Life Cycle Earnings in a Fast-Growing Economy
Zheng Michael SongDennis Tao Yang
Discussion by Christopher CarrollJanuary 7, 2011
Song and Yang China’s Saving Rate
Bottom Line
Very interesting paper; fabulous new data from China
“Highly Preliminary” – as is my critique.
Previous literature: Habit formation, precautionary saving,capital market imperfections, or other factors are needed toexplain remarkable increase in Chinese saving rates
Authors: It is explainable by an old-fashioned life cycle model
Intertemporally separable utility (no habits)Perfect foresight about income growthNo important capital market imperfections
Just need to account for changing pattern of income by age
Song and Yang China’s Saving Rate
Bottom Line
Very interesting paper; fabulous new data from China
“Highly Preliminary” – as is my critique.
Previous literature: Habit formation, precautionary saving,capital market imperfections, or other factors are needed toexplain remarkable increase in Chinese saving rates
Authors: It is explainable by an old-fashioned life cycle model
Intertemporally separable utility (no habits)Perfect foresight about income growthNo important capital market imperfections
Just need to account for changing pattern of income by age
Song and Yang China’s Saving Rate
Bottom Line
Very interesting paper; fabulous new data from China
“Highly Preliminary” – as is my critique.
Previous literature: Habit formation, precautionary saving,capital market imperfections, or other factors are needed toexplain remarkable increase in Chinese saving rates
Authors: It is explainable by an old-fashioned life cycle model
Intertemporally separable utility (no habits)Perfect foresight about income growthNo important capital market imperfections
Just need to account for changing pattern of income by age
Song and Yang China’s Saving Rate
Bottom Line
Very interesting paper; fabulous new data from China
“Highly Preliminary” – as is my critique.
Previous literature: Habit formation, precautionary saving,capital market imperfections, or other factors are needed toexplain remarkable increase in Chinese saving rates
Authors: It is explainable by an old-fashioned life cycle model
Intertemporally separable utility (no habits)Perfect foresight about income growthNo important capital market imperfections
Just need to account for changing pattern of income by age
Song and Yang China’s Saving Rate
Bottom Line
Very interesting paper; fabulous new data from China
“Highly Preliminary” – as is my critique.
Previous literature: Habit formation, precautionary saving,capital market imperfections, or other factors are needed toexplain remarkable increase in Chinese saving rates
Authors: It is explainable by an old-fashioned life cycle model
Intertemporally separable utility (no habits)Perfect foresight about income growthNo important capital market imperfections
Just need to account for changing pattern of income by age
Song and Yang China’s Saving Rate
Bottom Line
Very interesting paper; fabulous new data from China
“Highly Preliminary” – as is my critique.
Previous literature: Habit formation, precautionary saving,capital market imperfections, or other factors are needed toexplain remarkable increase in Chinese saving rates
Authors: It is explainable by an old-fashioned life cycle model
Intertemporally separable utility (no habits)Perfect foresight about income growthNo important capital market imperfections
Just need to account for changing pattern of income by age
Song and Yang China’s Saving Rate
Bottom Line
Very interesting paper; fabulous new data from China
“Highly Preliminary” – as is my critique.
Previous literature: Habit formation, precautionary saving,capital market imperfections, or other factors are needed toexplain remarkable increase in Chinese saving rates
Authors: It is explainable by an old-fashioned life cycle model
Intertemporally separable utility (no habits)Perfect foresight about income growthNo important capital market imperfections
Just need to account for changing pattern of income by age
Song and Yang China’s Saving Rate
Bottom Line
Very interesting paper; fabulous new data from China
“Highly Preliminary” – as is my critique.
Previous literature: Habit formation, precautionary saving,capital market imperfections, or other factors are needed toexplain remarkable increase in Chinese saving rates
Authors: It is explainable by an old-fashioned life cycle model
Intertemporally separable utility (no habits)Perfect foresight about income growthNo important capital market imperfections
Just need to account for changing pattern of income by age
Song and Yang China’s Saving Rate
Depends on What the Meaning of Is Is
My interpetation: Chinese behavior is not explained credibly by anoff-the-shelf model
By “off the shelf” I mean a model with:
Preferences that are stable over time and across countries
Time preference rate; intertemporal elasticity; bequest motive
Plausible assumptions about expectations of income growth
Why “stable” preferences?
Arbitrary variation in preferences can explain anything
What are “plausible” expectations about income growth?
Forecasts that could plausibly have been made at the time
Agents not allowed “sneak peek” at future outcomes (notperfect foresight)
Song and Yang China’s Saving Rate
Depends on What the Meaning of Is Is
My interpetation: Chinese behavior is not explained credibly by anoff-the-shelf model
By “off the shelf” I mean a model with:
Preferences that are stable over time and across countries
Time preference rate; intertemporal elasticity; bequest motive
Plausible assumptions about expectations of income growth
Why “stable” preferences?
Arbitrary variation in preferences can explain anything
What are “plausible” expectations about income growth?
Forecasts that could plausibly have been made at the time
Agents not allowed “sneak peek” at future outcomes (notperfect foresight)
Song and Yang China’s Saving Rate
Depends on What the Meaning of Is Is
My interpetation: Chinese behavior is not explained credibly by anoff-the-shelf model
By “off the shelf” I mean a model with:
Preferences that are stable over time and across countries
Time preference rate; intertemporal elasticity; bequest motive
Plausible assumptions about expectations of income growth
Why “stable” preferences?
