Sang-Wook (Stanley) Cho

Transcription

1 Beggar-thy-parents? A Lifecycle Model of Intergenerational Altruism Sang-Wook (Stanley) Cho University of New South Wales March 2009

2 Motivation & Question Since Becker (1974), several studies analyzing the size and the effect of parental transfers to children, both intervivos transfers to younger children, and bequests. Borrowing constraints may induce the parents to transfer resources to constrained children (who tend to be younger) rather than leaving it for later bequests.

3 Motivation Since Becker (1974), several studies analyzing the size and the effect of parental transfers to children, both intervivos transfers to younger children, and bequests. Borrowing constraints may induce the parents to transfer resources to constrained children (who tend to be younger) rather than leaving it for later bequests. The topic is important and deserves our attention.

4 Motivation Since Becker (1974), several studies analyzing the size and the effect of parental transfers to children, both intervivos transfers to younger children, and bequests. Borrowing constraints may induce the parents to transfer resources to constrained children (who tend to be younger) rather than leaving it for later bequests. The topic is important and deserves our attention.... by an anonymous referee who rejected the paper

5 Question We analyse intergenerational transfer of wealth over the lifecycle in the US vs. Korea using a standard lifecycle model with different transfer motives Explore the aggregate, distributional, and welfare impact of borrowing constraint The role of borrowing constraint in accounting for cross-country differences in wealth accumulation & timing of transfer Document counter-factual policy experiments in the light of borrowing constraint

6 Outline Some empirical evidence (rough data, no estimation) Benchmark lifecycle model with different transfer motives - bequest vs. intervivos transfer Simulation results Policy experiments

7 Age-Wealth Profiles in US vs. Korea US (PSID 2001) Korea (KLIPS ) Average Income = Age Figure 2.1 Age-profile of Wealth Accumulation (US vs Korea)

8 Aggregate Intervivos Transfer Table: Aggregate Intervivos Transfer - US, Korea Author Source Percentage TR TR Y Altonji et al. PSID 23% $ % (1997) (1988) (received) McGarry HRS 29% $ % (1999) (1992) (gave) Kim et al. KLIPS 21% $ % (2004) (2002) (received)

9 Age-Transfer Distribution Table: Age-Transfer Distribution - US, Korea Age Groups Fraction of households receiving transfers US 29.5% 22.6% 19.1% 16.5% 11.8% Korea 26.6% 28.6% 27.3% 19.1% 13.7% Transfer as a fraction of average household income US 4.3% 5.6% 6.5% 6.7% 4.8% Korea 5.2% 14.7% 9.9% 6.2% 9.2% Transfer as a fraction of average cohort income US 5.6% 5.2% 5.2% 5.2% 3.5% Korea 13.3% 14.0% 8.3% 5.5% 7.7%

10 Evidence of Borrowing Constraint? Table: Mortgage-GDP ratio and Loan-to-value ratios - US, Korea Year Outstanding Mortgage-to-GDP Ratio US 53% 52% 54% 56% 55% Korea 10% 11% 12% 12% 13% Loan-to-value ratio US 78% 80% 80% 78% 79% Korea 25% 26% 28% 28% 33%

11 Summarizing... Age-wealth profile shows that 1 wealth peaks earlier and decumulates faster after the peak in Korea than in the United States 2 younger Korean households hold more wealth than the corresponding cohorts in the United States, while for older households, the relation is reversed.

12 Summarizing... Age-transfer profile shows that 1 on average, more households in Korea receive positive parental transfers than in the United States 2 the relative importance of transfer income is larger for Korean households, especially in the younger cohorts aged less than 35. The degree of credit constraint in Korea is larger as evidenced from the size of the mortgage market and the down-payment requirements.

13 Main features Quantitative general equilibrium lifecycle model One-sided altruism (parent children) warm-glow bequest motive during retirement (parents aged vs. children aged 40-55) altruistic intervivos transfer prior to retirement (parents aged vs. children aged 25-40) Uninsurable labor income, uncertain lifetime (mortality risk)

14 Demographics & Technology 1 period in the model corresponds to 5 years. Agents enter working life at 25(j = 1), retire at 65(j = 9), live at most until 80(J = 12). Retired agents face mortality risk, upon death agents leave bequest, equally re-distributed to agents (aged 40-55) s j : exogenously given survival probability at age j + 1 conditional on being alive at age j. Final goods produced can be consumed (C), invested (I ) or spent by government (G).

15 Intergenerational Transfers INTERGENERATIONAL TRANSFERS Parents Work Retire j = 7, 8, 9 j = 10,11,12 Bequest Children Intervivos transfer j = 1, 2, 3 j = 4, 5, 6 Work Work

16 Preferences CRRA utility function U(c j ) = c1 σ j 1 σ Warm-glow bequest motive, De Nardi (2004). ϕ(q) = ϕ 1 [ 1 + q ϕ 2 ] 1 σ Intervivos transfer motivated by altruism parents care about children s utility with discount factor γ < 1.

17 Labor Productivity Exogenous age-efficiency profile ɛ j + stochastic productivity shocks y j (Markov process) First period productivity shock is inherited from parents according to a transition function Q yh. Afterwards, producvitity evolves according to transition function Q y. Total productivity given by ɛ j y j, wage rate given by ω.

