape of LA is more similar
to poor countries.
0.2
0
0
10
20
30
40
50
Age
Ronald Lee, UC Berkeley, 2009
60
70
80
90
Why is the Latin American consumption so
high relative to average labor income?
• Ratio of aggregate consumption to aggregate
labor income in National Income and Product
Accounts is very high in Latin America (5 out of
top 7 among NTA countries).
– Nonlabor income like remittances and natural
resources.
• Probably reflects very low national saving rates
in the region.
• Why is saving low? See Mario Gutiérrez (2007)
CEPAL study on this question.
• Could it be linked to generous public pensions?
Ronald Lee, UC Berkeley, 2009
Age Profiles of Consumption for
Five Latin American Countries
1.2
Ratio to av yl(30-49)
1
0.8
0.6
Consumption rises strongly at
old ages in Brazil.
BR
CL
CR
MX
UY
0.4
0.2
Generous public pensions?
Consumption falls in old age in
Mexico.
Low public pensions?
0
0
10
20
30
40
50
Age
Ronald Lee, UC Berkeley, 2009
60
70
80
90
Age Profiles of Labor Income and Consumption for
Four Rich, Four Poor, and Five Latin American Countries
1.4
Latin America (BR,CH,CR,MX,UY)
Rich (US,JP,SE,FI)
1.2
Ratio to av yl(30-49)
Consumption minus labor
income is “life cycle deficit”.
Poor (ID,IN,PH,KE)
1
0.8
0.6
0.4
0.2
0
0
10
20
30
40
50
Age
Ronald Lee, UC Berkeley, 2009
60
70
80
90
The Life Cycle Deficit (consumption - labor income) in rich,
poor, and Latin American countries
1.4
In Latin America, there is not much
surplus labor income (small age
range and low value), whereas life
cycle deficits are large.
(Consumption - Labor Income)/Av yl(3049)
1.2
1
0.8
0.6
0.4
0.2
0
0
10
20
30
40
50
60
70
80
90
100
-0.2
-0.4
Latin Am
-0.6
-0.8
Age
Ronald Lee, UC Berkeley, 2009
Rich
Poor
3. Age structure and economic
behavior: support ratios
• “Effective labor” is sum over age of
population times labor income.
• “Effective consumers” is similar.
• “Support Ratio” is ratio of effective labor to
effective consumers.
• Other things equal, consumption per
effective consumer is proportional to the
support ratio.
– Similarly for per capita income.
Ronald Lee, UC Berkeley, 2009
Support Ratios for Mexican Population (simulations and UN
projections, with average of 5 Latin American country profiles)
0.80
Aggregate YL / Aggregate C
0.75
0.70
0.65
0.60
0.55
0.50
1950
1970
1990
2010
2030
Ronald Lee, UC Berkeley, 2009
Year
2050
2070
2090
Support Ratios for Mexican Population (simulations and UN
projections, with average of 5 Latin American country profiles)
First dividend period
adds .7% per year to
consumption growth
from 1975 to 2025.
0.80
Aggregate YL / Aggregate C
0.75
0.70
0.65
0.60
0.55
0.50
1950
1970
1990
2010
2030
Ronald Lee, UC Berkeley, 2009
Year
2050
2070
2090
Support Ratios for Mexican Population (simulations and UN
projections, with average of 5 Latin American country profiles)
0.80
Aggregate YL / Aggregate C
0.75
0.70
0.65
Population aging.
Subtracts .4% per
year from
consumption growth
from 2025 to 2060
0.60
0.55
0.50
1950
1970
1990
2010
2030
Ronald Lee, UC Berkeley, 2009
Year
2050
2070
2090
Support Ratios for Mexican Population (simulations and UN
projections, with average of 5 Latin American country profiles)
0.80
Aggregate YL / Aggregate C
0.75
0.70
0.65
Contrast between
dividend phase and
aging is 1.1%/yr
0.60
0.55
0.50
1950
1970
1990
2010
2030
Ronald Lee, UC Berkeley, 2009
Year
2050
2070
2090
0.80
Support Ratios for Latin America
(Own-country populations, average of 5 country profiles)
BR
CL
0.75
CR
Support Ratio
MX
0.70
UY
0.65
MX is same as
last slide
0.60
Uruguay is different because
of its very early fertility
decline.
0.55
0.50
1950
1970
1990
2010
2030
2050
Ronald Lee, UC Berkeley, 2009
Year
2070
2090
0.80
Support Ratios for Latin America
(Own-country populations, average of 5 country profiles)
Approximate demographic
dividend phase
BR
CL
0.75
CR
Support Ratio
MX
0.70
UY
0.65
0.60
Approximate population
aging phase
0.55
0.50
1950
1970
1990
2010
2030
2050
Ronald Lee, UC Berkeley, 2009
Year
2070
2090
0.80
Support Ratios for Latin America
(Own-country populations, average of 5 country profiles)
BR
CL
0.75
CR
Support Ratio
MX
0.70
UY
0.65
Except for Uruguay, the
dividend contributes .50 to
.75 %/yr to consumption
growth 1970-2020.
