A Review: Estimating the Concentration of Wealth in America

Abstract

In 1986, the Joint Economic Committee of the U.S. Congress released a study indicating a sharp increase in the concentration of wealth in America. The committee estimated that the richest .5% of families held 25% of the wealth in 1963; by 1983, this proportion had risen to 35%. These findings came from Surveys of Consumer Finances (SCF) conducted by the University of Michigan Survey Research Center (SRC). The surveys were based on household samples, supplemented by samples of very rich persons selected by the Internal Revenue Service. Shortly after the report was released, the finding for 1983 was found to be in error. One respondent, who had received a very large weight in the analysis, had $200 million of wealth attributed to him when the correct figure was $2 million. When the figure was corrected, the estimated share of wealth held by the richest .5% dropped to 27%. The error was announced shortly after the release of the report, but many commentators had already offered opinions concerning President Reagan's economic policies. Even as recently as June 1987, the Texas Commissioner of Agriculture criticized the Reagan philosophy of greed and stated that fifteen years ago, the richest one percent of American families held 27 percent of the nation's wealth; today they are approaching and may soon surpass the 36 percent peak share attained in 1929 (Hightower, 1987). The error matters because it affects an important debate. Its magnitude is similar to the error committed by the Literary Digest (1936) when it underestimated Roosevelt's share of the vote in the 1936 presidential election by 20% and predicted that he would lose to Landon. Explanations for both errors have focused on survey procedures. For the Digest, blame is cast upon a poor sampling frame and low response rate (Freedman, Pisani, and Purves, 1978:302-304). For the wealth estimate, blame is given to an interviewer error (Smith, 1986).