Towards a Distribution of Household Income: Linking Survey Data to Administrative Data

Developing a national account based measure of the distribution of income from the commonly used Census based concept of money income has been the subject of earlier research—see Fixler and Johnson (2014) and Fixler, et al (2015) for example. A limitation of the earlier work is that the extrapolation from the survey data to the national account aggregate was based on “blow-up “factors that were constant across households. In this paper, we will explore using micro tax data to create income quintile specific blow up factors. More specifically, CPS data is linked to tax data by household in order to address misreporting and survey bias for several income categories. We find significant differences between the CPS and tax income for the same households, suggesting that simply replacing the survey income for the administrative income data is not satisfactory. Since the top incomes are significantly different, we create blow-up for the very top of the distribution, and recalculate distributional measures. Using these factors helps bridge the gap between micro data vs. macro statistics and also inform about results from other studies on aggregate income inequality, such as Piketty, Saez, and Zucman (2018).