Factors behind rising U.S. wealth inequality: a decomposition and policy implications
0 likes34 views
The document discusses the relationship between rising longevity and wealth inequality in the U.S., highlighting significant increases in life expectancy from retirement that impact wealth distribution. It presents a new method for decomposing wealth inequality into components related to saving behavior, population shares, and generational exchange, using data from the Survey of Consumer Finances spanning from 1949 to 2016. The findings suggest that while saving behavior tends to reduce inequality, changes in demographics significantly contribute to increased wealth inequality linked to longevity.
![Rising longevity and US wealth inequality
Marcin Lewandowski [FAME|GRAPE & University of Warsaw]
Krzysztof Makarski [FAME|GRAPE & Warsaw School of Economics]
Jan Svejnar [Columbia University & CERGE-EI]
Joanna Tyrowicz [FAME|GRAPE, University of Warsaw & IZA ]
13th Annual Lithuanian Conference on Economic Research
Vilnius, December 2024
1 / 30](https://image.slidesharecdn.com/presentation-241223131019-89b481ca/85/Factors-behind-rising-U-S-wealth-inequality-a-decomposition-and-policy-implications-1-320.jpg)
More Related Content
Similar to Factors behind rising U.S. wealth inequality: a decomposition and policy implications (20)
Factors behind rising U.S. wealth inequality: a decomposition and policy implications
- 1. Rising longevity and US wealth inequality Marcin Lewandowski [FAME|GRAPE & University of Warsaw] Krzysztof Makarski [FAME|GRAPE & Warsaw School of Economics] Jan Svejnar [Columbia University & CERGE-EI] Joanna Tyrowicz [FAME|GRAPE, University of Warsaw & IZA ] 13th Annual Lithuanian Conference on Economic Research Vilnius, December 2024 1 / 30
- 2. Rising longevity and wealth inequality in the US Life expectancy at retirement increaesd substantially Up 5+ years: 14.1 years (1950) → 19.5 years (2015) Theory: longevity affects wealth distribution 1. BEHAVIORAL: ↑ higher peak old-age saving → widening gap between young self and retiring self 2. STRUCTURAL: ↑ cohorts close to retirement relatively more numerous Question: does rising longevity matter quantitatively for wealth inequality? 2 / 30
- 3. Two dimensions of inequality: same birth cohort vs different birth cohorts 3 / 30
- 4. Two dimensions of inequality: same birth cohort vs different birth cohorts 4 / 30
- 5. Two dimensions of inequality: same birth cohort vs different birth cohorts 5 / 30
- 6. Our contribution • New decomposition of (changes in) wealth inequality • Saving behavior • Population shares • Generational exchange • Apply it to the SCF data for United States, spanning 1949-2016 • magnitudes of the three factors very large (compared to ∆ in wealth inequality) • demographics trump saving behavior • saving behavior reduces inequality • We link changes in wealth inequality to a rise in longevity • identify which cohorts contribute the most to wealth inequality 6 / 30
- 7. Plan of this talk Methods Take this method to the data Results Conclusions 7 / 30
- 9. Decomposing wealth inequality Total Inequality = Inequality Between Cohorts + Inequality Within Cohorts Gini can’t do this! but Generalized Entropy can GE(ϵ) = 1 Nϵ(ϵ − 1) N X i h ai ā ϵ − 1 i GE(ϵ) = GEwithin(ϵ) + GEbetween(ϵ) = 1 ϵ(ϵ − 1) c X i=1 h nc ai āc ϵ − 1 i + c X i=1 h nc · GEc (ϵ) · āc ā ϵi ∆GE(ϵ) = ∆GEwithin(ϵ) + ∆GEbetween(ϵ) 9 / 30
- 10. Factors of changes in within-cohort inequality ∆GEwithin(ϵ) = ∆population shares: = C X c=1 (n(t2,c) − n(t1,c)) · GEw (ϵ)(t1,c) · ā(t2,c) āt2 ϵ ∆saving behavior: = C X c=1 n(t1,c) · GEw (ϵ)(t1,c) · ā(t2,c) āt2 ϵ − ā(t1,c) āt1 ϵ ≎ generations = B X b=1 n(t2,b) · GEw (ϵ)(t2,c) · ā(t2,b) āt2 ϵ − D X d=1 n(t1,d) · GEw (ϵ)(t1,d) · ā(t1,d) āt1 ϵ ∆GEw = C X c=1 n(t1,c) · ā(t2,c) āt2 ϵ · GEw (ϵ)(t2,c) − GEw (ϵ)(t1,c) 10 / 30
- 11. Factors of changes in between-cohort inequality ∆GEbetween(ϵ) = ∆ population shares = 1 ϵ2 − ϵ C X c=1 (n(t2,c) − n(t1,c)) · ā(t2,c) āt2 ϵ ∆ saving behavior = 1 ϵ2 − ϵ C X c=1 n(t1,c) · ā(t2,c) āt2 ϵ − ā(t1,c) āt1 ϵ ≎ generations = 1 ϵ2 − ϵ B X b=1 n(t1,b) · ā(t2,b) āt2 ϵ − D X d=1 n(t1,d) · ā(t1,d) āt1 ϵ 11 / 30
- 12. Decomposing wealth inequality We formulate three propositions, for the role of 1. saving behavior 2. population structure 3. exchange of generations in the change of wealth inequality. 12 / 30