Arbitrary variation in preferences can explain anything
What are “plausible” expectations about income growth?
Forecasts that could plausibly have been made at the time
Agents not allowed “sneak peek” at future outcomes (notperfect foresight)
Song and Yang China’s Saving Rate
Depends on What the Meaning of Is Is
My interpetation: Chinese behavior is not explained credibly by anoff-the-shelf model
By “off the shelf” I mean a model with:
Preferences that are stable over time and across countries
Time preference rate; intertemporal elasticity; bequest motive
Plausible assumptions about expectations of income growth
Why “stable” preferences?
Arbitrary variation in preferences can explain anything
What are “plausible” expectations about income growth?
Forecasts that could plausibly have been made at the time
Agents not allowed “sneak peek” at future outcomes (notperfect foresight)
Song and Yang China’s Saving Rate
Depends on What the Meaning of Is Is
My interpetation: Chinese behavior is not explained credibly by anoff-the-shelf model
By “off the shelf” I mean a model with:
Preferences that are stable over time and across countries
Time preference rate; intertemporal elasticity; bequest motive
Plausible assumptions about expectations of income growth
Why “stable” preferences?
Arbitrary variation in preferences can explain anything
What are “plausible” expectations about income growth?
Forecasts that could plausibly have been made at the time
Agents not allowed “sneak peek” at future outcomes (notperfect foresight)
Song and Yang China’s Saving Rate
Depends on What the Meaning of Is Is
My interpetation: Chinese behavior is not explained credibly by anoff-the-shelf model
By “off the shelf” I mean a model with:
Preferences that are stable over time and across countries
Time preference rate; intertemporal elasticity; bequest motive
Plausible assumptions about expectations of income growth
Why “stable” preferences?
Arbitrary variation in preferences can explain anything
What are “plausible” expectations about income growth?
Forecasts that could plausibly have been made at the time
Agents not allowed “sneak peek” at future outcomes (notperfect foresight)
Song and Yang China’s Saving Rate
What version of a LC model is consistent with their data?
1 Today’s Chinese HH’s are much more pessimistic about theirown personal income growth than today’s American HH’s
2 HH’s in 2007 believed income growth in the future would belower than households believed growth would be in 1992.
In 15 years to 1992, avg income growth was a remarkable 5pct per yearin 15 years to 2007, avg income growth was an amazing 8 pctper year
1992-2007 certainly fastest sustained growth in Chinese history
Required beliefs: HH’s in 2007 (after astounding 1992-2007growth) are much more pessimistic about growth than HH’s atsimilar ages were in 1992
3 HH’s in 1992 were more pessimistic about income growththan HH’s in 1978
Economists have been accused of being willing to believe 8impossible things before breakfast. But it is slightly after breakfastnow, so you might be excused if you join me in not believing thesethings.
Song and Yang China’s Saving Rate
What version of a LC model is consistent with their data?
1 Today’s Chinese HH’s are much more pessimistic about theirown personal income growth than today’s American HH’s
2 HH’s in 2007 believed income growth in the future would belower than households believed growth would be in 1992.
In 15 years to 1992, avg income growth was a remarkable 5pct per yearin 15 years to 2007, avg income growth was an amazing 8 pctper year
1992-2007 certainly fastest sustained growth in Chinese history
Required beliefs: HH’s in 2007 (after astounding 1992-2007growth) are much more pessimistic about growth than HH’s atsimilar ages were in 1992
3 HH’s in 1992 were more pessimistic about income growththan HH’s in 1978
Economists have been accused of being willing to believe 8impossible things before breakfast. But it is slightly after breakfastnow, so you might be excused if you join me in not believing thesethings.
Song and Yang China’s Saving Rate
What version of a LC model is consistent with their data?
1 Today’s Chinese HH’s are much more pessimistic about theirown personal income growth than today’s American HH’s
2 HH’s in 2007 believed income growth in the future would belower than households believed growth would be in 1992.
In 15 years to 1992, avg income growth was a remarkable 5pct per yearin 15 years to 2007, avg income growth was an amazing 8 pctper year
1992-2007 certainly fastest sustained growth in Chinese history
Required beliefs: HH’s in 2007 (after astounding 1992-2007growth) are much more pessimistic about growth than HH’s atsimilar ages were in 1992
3 HH’s in 1992 were more pessimistic about income growththan HH’s in 1978
Economists have been accused of being willing to believe 8impossible things before breakfast. But it is slightly after breakfastnow, so you might be excused if you join me in not believing thesethings.
Song and Yang China’s Saving Rate
What version of a LC model is consistent with their data?
1 Today’s Chinese HH’s are much more pessimistic about theirown personal income growth than today’s American HH’s
2 HH’s in 2007 believed income growth in the future would belower than households believed growth would be in 1992.
In 15 years to 1992, avg income growth was a remarkable 5pct per yearin 15 years to 2007, avg income growth was an amazing 8 pctper year
1992-2007 certainly fastest sustained growth in Chinese history
Required beliefs: HH’s in 2007 (after astounding 1992-2007growth) are much more pessimistic about growth than HH’s atsimilar ages were in 1992
3 HH’s in 1992 were more pessimistic about income growththan HH’s in 1978
Economists have been accused of being willing to believe 8impossible things before breakfast. But it is slightly after breakfastnow, so you might be excused if you join me in not believing thesethings.