0.60
0.55
Population aging subtracts
.35 to .40%/yr 2020 to 2070..
0.50
1950
Contrast is around 1%/yr.
1970
1990
2010
2030
2050
Ronald Lee, UC Berkeley, 2009
Year
2070
2090
4. In older populations, more
consumption by elderly people is
funded somehow.
• Illustrate with Indonesia (young) and
Japan (old).
Ronald Lee, UC Berkeley, 2009
P e r c a p ita
c o n s u m p tio n o r la b o r
in c o m e
Per capita consumption and labor income
by age for Indonesia and Japan
1,000,000
Indonesia,
2002
800,000
600,000
400,000
200,000
0
20
40
60
80
100
Per capita consumption
or labor income in Yen
Age
500,000
Japan, 2004
400,000
300,000
200,000
100,000
0
20
40
60
80
Age
Indonesia NTA Maliki; Japan NTA
Ogawa
100
The aggregate lifecycle deficit at each age
(population by age times per capita age profiles)
Aggregate Life Cycle Deficit for Indonesia (2005) in Rupiah
30,000
20,000
10,000
-
0
20
40
60
80
100
(10,000)
(20,000)
(30,000)
Age
Aggregate Life Cycle Deficit for Japan (2004) in Yen
Express the total old age
aggregate life cycle deficit
as a share of aggregate
consumption at all ages.
Measures the importance
of consumption by the
elderly in excess of their
labor income.
5,000
Aggregated Consumption - Labor
Income
Aggregated Consumption - Labor Income
40,000
4,000
3,000
2,000
1,000
(1,000)
0
20
40
60
80
100
(2,000)
(3,000)
(4,000)
(5,000)
Age
Indonesia NTA Maliki; Japan NTA
Ogawa
The older the population,
the greater this share.
Calculate aggregate
deficit as share of total
consumption.
Aggregate old age deficit is funded by
assets and transfers (public and private)
Aggregate Life Cycle Deficit for Indonesia (2005) in Rupiah
Aggregated Consumption Labor Income
40,000
30,000
20,000
10,000
-
0
20
40
60
80
100
(10,000)
(20,000)
(30,000)
Age
Aggregate Life Cycle Deficit for Japan (2004) in Yen
Aggregated Consumption - Labor
Income
5,000
4,000
3,000
2,000
1,000
(1,000)
0
20
40
60
80
100
• Green arrows show
transfers from surplus
of prime working
years.
• Red arrows show
asset income
consumed by elderly
out of earlier savings.
(2,000)
(3,000)
(4,000)
(5,000)
Age
Indonesia NTA Maliki; Japan NTA
Ogawa
LCD 65+ as a Share of Total Consumption
0.25
Aggregate Old Age Consumption Deficit as a Share of Total
Consumption vs Proportion of Population 65+
JP
SE
Kenya: 2%
Japan: 22%
0.2
0.15
ES
SL
AT
HU FI FR
US
UY
Most difference due
to population aging.
Population aging
means more
assets.
0.1
CL TW
TH
BR
SK
CR
CN
IDMXIN
0.05
Old age deficit is
funded partly by
asset accumulation.
KY PH
y = 1.1921x - 0.0189
2
R = 0.9732
0
0
0.05
0.1
0.15
Share of Population 65+
Ronald Lee, UC Berkeley, 2009
0.2
0.25
Implications for capital
accumulation
• Population aging raises need to provide for old
age deficit.
• If met by more asset accumulation, then
population aging raises asset income and
perhaps labor productivity.
• If met by public or private transfers, then
population aging just raises the transfer burden
on workers.
• Policy should find right balance of assets and
transfers.
Ronald Lee, UC Berkeley, 2009
5. How is the elderly life cycle
deficit funded in different countries?
Ronald Lee, UC Berkeley, 2009
Funding the Old-age Deficit in the US
Components of Lifecycle Deficit, US 2003
70000
Asset income and dissaving
60000
Public Asset-Based Reallocations
Private Asset-Based Reallocations
Public Transfers
Private Transfers
50000
40000
Net public transfers (public
pensions, health care, etc.,
less taxes paid).
US $
30000
20000
10000
0
-10000
Net private transfers – interand intra-household
transfers.
-20000
-30000
0
3
6
9
12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60 63 66 69 72 75 78 81 84 87 90
Age
Ronald Lee, UC Berkeley, 2009
Shares of assets, public transfers,
and private transfers, differs from
country to country.
