A method of estimating the distribution of private wealth among American families, 1973.

Abstract

The method of estimation in this study relied on capitalization of certain flows recorded on income tax returns to estimate directly the gross value of certain classes of assets. Net wealth was then predicted for each family in a sample of 45,030, using estimated asset values as the independent variables in a linear regression equation. Coefficients of the assets were estimated from a sample of estate tax returns which provided net wealth and asset values for each return.

Families were ranked by net wealth levels into a frequency distribution from which a Lorenz curve was constructed and a Gini coefficient calculated. According to this estimate, in 1973 the lower half of the wealth distribution held 1% of net wealth and the third quadrant held roughly 10%. The upper decile of wealthholding families held almost 70% of personal wealth, and the Gini coefficient was .81. The distribution of assets and income was similarly analyzed by income class and by age group. Comparison of the wealth estimate with national balance sheet totals indicates that this method covers a substantial portion of total wealth and significantly enhances our knowledge of the entire distribution.

The dissertation describes the development of a method for estimating personal net wealth at the family level. Knowledge of the personal wealth distribution has been limited in the past to information gathered from estate tax returns, which include only the highest percentiles of the wealth distribution, and the results of a Federal Reserve Board-sponsored survey conducted in the early nineteen-sixties.