18 Market Arrangements Working agents can borrow upto some fraction of current period labor income. Agents in the early stages of lifecycle (j = 1, 2, 3) can also partially collaterialize parental assets. a j+1 κ 1 ωɛ j y j I j=1,2,3 κ 2 a 1 j+6 j = 1,..., 9 Retired households cannot borrow. a j+1 0 j = 10,..., 12

19 Government Re-distribution of bequest to children in periods j = 4, 5, 6. j=10,11,12 T = q m (dx) (1) m (dx) j=4,5,6 Social security benefit b m (dx) = τ ss j=10,11,12 j=1,...,9 Constant government expenditure from tax revenues. ωɛy m (dx) (2) τ c C + τ k rk = G (3)

20 Household Recursive Problem : Retired Parents (j = 10, 11, 12) V (j, a) = max c,a [ U(c) + sβv (j + 1, a ) + (1 s)ϕ(q) ] (4) subject to (1 + τ c )c + a (1 + r(1 τ k ))a + b c 0 q = a 0

21 Household Recursive Problem : Children ( j = 4, 5, 6) V ( j, ã, ỹ) = max c,ã [ U( c) + βev ( j + 1, ã, ỹ ) ] (5) subject to (1 + τ c ) c + ã (1 τ ss )ωɛỹ + (1 + r(1 τ k ))ã + T c 0 ã κ 1 ωɛỹ

22 Household Recursive Problem : Parents at (j = 7, 8, 9) V (j, a, y, ã, ỹ) = max c,a,tr U(c) + γu( c) + I j=7,8βev (j + 1, a, y, ã, ỹ ) +I j=9 βv (j + 1, a ) subject to (6) (1 + τ c )c + tr + a (1 τ ss )ωɛy + (1 + r(1 τ k ))a c 0 a κ 1 ωɛy

23 Household Recursive Problem : Children at ( j = 1, 2, 3) V ( j, a, y, ã, ỹ) = max c,ã U( c) + I j=1,2 βev ( j + 1, a, y, ã, ỹ ) +I j=3 βev ( j + 1, ã, ỹ ) subject to (7) (1 + τ c ) c + ã (1 τ ss )ωɛỹ + (1 + r(1 τ k ))ã + tr c 0 ã κ 1 ωɛỹ κ 2 a

24 Stationary Equilibrium government policy arrangements {T, G, τ SS, τ c, τ k } prices {r, w} value functions V (x) allotions c(x), a (x), tr(x) time-invariant distribution of agents, m (x), over the state variables x = {j, a, y, ã, ỹ} aggregate quantities {Y, C, K, L}

25 Stationary Equilibrium the functions V (x), c(x), a (x), tr(x) solve the dynamic maximization problem of the households. factor prices are equal to their marginal products: r = F K (K, L) δ (8) w = F L (K, L) (9) the government policy satisfies (1), (2), (3). m is the invariant distribution of households over the state variables for this economy. all markets clear.

26 Common Parameter Definition and Values Table: Common Parameter Definition and Values Parameters Definition Values σ Risk-aversion coefficient 1.5 α Capital income share δ Depreciation rate φ 1 Bequest parameter -9.5 φ 2 Bequest parameter 11.6 b Replacement ratio 0.4 τ c Consumption tax rate 0.05 ρ Persistence of income process 0.85

27 Country-specific Parameter Definition and Values Table: Country-Specific Parameter Definition and Values Parameters US Korea κ κ σy τ ss 9.0% 12.4% τ k 50.2% 13.0%

28 Parameters to match given set of targets Table: Parameters to match given set of targets Parameters Definition US Korea β Discount factor γ Pure altruism parameter

29 US Table: Summary Statistics of Benchmark Simulation - US Variable Benchmark Data Capital output ratio Transfer-to-bequest ratio Average Wealth

30 US Table: Summary Statistics of Benchmark Simulation - US Variable Benchmark Data Average flow of bequest (% of GDP) 1.6% 2.6% Transfer to (% of GDP) 0.54% 4.3% Transfer to (% of GDP) 1.94% 5.6% Transfer to (% of GDP) 2.66% 6.5% Gini (wealth)

31 Korea Table: Summary Statistics of Benchmark Simulation - Korea Variable Benchmark Data Capital output ratio Transfer-to-bequest ratio Average Wealth

32 Korea Table: Summary Statistics of Benchmark Simulation - Korea Variable Benchmark Data Average flow of bequest (% of GDP) 1.2% 1.9% Transfer to (% of GDP) 1.97% 5.2% Transfer to (% of GDP) 3.17% 14.7% Transfer to (% of GDP) 3.70% 9.9% Gini (wealth)

33 Table: Tightening Borrowing Constraint - US Parameter (κ 1 = 0.75) (κ 1 = 0.25) Capital output ratio Transfer-to-bequest ratio Average Wealth T tr tr tr Welfare Gini

34 Table: Relaxing Borrowing Constraint - Korea Parameter (κ 1 = 0.25) (κ 1 = 0.75) Capital output ratio Transfer-to-bequest ratio Average Wealth T tr tr tr Welfare Gini

35 Table: Parental Asset Collateralization - US Parameter (κ 2 = 0) (κ 2 = 0.25) Capital output ratio Transfer-to-bequest ratio Average Wealth T tr tr tr Welfare Gini

36 Table: Parental Asset Collateralization - Korea Parameter (κ 2 = 0) (κ 2 = 0.25) Capital output ratio Transfer-to-bequest ratio Average Wealth T tr tr tr Welfare Gini

37 Conclusion Life-cycle model with intergenerational linkages to study wealth transfers between parents and children. While the parents are of working age, the utility of the children s consumption enters the utility of the parents. When parents die, they leave bequests, and have an (exogenous) warm-glow utility from leaving a bequest. Higher borrowing constraints (that prevent children from borrowing to smooth their consumption) tend to generate larger transfers early in life (relative to bequests). This matches the empirical evidence from Korea (a stricter constraints country) and the US. We also look how transfers are impacted by allowing children to use parental wealth as a collateral to borrow.

38 Extensions Better calibration of Korean data Incorporate two-sided altruism (children supporting elderly parents) Incorporate human capital investment decision (and its interaction with intervivos transfer)