• The % shares of each add to 100%.
• We can show these on a “triangle graph”.
Ronald Lee, UC Berkeley, 2009
Triangle graph shows funding shares of life cycle deficit of elderly (65+,
consumption – labor income): Assets, Public Transfers, Private transfers
At each corner, 100% of deficit is
funded by that item.
PH nearly 100% Assets.
AT nearly 100% Public Transfers
On each side, one item is 0.
TH has 0 public trabnsfers.
JP has 0 private transfers.
Taiwan is near the
middle, with bbbequal
shares (1/3) from each.
Ronald Lee, UC Berkeley, 2009
Triangle graph for funding life cycle deficit of 65+: East Asian countries
show familial support of elderly (except Japan)
Japan relies 2/3 on public
transfers, rest on assets.
Family transfers are 0!
Other E. Asian rely less
than 1/3 on public transfers.
China relies 60% on family.
Ronald Lee, UC Berkeley, 2009
Triangle graph for funding life cycle deficit of 65+: Latin
American elderly make private transfers to others.
Brazil.s public transfers to
elderly are greater than any
other country’s.
Brazil, Mexico, and Uruguay
make larger downward
transfers than any other.
Mexico relies mostly
on assets. Why?
Mexican households
own no more assets
than in other LA
countries.
They use assets for
consumption
because unlike the
other NTA countries,
public pensions are
small.
Ronald Lee, UC Berkeley, 2009
Triangle graph for funding life cycle deficit of 65+: in US
and Europe elderly also transfer to others.
US relies less on public
transfers, more on
assets.
In Japan, younger
elderly make family
transfers, older elderly
receive them.
Ronald Lee, UC Berkeley, 2009
6. The demographic transition also
promotes investment in human
capital.
Ronald Lee, UC Berkeley, 2009
Human Capital
• Parents choose between number of children and
amount to invest per child (Quantity-quality
tradeoff)
• As economies develop parents opt for fewer
children and spend more per child (Becker;
Becker and Lewis, Willis).
• Aging (low fertility) will be accompanied by more
human capital, regardless of causal direction.
• The human capital “response” helps to offset the
negative effect of population aging on the
support ratio.
Ronald Lee, UC Berkeley, 2009
Empirical Relationship between
Human Capital and Fertility
• NTA measure of Human Capital (HK) investment
– NTA measures public and private spending per
capita at each age for health and for education
– Sum these for ages 0-17 for health and 0-26 for
education
– Normalize on average labor income ages 30-49
• Compare HK to Fertility in preceding five years
Ronald Lee, UC Berkeley, 2009
Cross-sectional Relationship
7
6
SE
JP
SI
TW
Estimated elasticity
d ln HK / d ln TFR
is -0.913
Human capital
5
MX
FR
ESKR US
AT
HU FI
BR
TH CR
DE
CL
UY
4
3
CN
2
PH
ID
IN
1
KE
0
0
1
2
3
Total Fertility Rate
Source: Lee and Mason, forthcoming,
European Journal of Population (2009).
Ronald Lee, UC Berkeley, 2009
4
5
6
Time Series Relationship
7
Number of
Observations
Japan
5
Taiwan
27
United States 23
6
Human capital spending
Estimated elasticities
Japan
-1.46
Taiwan
-1.40
United States -0.72
Taiwan 1977-2003
5
Japan 1984-2004
4
3
US 1960-2003
2
1
0
0
1
2
Total Fertility Rate
Ronald Lee, UC Berkeley, 2009
3
4
Population aging is accompanied by
increased investments in HK of children
• Raises the productivity and earnings of
labor force in future
• Substitutes HK for number of workers
• Offsets falling support ratios.
Ronald Lee, UC Berkeley, 2009
7. Policy can alleviate the economic
impact of population aging
• Population aging will certainly cause problems.
– Fiscal sustainability of public programs is a huge
problem for many Latin American countries. I did not
discuss this.
– Population aging will cause falling support ratios,
which will reduce consumption, other things equal, by
1%/yr relative to dividend phase.
– Some risk that that fiscal pressures of aging will
crowd out human capital investments in children.
Ronald Lee, UC Berkeley, 2009
But…
• Population aging raises the demand for capital by raising
the share of old age consumption in the economy.
– If old age consumption is funded at least in part by assets
rather than transfers, then population aging will lead to higher
capital/labor ratio.
– Policy should find the right balance between unfunded
public pensions and funded programs, either public or
private.
– Avoid mistakes of current industrial nations in this regard.
• The low fertility that causes population aging also goes
with increased HK investment in children, and higher
labor productivity in the future. Policy should support
investments in human capital of children.
Ronald Lee, UC Berkeley, 2009