Estimating trends in US income inequality using the Current Population Survey: the importance of controlling for censoring

Abstract

We analyze trends in US size-adjusted household income inequality between 1975 and 2004 using the most commonly used data source -the public use version of the March Current Population Survey. But, unlike most researchers, we also give substantial attention to the problems caused by the topcoding of each income source in the CPS data. Exploiting our access to Census Bureau internal CPS data, we examine estimates from data incorporating imputations for topcoded incomes derived from cell means and estimates from data multiplyimputed from parametric distribution models. Our analysis yields robust conclusions about inequality trends. The upward trend in US income inequality that began in the mid-1970s and increased in the 1980s slowed markedly after 1993.

Income Inequality, Mobility, and the Accumulation of Capital

“…0   and double the dispersion of the other two terms, i.e. , which roughly corresponds to the degree of dispersion reported by Burkhauser et al (2011) for recent years when they use the uncensored CPS data. benchmark case is reported in the first panel.…”

The Dynamics of Personal Income Distribution and Inequality in the United States

“…Furthermore, in this paper we did not correct for topcoded incomes and did not make any additional adjustments to the censoring of incomes, as applied to by the Census Bureau to high incomes. We propose to use income imputation methodologies, similar to the the ones developed in Piketty and Saez (2003) and Burkhauser et al (2011), in order to improve the resolution of top incomes. Moreover, given that CPS does not provide good data for the top end of the income distribution, it should be possible to obtain additional information from the tax returns data provided by the IRS.…”

“…The bin with the highest incomes is open-ended and includes all persons with incomes above the maximum income threshold. We should also point out the well-known problem of topcoding of high incomes, which may significantly affect the overall estimates (Larrimore et al, 2008;Burkhauser et al, 2011). This paper, however, presents a preliminary attempt to test our model and so we avoid discussion about correcting topcoded incomes and their cell means.…”

The Great Recession and Racial Inequality: Evidence from Measures Of Economic Well-Being

“…One estimate is that these changes accounted for half of the increase in the inequality in household money income between 1992 and 1993 or about one Gini point (Ryscavage 1995). In addition, top-coding of property income items (such as dividends) also underwent changes (Burkhauser et al 2011). underestimation of inequality prior to 1994, there is still evidence of a trend of increasing inequality since the 1980s. However, inequality in both MI and LIMEW (as measured by the Gini ratio) has remained stable since 2000, including during the period of the Great Recession.…”