Song and Yang China’s Saving Rate
What version of a LC model is consistent with their data?
1 Today’s Chinese HH’s are much more pessimistic about theirown personal income growth than today’s American HH’s
2 HH’s in 2007 believed income growth in the future would belower than households believed growth would be in 1992.
In 15 years to 1992, avg income growth was a remarkable 5pct per yearin 15 years to 2007, avg income growth was an amazing 8 pctper year
1992-2007 certainly fastest sustained growth in Chinese history
Required beliefs: HH’s in 2007 (after astounding 1992-2007growth) are much more pessimistic about growth than HH’s atsimilar ages were in 1992
3 HH’s in 1992 were more pessimistic about income growththan HH’s in 1978
Economists have been accused of being willing to believe 8impossible things before breakfast. But it is slightly after breakfastnow, so you might be excused if you join me in not believing thesethings.
Song and Yang China’s Saving Rate
What version of a LC model is consistent with their data?
1 Today’s Chinese HH’s are much more pessimistic about theirown personal income growth than today’s American HH’s
2 HH’s in 2007 believed income growth in the future would belower than households believed growth would be in 1992.
In 15 years to 1992, avg income growth was a remarkable 5pct per yearin 15 years to 2007, avg income growth was an amazing 8 pctper year
1992-2007 certainly fastest sustained growth in Chinese history
Required beliefs: HH’s in 2007 (after astounding 1992-2007growth) are much more pessimistic about growth than HH’s atsimilar ages were in 1992
3 HH’s in 1992 were more pessimistic about income growththan HH’s in 1978
Economists have been accused of being willing to believe 8impossible things before breakfast. But it is slightly after breakfastnow, so you might be excused if you join me in not believing thesethings.
Song and Yang China’s Saving Rate
What version of a LC model is consistent with their data?
1 Today’s Chinese HH’s are much more pessimistic about theirown personal income growth than today’s American HH’s
2 HH’s in 2007 believed income growth in the future would belower than households believed growth would be in 1992.
In 15 years to 1992, avg income growth was a remarkable 5pct per yearin 15 years to 2007, avg income growth was an amazing 8 pctper year
1992-2007 certainly fastest sustained growth in Chinese history
Required beliefs: HH’s in 2007 (after astounding 1992-2007growth) are much more pessimistic about growth than HH’s atsimilar ages were in 1992
3 HH’s in 1992 were more pessimistic about income growththan HH’s in 1978
Economists have been accused of being willing to believe 8impossible things before breakfast. But it is slightly after breakfastnow, so you might be excused if you join me in not believing thesethings.
Song and Yang China’s Saving Rate
What version of a LC model is consistent with their data?
1 Today’s Chinese HH’s are much more pessimistic about theirown personal income growth than today’s American HH’s
2 HH’s in 2007 believed income growth in the future would belower than households believed growth would be in 1992.
In 15 years to 1992, avg income growth was a remarkable 5pct per yearin 15 years to 2007, avg income growth was an amazing 8 pctper year
1992-2007 certainly fastest sustained growth in Chinese history
Required beliefs: HH’s in 2007 (after astounding 1992-2007growth) are much more pessimistic about growth than HH’s atsimilar ages were in 1992
3 HH’s in 1992 were more pessimistic about income growththan HH’s in 1978
Economists have been accused of being willing to believe 8impossible things before breakfast. But it is slightly after breakfastnow, so you might be excused if you join me in not believing thesethings.
Song and Yang China’s Saving Rate
What version of a LC model is consistent with their data?
1 Today’s Chinese HH’s are much more pessimistic about theirown personal income growth than today’s American HH’s
2 HH’s in 2007 believed income growth in the future would belower than households believed growth would be in 1992.
In 15 years to 1992, avg income growth was a remarkable 5pct per yearin 15 years to 2007, avg income growth was an amazing 8 pctper year
1992-2007 certainly fastest sustained growth in Chinese history
Required beliefs: HH’s in 2007 (after astounding 1992-2007growth) are much more pessimistic about growth than HH’s atsimilar ages were in 1992
3 HH’s in 1992 were more pessimistic about income growththan HH’s in 1978
Economists have been accused of being willing to believe 8impossible things before breakfast. But it is slightly after breakfastnow, so you might be excused if you join me in not believing thesethings.
