Zip-code and surname analysis can impute racial tax data

Evidence

One of the simplest, if imperfect, ways for the IRS Statistics of Income division (SOI) to access this information would be to apply a zip code analysis coupled with a surname analysis, where last names are used to broadly classify taxpayers by ethnicity, to impute race and ethnicity data. Another option is to add questions about race and ethnicity to existing surveys—the Comprehensive Taxpayer Attitude Survey, which collects opinions on taxpayer compliance and trust in the IRS, is already conducted regularly (though the SOI would have to ask for Social Security numbers if the answers were to be matched with specific tax returns). However, even without IRS participation, we could take advantage of other data collection that is currently in place—the Survey of Consumer Finances, for example, conducted by the Federal Reserve, already collects race and ethnicity data. They’d simply need to begin including tax return (1040) items in their own survey and publishing the results. — Dorothy A. Brown, The Whiteness of Wealth, 518

Categorization

• XREF: Directly answers the reader’s tracking of how she argues around the IRS data hole — concrete fallback proxies for missing racial data.