A new proposed rule submitted by the Census Bureau seeks to restrict the inclusion of certain demographic questions in the decennial census (next due in 2030). In this compliance alert, we consider the potential impact of the proposal on key analysis practices such as benchmarking.
Summary
On Thursday, 10 September, the Census Bureau submitted a proposal to add two new parts to the Code of Federal Regulations governing the decennial census of population and housing. The second of these concerns “establishing standards and restrictions for the inclusion of certain demographic questions”—specifically, it would prohibit questions about race and ethnicity, as well as sexual orientation.
The Census Bureau argues that the collection of such data does not advance “the fundamental, Constitutional purpose of the decennial census” (i.e., ascertaining the population count for apportionment). Furthermore, it observes that a greater number of questions in the census tends to correlate with lower response rates.
It also cites “Substantial doubts about the meaning, significance, and impact of attempting to categorize persons by race and ethnicity”, pointing to Students for Fair Admissions, Inc. v. President & Fellows of Harvard, and Executive Order 14151 (Ending Radical and Wasteful Government DEI Programs and Preferencing) as examples of recent shifts in the law.
Anyone who wishes to comment on the proposed rule has until October 13, 2026, to submit their input.
Who Is Impacted and What Actions Are Required?
Should the proposed rule pass into law, the change will have an impact on organizations that rely exclusively on Census demographic data for workforce analytics, EEO benchmarking, pay equity studies, diversity reporting, and labor market analyses.
However, it should be noted that the Census is only one of a range of potential sources for benchmarking data. The proposal will not restrict the Census Bureau’s ability to collect demographic information through other surveys, including, most significantly, the American Community Survey (ACS).
Nonetheless, organizations may have found the different characteristics of Census data and ACS data useful for certain purposes. Census data is obviously refreshed less frequently and already involves fewer questions, however, it is based on a significantly higher number of responses. By contrast, ACS data involves more questions, is updated annually, but only provides estimates based on 12 and 60-month averages extrapolated from small samples (and its 1-year estimates only study areas with populations of 65,000 and more).
For now, organizations should closely monitor the rulemaking process and begin exploring the viability of replacing Census demographic data with ACS data or other alternatives.
How Affirmity Can Help
Affirmity software allows you to use multiple availability data sets, with ACS data among them. In the event that the proposed law passes, Affirmity clients will be able to make data set selection adjustments accordingly.
Affirmity is here to help you maintain data readiness and to benchmark your workforce against available talent pools. ACS data is available across our software and services portfolio, including:
- Workforce analysis software and services
- Workforce compliance services
- Compensation and pay equity analysis
Get the internal and external demographic information you need for a risk-free organization strategy: Contact our team of experts today.
About the Author
Kim Hendon oversees account management and sales for Affirmity. She is responsible for building successful, long-term partnerships with clients and generating new business. Having served with the company for more than 25 years, Ms. Hendon has in-depth knowledge and broad experience in all areas of workforce analytics and HR compliance.
Ms. Hendon assists clients with the planning and development of workforce compliance and non-discrimination programs, as well as employee engagement initiatives. She holds a Bachelor of Arts in Speech Communication and a Master’s in Business Administration. Connect with her on LinkedIn.