The Relationship Between Altruism and Equal Division Evidence

The Relationship Between Altruism and Equal Division: Evidence From Inter Vivos Transfer Behavior Elin Halvorsen elin. halvorsen@ssb. no Thor O. Thoresen The opinions presented here are those of the discussant alone and may not represent the official position of the Internal Revenue Service or the U. S. Treasury Department.

Presentation n n n Introduction and background Norwegian wealth transmission laws and inheritance tax system The model The data Results Additional comments

Motivations for Intergenerational Transfers n n Accidental Intentional ¨ Altruistic ¨ Exchange or strategic ¨ Other n Equal sharing ¨ Societal norm ¨ Indicator of parental affection

Norwegian Wealth Transmission Laws n Children are guaranteed 2/3 of parents’ estate for amounts of up to 1 million kroner (kr) ¨ The 2/3 portion must be divided evenly among children n Remaining 1/3 may be distributed according to preferences ¨ Children, other family members ¨ Charities n Only 25 percent of Norwegians transfer assets through a will

Inter Vivos Transfer Restrictions Gifts deemed “advance bequests” are subject to transmission laws and inheritance tax n Gifts not considered “advance bequests” n ¨ Relatively small amounts ¨ Not subject to equal sharing ¨ Tax-exempt

Inheritance Tax Structure Applies to cumulative gifts and inheritance n Special valuation rules to facilitate transfer of family businesses n Tax rates are progressive: n <200, 000<500, 000 ≥ 500, 000 Inter-spousal 0% 0% 0% Child/Parent 0% 8% 20%

The Model n n Authors seek to test the influence of an “Equal Sharing Norm” on the altruistic model. Two testable hypotheses ¨ The degree of income compensation should be stronger in one-child families ¨ The altruism motive should dominate when income differences among children are large

The Model n Parent consumption: cp = ep − ∑bi i=1 Child consumption: ci = ei + bi i = 1, …, n n n Parent Utility: U = ln cp + α ∑ ln ci − ∑ γ/2 (bi – b)2 i=1 n n n Max: U = ln (ep − ∑ bi ) + α ∑ ln (ei + bi ) − ∑ γ/2 (bi – b)2 i=1 i=1 Subject to: bi > 0

Implications of model n In absence of fairness attitude (γ=0), parents balance gifts to equalize children’s consumption, taking income as given: b 2 – b 1 = e 1 – e 2 n If parent’s derive utility from equal division (γ>0) then: b 2 – b 1 < e 1 – e 2 n The larger the fairness parameter, γ, compared to altruism, α, the less parents compensate income differences between children n Transfer-income derivative: δbi – δbi = 1

The Data n n n Use data from a 2001 survey conducted by Norwegian Social Research (NOVA) 1, 877 households where HH was interviewed Wealth, income, employment for household and constituent members Educational attainment, economic situation, etc. , for household member’s parents, in-laws, all grown children Transfers given and received within previous 12 months

Sub-sample : HH with grown children n Respondent characteristics (donor role) ¨ ¨ ¨ ¨ n Sample size: 2, 021 parent/child pairs Mean age respondent: 61 (40 - 93) Mean number of children: 3 Mean household income: 347, 000 kr (0 – 1, 200, 000) Mean household net worth: 1, 729, 000 kr (-245, 000 – 11, 400, 000) 23% made gifts Mean gift: 30, 000 kr (80 -85% < 40, 000 kr) Child characteristics ¨ ¨ ¨ Mean age: 43 (18 - 73) 49% female, 51% male 63% married/cohabitant; 56% had children 33% college/university degree 18% unemployed; 15% students Economic situation: 6% bad, 43% well

Sub-sample: HH with living parents n Respondent characteristics (donee role) ¨ Sample size: 1, 263 Mean age: 38 (18 -74) 70% married/cohabitant; 56% had children 16% unemployed; 13% students Economic situation: 10% bad, 54% well Mean household income: 421, 000 kr (0 – 6, 600, 000) Mean household net worth: 1, 290, 000 kr (-740, 000 – 13, 020, 000) 19% received gifts ¨ Mean 22, 000 kr (80 -85% gifts < 40, 000 kr) ¨ ¨ ¨ ¨ n Parent characteristics Mean age parent: 65 (35 – 105) ¨ Mean number of children: 3. 3 ¨ 51% married ¨

Parent attitudes toward inter vivos transfers Q. What kind of economic obligation do you think parents should have towards their grown children? All parents Donor parents Only in emergencies. 68. 64 Education, house and family formation. 26. 33 Same living standard as parents. 04. 02 Q. When parents who have more than one child give economic support, how do you think they should divide the resources? Equal sharing According to need To the most helpful child All parents Donor parents. 73. 67. 23. 29. 01

Econometric Analysis n Econometric Model b*ij = η 1 epj + η 2 ei +X β + uij bij = b*ij if b*ij > 0 0 if b*ij ≤ 0 i = 1 …, N; j = 1, …, P Expect η 1 – η 2 = 1 Sample observations with no gifts include both future donors/recipients and those who will never make gifts Do not have income and wealth for child and parent together, so use both samples Use proxies for permanent and transitory income { n n

Tobit Findings n Determinants of inter vivos gifts ¨ Significant negative effect of recipient income on amount of inter vivos gifts ¨ Transfers decline as parents age ¨ Transfers smaller and more frequent when children younger ¨ Size of transfers decrease as number of children increase n Transfer-income derivative findings Transfer derivative higher for only-children = -. 27 ¨ Derivative for children with siblings = -. 03 ¨

Fitted Spline for Inter Vivos Gifts by Recipient Household Income Flattens above child income mean Household income (in thousands)

Income and Transfer Differences Among Siblings n Parent is unit of observation n Model: (bij – bj ) = γ Zpj + βk (xkij – xkj ) + εij One child is the reference point Includes variables indicating contact with, and assistance provided to, parent Results show only relevant variables are child’s relative economic situation and employment status Unequal sharing declines with parent age and income, but increases with number of children n n

Author Conclusions n n n Parents are faced with trade-off between compensating income differences among children and being fair Recipient income derivative higher for one-child families Recipient income derivative nonlinear for multiple-child families ¨ Evidence suggests compensation when recipient income is below the median n Comparing sibling characteristics suggests relative economic situation is main determinant of unequal division of gifts

Final Thoughts n n Use survey data on propensity to give? Does wealth taxation play a role? Reduce taxable inheritance ¨ Transfer anticipated future gains ¨ n n Effects of uncertainty? (Kotlikoff, et al. , 2003) Data challenges ¨ Time frame -12 months is short ¨ Estimate of accumulated/expected transfers; panel? ¨ Definition of gifts very broad n n Parents’ ability to make gifts Consider interplay of motivations