- 13. Decomposing wealth inequality Proposition 3: Generational exchange The change in wealth inequality brought about by direct effects of a change in the population structure and measured by the Generalized Entropy index between periods t1 and t2, (∆GE(ϵ)), can be decomposed into: • a between-cohort effect of the generational exchange (inequality contributed by entering cohorts, indexed by e and by deceased cohorts, indexed by d) 1 ϵ2 − ϵ B X b=1 nt2,b · at2,b at2 ϵ − D X d=1 nt1,d · at1,d at1 ϵ # • and a within-cohort effect of the change in population shares of surviving cohorts B X b=1 nt2,b · GEw (ϵ)t2,c · at2,b at2 ϵ − D X d=1 nt1,d · GEw (ϵ)t1,c · at1,d at1 ϵ 13 / 30
- 14. Decomposing wealth inequality Proposition 2: Population shares The change in wealth inequality brought about by direct effects of a change in the population structure and measured by the Generalized Entropy index between periods t1 and t2, (∆GE(ϵ)), can be decomposed into: • a between-cohort change in population shares of surviving cohorts 1 ϵ2 − ϵ C X c=1 (n(t2,c) − n(t1,c)) · a(t2,c) at2 ϵ • and a within-cohort effect of the change in population shares of surviving cohorts C X c=1 (n(t2,c) − n(t1,c)) · GEw (ϵ)(t1,c) · ā(t2,c) āt2 ϵ 14 / 30
- 15. Decomposing wealth inequality Proposition 1: Saving behavior The change in wealth inequality brought about by a change in individuals’ saving behavior and measured by the Generalized Entropy index between periods t1 and t2, (∆GE(ϵ)), can be decomposed into: • a between-cohort change in average cohort wealth (saving behavior) of surviving cohorts 1 ϵ2 − ϵ C X c=1 n(t1,c) · a(t2,c) at2 ϵ − a(t1,c) at1 ϵ • a within-cohort change in average cohort wealth (saving behavior) of surviving cohorts C X c=1 n(t1,c) · GEw (ϵ)(t1,c) · ā(t2,c) āt2 ϵ − ā(t1,c) āt1 ϵ • and a residual change in within-cohort GE index (saving behavior) of the surviving cohorts C X c=1 n(t1,c) · GEw (ϵ)(t2,c) − GEw (ϵ)(t1,c) · ā(t2,c) āt2 ϵ 15 / 30
- 16. Decomposing wealth inequality – a summary Three propositions yield seven components: 1. within-cohort related to saving behavior ✓for infinitely lived agents 2. within-cohort implied by changing population shares requires OLG 3. within-cohort implied by changing exchange of generations requires OLG+ 4. within-cohort resulted from overall changes in wealth distribution (first-order approximation, further decomposable) perhaps ✓for infinitely lived agents? 5. between-cohort driven by population structure requires OLG+ 6. between-cohort contribution from exchange of generations requires OLG+ 7. between-cohort contribution from saving behavior requires OLG+ 16 / 30
- 17. Take this method to the data 17 / 30
- 18. Data source Survey of Consumer Finances (SCF+) data for 1950-2020 by Kuhn et al (2020) • Assets - Financial assets (including defined-contribution retirement plans), real estate, cars • Liabilities - Personal debt and hosusing debt • Our measure - Net wealth Demographic characteristics match Current Population Survey and U.S. Census data 18 / 30
- 19. Choose the ϵ parameter: match Gini 19 / 30
- 20. But this ϵ parameter does well also for other measures: top10 20 / 30
- 21. But this ϵ parameter does well also for other measures: top1 21 / 30
- 22. Results 22 / 30
- 23. Saving behavior reduces inequality, but demographic changes raise it 23 / 30
- 24. This is a consistent pattern, only magnitudes change (driving net changes) 24 / 30
- 25. The full decomposition • Within and between off-set each other, but magnitudes are HUGE • Between- due to changes in population structure LARGE and outweighs within- • Within- saving behavior amplifies inequality, but small and outweighed by between- 25 / 30
- 26. Do cohorts with higher longevity contribute to overall inequality? Deaton Paxson (1994) decomposition Wealthi = βc birth cohortc + βy yeary + βaagea + ϵi Recentered Influence Functions Firpo, Fortin Lemieux (2009) RIF{wealthi , GE(ϵ)} = βc birth cohortc + βaagea + βy yeary + ϵi (we also do it for Gini) 26 / 30
- 27. Yes! 27 / 30
- 29. Demography and life-cycle matters • Demographics is key in changes of wealth inequality 1. Model mis-specification 2. Adequate targets to match 3. Rise in wealth inequality linked to increased longevity • Saving behavior reduced inequality in the US (in line with the predictions of OLG model) • We offer a new method for studying dynamics of wealth inequality 29 / 30
- 30. Thank you for your attention! Questions or suggestions? w: grape.org.pl t: grape org f: grape.org e: [email protected] 30 / 30