Song and Yang China’s Saving Rate
Accounting for Growth
Alternatives:1 Individual (idiosyncratic)
Affects only one specific household/person
2 Generational (cohort)
Affects similarly every household/person born at the same date
3 Societal (time)
Affects every household/person same way at same time
For aggregate questions (‘national saving rate’), (2) and (3) are key
Song and Yang China’s Saving Rate
Accounting for Growth
Alternatives:1 Individual (idiosyncratic)
Affects only one specific household/person
2 Generational (cohort)
Affects similarly every household/person born at the same date
3 Societal (time)
Affects every household/person same way at same time
For aggregate questions (‘national saving rate’), (2) and (3) are key
Song and Yang China’s Saving Rate
Accounting for Growth
Alternatives:1 Individual (idiosyncratic)
Affects only one specific household/person
2 Generational (cohort)
Affects similarly every household/person born at the same date
3 Societal (time)
Affects every household/person same way at same time
For aggregate questions (‘national saving rate’), (2) and (3) are key
Song and Yang China’s Saving Rate
Accounting for Growth
Alternatives:1 Individual (idiosyncratic)
Affects only one specific household/person
2 Generational (cohort)
Affects similarly every household/person born at the same date
3 Societal (time)
Affects every household/person same way at same time
For aggregate questions (‘national saving rate’), (2) and (3) are key
Song and Yang China’s Saving Rate
Accounting for Growth
Alternatives:1 Individual (idiosyncratic)
Affects only one specific household/person
2 Generational (cohort)
Affects similarly every household/person born at the same date
3 Societal (time)
Affects every household/person same way at same time
For aggregate questions (‘national saving rate’), (2) and (3) are key
Song and Yang China’s Saving Rate
Accounting for Growth
Alternatives:1 Individual (idiosyncratic)
Affects only one specific household/person
2 Generational (cohort)
Affects similarly every household/person born at the same date
3 Societal (time)
Affects every household/person same way at same time
For aggregate questions (‘national saving rate’), (2) and (3) are key
Song and Yang China’s Saving Rate
Accounting for Growth
Alternatives:1 Individual (idiosyncratic)
Affects only one specific household/person
2 Generational (cohort)
Affects similarly every household/person born at the same date
3 Societal (time)
Affects every household/person same way at same time
For aggregate questions (‘national saving rate’), (2) and (3) are key
Song and Yang China’s Saving Rate
Accounting for Growth
Alternatives:1 Individual (idiosyncratic)
Affects only one specific household/person
2 Generational (cohort)
Affects similarly every household/person born at the same date
3 Societal (time)
Affects every household/person same way at same time
For aggregate questions (‘national saving rate’), (2) and (3) are key
Song and Yang China’s Saving Rate
Growth Decomposition
Friedman-definition “permanent income”: Avg for “similar” people
For person age 25 in 1992:
ppp1992,25 = Date1992 + GenerationY + Wrinkles1992
For person age 25 in 2007:
ppp2007,25 = Date2007 + GenerationZ + Wrinkles2007
For person age 40 in 2007
ppp2007,40 = Date2007 + GenerationY + Wrinkles2007
Data are cross-section, both date and wrinkles have changed:
same person︷ ︸︸ ︷ppp2007,40 − ppp1992,25 = ∆Date +
=0︷ ︸︸ ︷∆Generation +∆Wrinkles
Song and Yang China’s Saving Rate
Growth Decomposition
Friedman-definition “permanent income”: Avg for “similar” people
For person age 25 in 1992:
ppp1992,25 = Date1992 + GenerationY + Wrinkles1992
For person age 25 in 2007:
ppp2007,25 = Date2007 + GenerationZ + Wrinkles2007
For person age 40 in 2007
ppp2007,40 = Date2007 + GenerationY + Wrinkles2007
Data are cross-section, both date and wrinkles have changed:
same person︷ ︸︸ ︷ppp2007,40 − ppp1992,25 = ∆Date +
=0︷ ︸︸ ︷∆Generation +∆Wrinkles
Song and Yang China’s Saving Rate
Growth Decomposition
Friedman-definition “permanent income”: Avg for “similar” people
For person age 25 in 1992:
ppp1992,25 = Date1992 + GenerationY + Wrinkles1992
For person age 25 in 2007:
ppp2007,25 = Date2007 + GenerationZ + Wrinkles2007
For person age 40 in 2007
ppp2007,40 = Date2007 + GenerationY + Wrinkles2007
Data are cross-section, both date and wrinkles have changed:
same person︷ ︸︸ ︷ppp2007,40 − ppp1992,25 = ∆Date +
=0︷ ︸︸ ︷∆Generation +∆Wrinkles
Song and Yang China’s Saving Rate
Growth Decomposition
Friedman-definition “permanent income”: Avg for “similar” people
For person age 25 in 1992:
ppp1992,25 = Date1992 + GenerationY + Wrinkles1992
For person age 25 in 2007:
ppp2007,25 = Date2007 + GenerationZ + Wrinkles2007
For person age 40 in 2007
ppp2007,40 = Date2007 + GenerationY + Wrinkles2007
Data are cross-section, both date and wrinkles have changed:
same person︷ ︸︸ ︷ppp2007,40 − ppp1992,25 = ∆Date +
=0︷ ︸︸ ︷∆Generation +∆Wrinkles
Song and Yang China’s Saving Rate
Growth Decomposition
Friedman-definition “permanent income”: Avg for “similar” people
For person age 25 in 1992:
ppp1992,25 = Date1992 + GenerationY + Wrinkles1992
For person age 25 in 2007:
ppp2007,25 = Date2007 + GenerationZ + Wrinkles2007
For person age 40 in 2007
ppp2007,40 = Date2007 + GenerationY + Wrinkles2007
Data are cross-section, both date and wrinkles have changed:
same person︷ ︸︸ ︷ppp2007,40 − ppp1992,25 = ∆Date +
=0︷ ︸︸ ︷∆Generation +∆Wrinkles
Song and Yang China’s Saving Rate
Why It Matters for Consumption
LC/PIH says C is based on expectations about lifetime income
Perfect foresight infinite horizon constant growth:
Ct = (r − ρ−1(r − ϑ))(Bt + Ht)
Ht =Pt
r − γ
Summers (1981)/Tobin (1967): Human Wealth effect is huge(n.b.: If r < γ, Ht =∞)
Song and Yang China’s Saving Rate
Why It Matters for Consumption
LC/PIH says C is based on expectations about lifetime income
Perfect foresight infinite horizon constant growth:
Ct = (r − ρ−1(r − ϑ))(Bt + Ht)
Ht =Pt
r − γ
Summers (1981)/Tobin (1967): Human Wealth effect is huge(n.b.: If r < γ, Ht =∞)
Song and Yang China’s Saving Rate
Why It Matters for Consumption
LC/PIH says C is based on expectations about lifetime income
Perfect foresight infinite horizon constant growth:
Ct = (r − ρ−1(r − ϑ))(Bt + Ht)
Ht =Pt
r − γ
Summers (1981)/Tobin (1967): Human Wealth effect is huge(n.b.: If r < γ, Ht =∞)
Song and Yang China’s Saving Rate
Big Problem: G ↑⇒ S ↑
Cross-section: Modigliani (1970s, 1986)
Granger-causality: Carroll and Weil (1994), Attanasio et al(2000), Rodrik (2000)
Lots of other papers now (cf. Gourinchas and Jeanne (2007))
More convincing: ‘Narrative approach’
Japan, Korea, Hong Kong, Taiwan, Singapore, ....
