Upload
ivor-french
View
45
Download
0
Embed Size (px)
DESCRIPTION
Discussion of Changing Progressivity as a Means of Risk Protection in Investment-Based Social Security. Michael Hurd RAND and NBER. Can risk of low rates of return from Private Retirement Accounts be partially offset by increased progressivity in (reduced) Social Security program?. - PowerPoint PPT Presentation
Citation preview
Discussion of
Changing Progressivity as a Means of Risk Protection in
Investment-Based Social Security
Michael Hurd
RAND and NBER
Can risk of low rates of return from Private Retirement Accounts be
partially offset by increased progressivity in (reduced) Social
Security program?
Interesting alternative to pure insurance against bad outcomes on rates of return
Done through simulation
Here are the steps
Simulation of lifetime earnings
• Cohort born in 1973 age 30 in 2003
• Take distribution of earnings from cross-section estimate of log normal
• First order Markov with high persistence
Age 30 Age 67
Forecast and back-cast earnings from age 30 in 20003
Low earning person in 2003
Range of wage outcomes
Age 30 Age 67
Forecast and back-cast earnings from age 30 in 20003
Age 30 Age 67
Forecast and back-cast earnings from age 30 in 20003
At each age
• save 2% of earnings
• into PRA in bond-stock portfolio
• augment PRA by random draw from historical rates of return
Calculate Social Security benefits for each path based on current law
Repeat many times
Distribution of Social Security Benefits
Scheduled: E(B) = $21.8k
But current law not sustainable, so reduce benefits by 40%
Reduced Distribution of Social Security Benefits to close actuarial gap
Reduced by 40%: E(B) = $13.1k
Scheduled: E(B) = $21.8k
Redistribute to increase progressivity…example
E(B) = $13.1k
E(B) = $13.1k
Topped up at, say, 25th percentile and then reduced
Then add in private investment account: Distribution of Social Security Benefits + PRA
E(B) = $13.1k 2% annual earnings invested in ½ stocks and ½ bonds, annuitized plus Soc. Sec.
Calculate E(U)
Repeat for other investment programs
2% annual earnings invested in ½ stocks and ½ bonds, annuitized plus Soc. Sec.
2% annual earnings invested in bonds, annuitized plus Soc. Sec.
Calculate expected utility for these and variants
Compare utilities (certainty equivalents) for variation in
• Risk aversion
• Rates of return on equities
• Fraction of earnings in Private Retirement Accounts
My main question:Whose utility?
Thought experiment:
• Draw a worker from population.
• Calculate wage path
• Calculate saving path with stochastic rates of return
• Calculate utility
• Repeat and average
Whose utility
Stochastic elements: wage level w, wage growth (u), rates of return (v)
What is calculated:
, , , |( ) (( ) | )w u v u v w wU c dP U c w dP dP
Thus average (over workers) of individual expected utility
• Not utility of any individual• Is this quantity a desirable social
objective?
, , , |( ) (( ) | )w u v u v w wU c dP U c w dP dP
Alternative…utility of individuals
• Begin with a worker (w)
• Calculate wage and saving path
• Utility for that worker
• Replicates for that worker
• E(U)
Utility of individuals
• Repeat for many workers
• Study distribution of E(U)– E.g. How many workers have their E(U)
improved and how many reduced under each alternative
– Effects on workers in lower part of wage distribution compared with workers in upper part.
Additional advantage
• Workers indexed by initial wage
• Use more realistic life-cycle wage paths
Real earnings of cohort born in 1940-45
0
200
400
600
800
1000
1200
1400
25-29 35-39 45-49
Age
16+ years education
9-12 years education
Question about utility calculation
• Consumption in retirement equals Social Security plus annuitized Private Retirement Account
• But most have other resources• Variation in marginal utility from earnings
uncertainty and investment returns will vary with other resources
• Other resources positively correlated with initial wage and therefore position in Social Security benefit distribution
Conclusion
• Package of increasing progressivity in Social Security along with PRAs can lead to higher resources along with some protection for low wage or unlucky workers.
• Moves in direction of some European public pension systems– Politically feasible somewhere in world
• Excellent contribution to debate
Conclusions (cont.)
But reservations about utility calculation
• Ex ante random worker– Interest in distribution of E(U)
• No accounting for differing rates of growth in earnings
• No accounting for other resources