Informing Aging Policy:
Microeconomic Panel Data
on the Older Population
©Olivia
S. Mitchell
The Pension Research Council
mitchelo@wharton.upenn.edu
The US Health and Retirement Study
(HRS http://hrsonline.isr.umich.edu/ )
• Bi-annual survey of age 50+ from 1992 (to ??)
– N~22,000
– 90-minute interview each for husband and wife
– Follow into nursing home, proxy interview if demented
or if dead
• High quality data essential for policy analysis:
–
–
–
–
–
Economic/labor market status and history;
Health status/history, healthcare utilization, costs;
Family transfers (time and money);
Psychological/expectations;
Links to government benefit programs
ÆHuge Policy Impact:
• Social Security Commission
(www.csss.gov): levels/distribution of financial
capital, social capital, health capital
• Medical and old-age systems: analysis
of health policy reforms (health and LTC)
• Congress: saving/dissaving, asset
allocation, credit behavior
• Researchers: life cycle focus (vs cross
section)
HRS Longitudinal Sample Design
AGE
90
AHEAD
85
CODA
80
75
HRS
70
War
Babies
65
60
Early
Boomers
55
Mid
Boomers
50
92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10
YEAR
Pre-retirement wealth varies substantially
Health and Retirement Study (Michigan)
200%
Assets in Fifths
150%
100%
50%
0%
1
2
3 4
5
6
7
8 9 10 11 12 13 14 15 16 17 18 19 20
-50%
Wealth Ventile
-100%
US: Moore/Mitchell 00
Net Housing Wealth
Net Financial Wealth
Social Security
Pension Wealth
Many face large saving shortfalls:
Nonmarried
Earnings Decile
10
8
6
4
2
0.4
Earnings Decile
10
8
6
Prescribed Savings Rate
4
0.2
2
0
0.4
2
4
6
Prescribed Savings Rate
8
Wealth Decile
0.2
10
0
To retire at 62 & smooth consumption:
Æmust save 15-20% more (half at 65)
High earners also have largest shortfalls
Mitchell/Moore
2
4
6
8
Wealth Decile
Married
10
Detailed survey on 50+ population
essential for Japan as most-rapidly aging
nation
ÆExisting datasets unsuited:
– Not public
– Not panel
– Small N
– Focus on household head
– Lack coherent financial, health, and social
capital picture (including family transfers)
– Lack consumption & saving information
– Must link tax and benefit (e.g. pension, LTC)
information
Considerations:
• Survey design:
– Ask questions similar to HRS, SHARE
– Don’t focus only on household head!
– Response must be ~80% for international credibiltiy;
consider incentives?
• Funding:
– Need large enough N to draw conclusions
– Cost rises with sample but so does reliability
• Confidentiality:
– Data can be masked and made available to
researchers
• Timing:
– Start now, since building a panel takes time
HRS: hypothetical vs actual lifetime pay
Distribution of Average Lifetime Earnings
100000
90000
Low Scaled
($8726)
80000
Medium Scaled
($19,395)
High Scaled
($31,039)
Average Lifetime Earning
70000
60000
50000
40000
30000
20000
10000
0
1
5
9
13
17
21
25
29
33
37
41
45
49
53
Percentile
57
61
65
69
73
77
81
85
89
93
97
101
Importance cannot be overestimated:
• Global intellectual impact:
– 000s of published papers and books
• Similar surveys underway in:
– Chile, Australia, Korea, Mexico, U.K., rest of
Europe,
– And now…Japan!
Thank you!
For more information:
• Wharton’s Pension Research
Council:
http://prc.wharton.upenn.edu/prc/prc.html
• Books and working papers:
http://rider.wharton.upenn.edu/~prc/publication.html