OECD countries post-1974
Japan post-1990
Song and Yang China’s Saving Rate
Big Problem: G ↑⇒ S ↑
Cross-section: Modigliani (1970s, 1986)
Granger-causality: Carroll and Weil (1994), Attanasio et al(2000), Rodrik (2000)
Lots of other papers now (cf. Gourinchas and Jeanne (2007))
More convincing: ‘Narrative approach’
Japan, Korea, Hong Kong, Taiwan, Singapore, ....
OECD countries post-1974
Japan post-1990
Song and Yang China’s Saving Rate
Big Problem: G ↑⇒ S ↑
Cross-section: Modigliani (1970s, 1986)
Granger-causality: Carroll and Weil (1994), Attanasio et al(2000), Rodrik (2000)
Lots of other papers now (cf. Gourinchas and Jeanne (2007))
More convincing: ‘Narrative approach’
Japan, Korea, Hong Kong, Taiwan, Singapore, ....
OECD countries post-1974
Japan post-1990
Song and Yang China’s Saving Rate
Big Problem: G ↑⇒ S ↑
Cross-section: Modigliani (1970s, 1986)
Granger-causality: Carroll and Weil (1994), Attanasio et al(2000), Rodrik (2000)
Lots of other papers now (cf. Gourinchas and Jeanne (2007))
More convincing: ‘Narrative approach’
Japan, Korea, Hong Kong, Taiwan, Singapore, ....
OECD countries post-1974
Japan post-1990
Song and Yang China’s Saving Rate
Big Problem: G ↑⇒ S ↑
Cross-section: Modigliani (1970s, 1986)
Granger-causality: Carroll and Weil (1994), Attanasio et al(2000), Rodrik (2000)
Lots of other papers now (cf. Gourinchas and Jeanne (2007))
More convincing: ‘Narrative approach’
Japan, Korea, Hong Kong, Taiwan, Singapore, ....
OECD countries post-1974
Japan post-1990
Song and Yang China’s Saving Rate
Big Problem: G ↑⇒ S ↑
Cross-section: Modigliani (1970s, 1986)
Granger-causality: Carroll and Weil (1994), Attanasio et al(2000), Rodrik (2000)
Lots of other papers now (cf. Gourinchas and Jeanne (2007))
More convincing: ‘Narrative approach’
Japan, Korea, Hong Kong, Taiwan, Singapore, ....
OECD countries post-1974
Japan post-1990
Song and Yang China’s Saving Rate
Escape? Modigliani
Suppose aggregate growth has no effect on individual growth
Compare two countries:
A has zero pty growth
C has 8 percent pty growth
Compare two individuals in those countries:
Age 60 in 2010
60-year-olds in both have same income they did at age 25
25-year-olds in both countries expect same growth that25-year-olds expected in 1975
Agg growth all comes from higher level of initial income uponentry at 25
Song and Yang China’s Saving Rate
Escape? Modigliani
Suppose aggregate growth has no effect on individual growth
Compare two countries:
A has zero pty growth
C has 8 percent pty growth
Compare two individuals in those countries:
Age 60 in 2010
60-year-olds in both have same income they did at age 25
25-year-olds in both countries expect same growth that25-year-olds expected in 1975
Agg growth all comes from higher level of initial income uponentry at 25
Song and Yang China’s Saving Rate
Escape? Modigliani
Suppose aggregate growth has no effect on individual growth
Compare two countries:
A has zero pty growth
C has 8 percent pty growth
Compare two individuals in those countries:
Age 60 in 2010
60-year-olds in both have same income they did at age 25
25-year-olds in both countries expect same growth that25-year-olds expected in 1975
Agg growth all comes from higher level of initial income uponentry at 25
Song and Yang China’s Saving Rate
Escape? Modigliani
Suppose aggregate growth has no effect on individual growth
Compare two countries:
A has zero pty growth
C has 8 percent pty growth
Compare two individuals in those countries:
Age 60 in 2010
60-year-olds in both have same income they did at age 25
25-year-olds in both countries expect same growth that25-year-olds expected in 1975
Agg growth all comes from higher level of initial income uponentry at 25
Song and Yang China’s Saving Rate
Escape? Modigliani
Suppose aggregate growth has no effect on individual growth
Compare two countries:
A has zero pty growth
C has 8 percent pty growth
Compare two individuals in those countries:
Age 60 in 2010
60-year-olds in both have same income they did at age 25
25-year-olds in both countries expect same growth that25-year-olds expected in 1975
Agg growth all comes from higher level of initial income uponentry at 25
Song and Yang China’s Saving Rate
Escape? Modigliani
Suppose aggregate growth has no effect on individual growth
Compare two countries:
A has zero pty growth
C has 8 percent pty growth
Compare two individuals in those countries:
