Updates to the Sampling of Wealthy Families in the Survey of Consumer Finances

Participation in household surveys has fallen over time, making it harder to produce a household survey-like the Survey of Consumer Finances (SCF)-in a timely manner. To address these challenges, the reference year of the sampling frame data for the 2016 SCF wealthy oversample was shifted back one year, allowing the oversample to be selected earlier than the past. In implementing this change, though, we risk identifying an outdated set of families and introducing variability in the sampling process. However, we show that the set of families selected in the new frame are observationally equivalent to those that would have been selected from a past frame, and that the increased variability of wealth estimates is compensated-for with the use of more comprehensive data than in the past. Other aspects of the SCF sampling process are revisited, too. We continue to find support for using permanent income in the sampling process, rather than annual income. We also estimate the geographic distribution of wealthy families and show that the current distribution is similar to the past. We propose adding one geographic area to the oversample, though, and supplementing by 100 the set of sampled families.