Exploring the possibilities and boundaries of survey data for the analysis of wealth and wealth transfers

Abstract

In Germany, a flagrant lack of official register or tax data for scholarly use leads to a situation wherein survey data is the last remaining source of evidence about the distribution of wealth. Two of the four research chapters in this thesis aim to evaluate methods for the improvement of available survey data. The other two contributions discuss the possibilities and limitations of survey data for the analysis of the joint distribution of wealth and wealth transfers. The first research contribution “Longitudinal wealth data and multiple imputation -- An evaluation study” is a simulation exercise that intends to improve the imputation of missing values in wealth surveys. Considering trend and inequality estimates, the univariate row-and-column methods performs surprisingly well. Its combination with MICE as fallback procedure unanimously improves the imputation quality for all asset types considered. However, researchers interested in wealth mobility might prefer the imputation with MICE as basic and fallback, as it best replicates the mobility structures observed in the original data. The chapter “Estimating top wealth shares using survey data -- An empiricist's guide” picks up where the previous one left off: the treatment of missing values (item non-response) is not sufficient if individuals or complete households refuse to participate in a survey (unit non-response). In a series of Monte Carlo experiments this contribution shows that using maximum likelihood techniques to simulate a Pareto distributed top tail does not improve the estimates, as aggregate wealth and shares are still biased downward. The addition of rich list data does improve--and potentially overestimate--top wealth shares, while still yielding aggregate numbers that are too low. The fourth chapter “Breaking down Germany's private wealth into inheritance and personal efforts -- A distributional analysis” takes advantage of the records of both household wealth and wealth transfers received from inheritances and gifts in the questionnaire of the survey Panel on Household Finances, which enables us to compute the significance of inheritance for different quantiles of the distribution of wealth in Germany. We show that wealth inequality, at least for 99 % of the German population, is hardly the result of an unequal distribution of inherited wealth: the ratio is one-third and barely changes along the distribution of wealth. This observation is nigh on identical for the sub population of retirees. The addition of pension wealth reduces the significance of inherited wealth for a household's financial situation particularly in the bottom half of the wealth distribution. The final research contribution “Comparing the joint distribution of intergenerational transfers, income and wealth across the Euro area” expands upon the previous chapter with additional analyses in a European context.