Age 60 in 2010
60-year-olds in both have same income they did at age 25
25-year-olds in both countries expect same growth that25-year-olds expected in 1975
Agg growth all comes from higher level of initial income uponentry at 25
Song and Yang China’s Saving Rate
Escape? Modigliani
Suppose aggregate growth has no effect on individual growth
Compare two countries:
A has zero pty growth
C has 8 percent pty growth
Compare two individuals in those countries:
Age 60 in 2010
60-year-olds in both have same income they did at age 25
25-year-olds in both countries expect same growth that25-year-olds expected in 1975
Agg growth all comes from higher level of initial income uponentry at 25
Song and Yang China’s Saving Rate
Implied Cross-Section of (Lifetime) Income By Age
� A
� C
25 30 35 40 45 50 55 60Age
0.2
0.4
0.6
0.8
1.0Income
Cross Section Age�Earnings Profile
Song and Yang China’s Saving Rate
Problem: Modigliani/Friedman ⇒ C = P
Song and Yang China’s Saving Rate
Problem: Modigliani/Friedman ⇒ C = P
Song and Yang China’s Saving Rate
Simplest Alternative: It’s all date effects⇒ similar cross-section age-income across countries
Empirically, roughly true (certainly better than Modigliani)
Song and Yang China’s Saving Rate
Simplest Alternative: It’s all date effects⇒ similar cross-section age-income across countries
Empirically, roughly true (certainly better than Modigliani)
Song and Yang China’s Saving Rate
Simplest Alternative: It’s all date effects⇒ similar cross-section age-income across countries
Empirically, roughly true (certainly better than Modigliani)
Song and Yang China’s Saving Rate
Does Precautionary Saving Change Things?
No. Carroll and Toche (2009):
a ≈(
1
(γ − r) + ϑ(1 + (γ/f)(1− (γ/f)ω))
)(1)
and staring at this (or solving models) convinces you thatγ ↑⇒ a ↓ (Summers was right)
... but if growth increases are always accompanied by increases inf, then maybe ...
Carroll and Jeanne (2009)
Sandri (2009)
Song and Yang China’s Saving Rate
Does Precautionary Saving Change Things?
No. Carroll and Toche (2009):
a ≈(
1
(γ − r) + ϑ(1 + (γ/f)(1− (γ/f)ω))
)(1)
and staring at this (or solving models) convinces you thatγ ↑⇒ a ↓ (Summers was right)
... but if growth increases are always accompanied by increases inf, then maybe ...
Carroll and Jeanne (2009)
Sandri (2009)
Song and Yang China’s Saving Rate
Does Precautionary Saving Change Things?
No. Carroll and Toche (2009):
a ≈(
1
(γ − r) + ϑ(1 + (γ/f)(1− (γ/f)ω))
)(1)
and staring at this (or solving models) convinces you thatγ ↑⇒ a ↓ (Summers was right)
... but if growth increases are always accompanied by increases inf, then maybe ...
Carroll and Jeanne (2009)
Sandri (2009)
Song and Yang China’s Saving Rate
Authors’ story: China is Modigliani on Steriods!
Show clockwise rotation in cross-section profile in ChinaCurrent 25-year-olds
Much higher income level than their parents hadSimilar income at 25 to current 55-year-oldsBut expect worse income growth than their parents expectedShould show implied beliefs about absolute income growthover lifetime, not cross-section
Case:
Pension system phase-outHealth care system privatizedNo kids to take care of youPeople just look to older HHs, don’t think about growth
Song and Yang China’s Saving Rate
Authors’ story: China is Modigliani on Steriods!
Show clockwise rotation in cross-section profile in ChinaCurrent 25-year-olds
Much higher income level than their parents hadSimilar income at 25 to current 55-year-oldsBut expect worse income growth than their parents expectedShould show implied beliefs about absolute income growthover lifetime, not cross-section
Case:
Pension system phase-outHealth care system privatizedNo kids to take care of youPeople just look to older HHs, don’t think about growth
Song and Yang China’s Saving Rate
Authors’ story: China is Modigliani on Steriods!
Show clockwise rotation in cross-section profile in ChinaCurrent 25-year-olds
Much higher income level than their parents hadSimilar income at 25 to current 55-year-oldsBut expect worse income growth than their parents expectedShould show implied beliefs about absolute income growthover lifetime, not cross-section
Case:
Pension system phase-outHealth care system privatizedNo kids to take care of youPeople just look to older HHs, don’t think about growth
Song and Yang China’s Saving Rate
Authors’ story: China is Modigliani on Steriods!
