LWS Final Conference, Roma, Italia, July 5-7 th, 2007 “The Luxembourg Wealth Study: Enhancing Comparative Research on Household Finance
Introduction Motivation Past Empirical Evidence Methods Decomposition
Figure 1: Net Worth ¹ by Martial Status and Gender, Germany 2002 ¹Estimates derived from multiply imputed data together with a 95% confidence interval Source: German SOEP 2002; authors’ calculations.
Simple model of accumulation of assets in period t+1 can be expressed in the following way r - gross rate of return on investments Y t - income in period t C t - consumption in period t. Differences in wealth accumulation could result from: › households differ in the amount they save (Y-C) › households enter the period with different stocks of assets (A) – perhaps due to inheritance. › households receive different rates of return due to differences in portfolio allocation reflecting individual variation in risk preference.
Gender differences in wealth accumulation could result from: › households differ in the amount they save (Y-C) Position in life-cycle (age) Income ( labor market participation (Warren et al 2001); Earnings (for ex. Blau & Kahn 1997, 2000) Differences in precautionary savings due to liquidity constraints › households enter the period with different stocks of assets (A) – perhaps due to inheritance. Differences in probability of owning a home; discrimination in mortgage lending (Ladd 1998); differences by family type (Sedo & Kossoudji 2004) › households receive different rates of return due to differences in portfolio allocation reflecting individual variation in risk preference. Women invest more conservatively (eg. Jiankokopolos & Bernasek 1998) Additionally: › Gender differences in family type e.g. marriage patterns (Zagorsky 1999).
Male TOTAL Male married Male single - never married Female TOTAL Female married Female single - never married % recent inheritance (since 1997) Amt recent inheritance (median, in €)15,400 35,80012,800 % expected inheritance Gender specific Population Share in %
Table 3: Relative Gender Wealth Gap (Men/Women) based on average wealth holdings by Marital Status, Germany 2002 Wealth Component TOTALMarriedCohabitingSingle - divorced/ separated Single - widowed Single - never married Housing Other Property Financial assets Insurance/ Private pensions Business assets Tangible assets Debt Total Shaded cells indicate significant deviation (p<=0,05). Note: Calculations are based on multiply imputed data Source: SOEP 2002.
Focus on married and couple households F>M: Woman's wealth exceeds man's wealth in a couple household F=M: Equal sharing among partners in a couple household F<M: Male-in-charge: Man's wealth exceeds woman's wealth in a couple household NO: No wealth in a couple household Top/bottom code net worth
SOEP Wealth module in 2002 LWS (household level) Individual level (all HH members >16): n= ( HH) Multiple Imputation (5 replicates) Own Property Other Property Financial Assets (>2.500 €)Total Assets Private Pensions Business Assets Tangible Assets (>2.500 €) Net Worth Main Property Debt Other Property DebtTotal Debt Consumer Debt (>2.500 €) Not included: cars, public pension entitlements, durables
Table 4a. Distribution of different type of sharing couple households by wealth quantiles. Wealth quantileSharing Type F>MEqM>F NoTotal Total
Table 4b. Average wealth gap in couple household by different type of sharing couple households and wealth quantiles. Wealth quantile Sharing Type F>MEqM>FTotal Total
Table 4e. Prevalence of the self employed by different type of sharing couple households and wealth quantiles. Wealth quantile Sharing Type F>MEqM>F Total Total
Table 4f. Prevalence of homeowners by different type of sharing couple households and wealth quantiles. Wealth quantile Sharing Type F>MEqM>F Total Total
Decompose the gender wealth gap Mean difference › Into portions attributable to differences in the distribution of endowments and differences in the return to these endowments (wealth function) or way in which the endowments are transformed into wealth (Blinder-Oaxaca) Juhn, Murphy, Pierce (1991) method › Decomposes the component due to differences in coefficients into two parts: unobserved ability and differences in the wealth function (discrimination)
Women Men VariablesCoef. Mean Coef. Mean migback-42523* *0.14 partner-25520* **0.11 loc89east-47858* *0.20 badhlth-12711* *0.18 goodhlth11011* kids ** **0.12 _Iedu_222981* *0.48 _Iedu_335848* *0.14 _Iedu_451965* *0.22 notlabor1846* *11.16 over651262* expft021718* *25.78 exppt022585* expue expmiss74214* *0.01 selfemp71177* *0.08 autonom16203* **0.24 hiedu_f30886* hiedu_m hiedu_p inheri149425* *0.08 inheri246541* *0.08 Ind tot inc dec115162* ***0.01 Ind tot inc dec **0.02 Ind tot inc dec Ind tot inc dec **0.06 Ind tot inc dec Ind tot inc dec *0.13 Ind tot inc dec *0.14 Ind tot inc dec *0.16 Ind tot inc dec * *0.21 _cons-17822** * N7803 Adj R-squared Note * denotes statistical significance at the 1% level; ** at the 5% level; and *** at the 10% level
Wealth gap Average wealth Amount of gap explained with differences in characteristics Women's av. wealth if male characteristics Amount of gap explained with differences in coefficients Women's av. wealth if male's characteristics and wealth function Interaction Average wealth Women Men st err Wealth gap Average wealth Amount of gap explained with differences in coefficients Men's av. wealth if women's wealth function Amount of gap explained with differences in characteristics Men's av. wealth if women's wealth function and characteristics Interaction Average wealth Men Women st.err
Wealth gap Average wealth Amount of gap explained with differences in characteristics Women's av. wealth if male characteristics Amoount of gap explained with differences in coefficients Women's av. wealth if male's characteristics and wealth function Unexplained Average wealth Women Men JMP Women Men Oaxaca Note: Women as reference group
Table 8. Wealth decomposition results across the wealth distribution (Juhn-Murphy-Pierce). Wealth gap Amount of gap explained with differences in characteristics Amount of gap explained with differences in coefficients Unexplained 10th th th th th
Wealth gap Amount of gap explained with differences in characteristics Amount of gap explained with differences in coefficients Unexplained P90-P P90-P P50-P P75-P
Oaxaca decomposition suggests that half of the gap is due to differences in characteristics between men and women and half due to differences in coefficients JMP takes a closer look at the gap across the wealth distribution › Bottom: large part of the gap is due to differences in characteristics (diminishes) › Bottom: the way women accumulate wealth has a reducing effect on the gap › This is not the case past the median
Consistent result that a big portion of the gap is due to characteristics Quantile regressions Robustness checks for equal sharing assumption marriage patterns of men vs women Life-time work histories Control for differences within couples DiNardo, Fortin, Lemieux decomposition partition wealth vector: Labor market experience, education level, intergenerational characteristics and demographic characteristics Future work