Show clockwise rotation in cross-section profile in ChinaCurrent 25-year-olds
Much higher income level than their parents hadSimilar income at 25 to current 55-year-oldsBut expect worse income growth than their parents expectedShould show implied beliefs about absolute income growthover lifetime, not cross-section
Case:
Pension system phase-outHealth care system privatizedNo kids to take care of youPeople just look to older HHs, don’t think about growth
Song and Yang China’s Saving Rate
Authors’ story: China is Modigliani on Steriods!
Show clockwise rotation in cross-section profile in ChinaCurrent 25-year-olds
Much higher income level than their parents hadSimilar income at 25 to current 55-year-oldsBut expect worse income growth than their parents expectedShould show implied beliefs about absolute income growthover lifetime, not cross-section
Case:
Pension system phase-outHealth care system privatizedNo kids to take care of youPeople just look to older HHs, don’t think about growth
Song and Yang China’s Saving Rate
Authors’ story: China is Modigliani on Steriods!
Show clockwise rotation in cross-section profile in ChinaCurrent 25-year-olds
Much higher income level than their parents hadSimilar income at 25 to current 55-year-oldsBut expect worse income growth than their parents expectedShould show implied beliefs about absolute income growthover lifetime, not cross-section
Case:
Pension system phase-outHealth care system privatizedNo kids to take care of youPeople just look to older HHs, don’t think about growth
Song and Yang China’s Saving Rate
Authors’ story: China is Modigliani on Steriods!
Show clockwise rotation in cross-section profile in ChinaCurrent 25-year-olds
Much higher income level than their parents hadSimilar income at 25 to current 55-year-oldsBut expect worse income growth than their parents expectedShould show implied beliefs about absolute income growthover lifetime, not cross-section
Case:
Pension system phase-outHealth care system privatizedNo kids to take care of youPeople just look to older HHs, don’t think about growth
Song and Yang China’s Saving Rate
Authors’ story: China is Modigliani on Steriods!
Show clockwise rotation in cross-section profile in ChinaCurrent 25-year-olds
Much higher income level than their parents hadSimilar income at 25 to current 55-year-oldsBut expect worse income growth than their parents expectedShould show implied beliefs about absolute income growthover lifetime, not cross-section
Case:
Pension system phase-outHealth care system privatizedNo kids to take care of youPeople just look to older HHs, don’t think about growth
Song and Yang China’s Saving Rate
Authors’ story: China is Modigliani on Steriods!
Show clockwise rotation in cross-section profile in ChinaCurrent 25-year-olds
Much higher income level than their parents hadSimilar income at 25 to current 55-year-oldsBut expect worse income growth than their parents expectedShould show implied beliefs about absolute income growthover lifetime, not cross-section
Case:
Pension system phase-outHealth care system privatizedNo kids to take care of youPeople just look to older HHs, don’t think about growth
Song and Yang China’s Saving Rate
Authors’ story: China is Modigliani on Steriods!
Show clockwise rotation in cross-section profile in ChinaCurrent 25-year-olds
Much higher income level than their parents hadSimilar income at 25 to current 55-year-oldsBut expect worse income growth than their parents expectedShould show implied beliefs about absolute income growthover lifetime, not cross-section
Case:
Pension system phase-outHealth care system privatizedNo kids to take care of youPeople just look to older HHs, don’t think about growth
Song and Yang China’s Saving Rate
Authors’ story: China is Modigliani on Steriods!
Show clockwise rotation in cross-section profile in ChinaCurrent 25-year-olds
Much higher income level than their parents hadSimilar income at 25 to current 55-year-oldsBut expect worse income growth than their parents expectedShould show implied beliefs about absolute income growthover lifetime, not cross-section
Case:
Pension system phase-outHealth care system privatizedNo kids to take care of youPeople just look to older HHs, don’t think about growth
Song and Yang China’s Saving Rate
Assessment
I don’t buy it
Unexplained: Why does this always happen?
Japan, Korea, Hong Kong, Taiwan, Singapore, ... now ChinaAge/income profile always rotates? Not likely ...
Leaves high absolute level of Chinese saving rates unexplained
Song and Yang China’s Saving Rate
Assessment
I don’t buy it
Unexplained: Why does this always happen?
Japan, Korea, Hong Kong, Taiwan, Singapore, ... now ChinaAge/income profile always rotates? Not likely ...
Leaves high absolute level of Chinese saving rates unexplained
Song and Yang China’s Saving Rate
Assessment
I don’t buy it
Unexplained: Why does this always happen?
Japan, Korea, Hong Kong, Taiwan, Singapore, ... now ChinaAge/income profile always rotates? Not likely ...
Leaves high absolute level of Chinese saving rates unexplained
Song and Yang China’s Saving Rate
Assessment
I don’t buy it
Unexplained: Why does this always happen?
Japan, Korea, Hong Kong, Taiwan, Singapore, ... now ChinaAge/income profile always rotates? Not likely ...
Leaves high absolute level of Chinese saving rates unexplained
Song and Yang China’s Saving Rate
Assessment
I don’t buy it
Unexplained: Why does this always happen?
Japan, Korea, Hong Kong, Taiwan, Singapore, ... now ChinaAge/income profile always rotates? Not likely ...
Leaves high absolute level of Chinese saving rates unexplained
Song and Yang China’s Saving Rate
Patrick Henry, Thomas Sargent, and Economic Man
“I know no way of predicting the future but by the past” – PatrickHenrySargent: Model expectations as coming from regression analysis ofpast dataLaibson et al (current JEP): “Natural Expectations”Authors: “Myopic expectations” – need to go further
Song and Yang China’s Saving Rate
Patrick Henry, Thomas Sargent, and Economic Man
“I know no way of predicting the future but by the past” – PatrickHenrySargent: Model expectations as coming from regression analysis ofpast dataLaibson et al (current JEP): “Natural Expectations”Authors: “Myopic expectations” – need to go further
Song and Yang China’s Saving Rate
Patrick Henry, Thomas Sargent, and Economic Man
“I know no way of predicting the future but by the past” – PatrickHenrySargent: Model expectations as coming from regression analysis ofpast dataLaibson et al (current JEP): “Natural Expectations”Authors: “Myopic expectations” – need to go further
Song and Yang China’s Saving Rate
Patrick Henry, Thomas Sargent, and Economic Man
“I know no way of predicting the future but by the past” – PatrickHenrySargent: Model expectations as coming from regression analysis ofpast dataLaibson et al (current JEP): “Natural Expectations”Authors: “Myopic expectations” – need to go further
Song and Yang China’s Saving Rate
Patrick Henry, Thomas Sargent, and Economic Man
“I know no way of predicting the future but by the past” – PatrickHenrySargent: Model expectations as coming from regression analysis ofpast dataLaibson et al (current JEP): “Natural Expectations”Authors: “Myopic expectations” – need to go further
Song and Yang China’s Saving Rate
Precautionary Saving?
Carroll and Jeanne (2009) Chamon, Prasad, and Liu (2010):
Big increase in precautionary saving
Effective phasing out of pensionsEffective privatization of health careUncertain political/economic future
Even so:
ρ ≈ 5
β ≈ 1
Song and Yang China’s Saving Rate
Precautionary Saving?
Carroll and Jeanne (2009) Chamon, Prasad, and Liu (2010):
Big increase in precautionary saving
Effective phasing out of pensionsEffective privatization of health careUncertain political/economic future
Even so:
ρ ≈ 5
β ≈ 1
Song and Yang China’s Saving Rate
Precautionary Saving?
Carroll and Jeanne (2009) Chamon, Prasad, and Liu (2010):
Big increase in precautionary saving
Effective phasing out of pensionsEffective privatization of health careUncertain political/economic future
Even so:
ρ ≈ 5
β ≈ 1
Song and Yang China’s Saving Rate
Precautionary Saving?
Carroll and Jeanne (2009) Chamon, Prasad, and Liu (2010):
Big increase in precautionary saving
Effective phasing out of pensionsEffective privatization of health careUncertain political/economic future
Even so:
ρ ≈ 5
β ≈ 1
Song and Yang China’s Saving Rate
Precautionary Saving?
Carroll and Jeanne (2009) Chamon, Prasad, and Liu (2010):
Big increase in precautionary saving
Effective phasing out of pensionsEffective privatization of health careUncertain political/economic future
Even so:
ρ ≈ 5
β ≈ 1
Song and Yang China’s Saving Rate
Precautionary Saving?
Carroll and Jeanne (2009) Chamon, Prasad, and Liu (2010):
Big increase in precautionary saving
Effective phasing out of pensionsEffective privatization of health careUncertain political/economic future
Even so:
ρ ≈ 5
β ≈ 1
Song and Yang China’s Saving Rate
Precautionary Saving?
Carroll and Jeanne (2009) Chamon, Prasad, and Liu (2010):
Big increase in precautionary saving
Effective phasing out of pensionsEffective privatization of health careUncertain political/economic future
Even so:
ρ ≈ 5
β ≈ 1
Song and Yang China’s Saving Rate
Precautionary Saving?
Carroll and Jeanne (2009) Chamon, Prasad, and Liu (2010):
Big increase in precautionary saving
Effective phasing out of pensionsEffective privatization of health careUncertain political/economic future
Even so:
ρ ≈ 5
β ≈ 1
Song and Yang China’s Saving Rate
Precautionary Saving?
Carroll and Jeanne (2009) Chamon, Prasad, and Liu (2010):
Big increase in precautionary saving
Effective phasing out of pensionsEffective privatization of health careUncertain political/economic future
Even so:
ρ ≈ 5
β ≈ 1
Song and Yang China’s Saving Rate
Habits (Carroll, Overland, and Weil (2000))
Habits + Precautionary
Political economy (?)
Corporate governance (?)
...
Song and Yang China’s Saving Rate
Habits (Carroll, Overland, and Weil (2000))
Habits + Precautionary
Political economy (?)
Corporate governance (?)
...
Song and Yang China’s Saving Rate
Habits (Carroll, Overland, and Weil (2000))
Habits + Precautionary
Political economy (?)
Corporate governance (?)
...
Song and Yang China’s Saving Rate
Habits (Carroll, Overland, and Weil (2000))
Habits + Precautionary
Political economy (?)
Corporate governance (?)
...
Song and Yang China’s Saving Rate
Habits (Carroll, Overland, and Weil (2000))
Habits + Precautionary
Political economy (?)
Corporate governance (?)
...
Song and Yang China’s Saving Rate
A Big Question
Very valuable new data
Much work remains to be done
Song and Yang China’s Saving Rate
A Big Question
Very valuable new data
Much work remains to be done
Song and Yang China’s Saving Rate