The Biden administration is proposing to issue a new directive on how U.S. agencies collect and publish data on race and ethnicity. Discussions among some technical users of the data have identified inconsistencies and gaps in how the races have been defined and differentiated and concerns about the design for collection. But the proposed directive, issued by the Office of Management and Budget in January, deserves a wider audience because it threatens individual liberty. Comments on the proposal are due by April 27.

The proposal would increase the number of primary racial groups for data collection to seven from five. First, it would split the current category of “white” into two—one called “white” for people with European ancestry and another called “Middle East and North Africa,” or MENA. Second, it would redefine “Hispanic”—currently an independent ethnic classification—as a race. That yields seven races: white, black, Asian, Hispanic, American Indian, Native Hawaiian and MENA. Respondents to government surveys or forms would be asked to select one or more races from the list. This would replace the current method by which people are first asked whether they are “Hispanic or Latino,” and then asked to select races from the original five options.

Besides increasing the number of primary races, the directive would require that most racial data be collected at finer levels of detail, typically specifying a person’s country of origin and sometimes an ethnic group within or across countries such as Hmong, Roma, Bantu or Kurdish. Among those listed as black or African-American, it would require classifying separately descendants from slaves in the U.S. Descendants of Caribbean or Latin American slaves would be classified only by the country from which they or their ancestors immigrated.

The proposed OMB directive claims to be for obtaining self-identification, but in practice government officials apply it to determine eligibility for benefits such as those from the Small Business Administration. Despite the directive’s statement that “the categories are not to be used for determining the eligibility of population groups for participation in any Federal programs,” in other places it says otherwise: (1) “Data could be used to allocate program or initiative benefits”; (2) the proposal “provides a minimum set of categories that all Federal agencies must use regardless of the collection mechanism (e.g., … program benefit applications)”; (3) “MENA population counts could be used to allocate needed resources”; and (4) “foremost consideration should be given to data aggregations by race and ethnicity that are useful for statistical analysis, program administration and assessment, and enforcement” (emphasis added).

The bureaucracy behind this proposal is focused on being able to control, reward and punish the population by racial classification. American history doesn’t reassure us that such classifications are benign. Even after the ratification of the 13th, 14th and 15th amendments, government continued to create and enforce Jim Crow restrictions and real-estate red-lining on American citizens classified as black, in addition to sending those classified as Japanese to internment camps. And it hasn’t stopped there. In his book “Classified,” George Mason University law professor David Bernstein documents some of the many ways that the racial classification schemes flowing from the current version of this statistical policy directive are used to distribute rewards to some and punishments to others.

Government collecting and acting on individuals’ beliefs about their race or ethnicity is antithetical to America’s founding principles. We are a country of “all men are created equal” and “e pluribus unum.” This initiative is trying to create even more categories by which people can be divided, separated, discriminated against or given special favors. When government collects and classifies individual data by personal racial and ethnic characteristics, it lays a foundation for discrimination, even oppression.

Current census data show about 20% of new marriages are between people of different races or ethnicities, and for some groups almost half cross racial or ethnic lines. OMB uses this trend to justify collecting more data on race to accommodate “a growing number of people who identify as more than one race or ethnicity.” The office misses the point of its own argument. If a growing proportion of the population has multiple racial or ethnic roots, that means when establishing the most intimate and significant of human relationships—the family—race and ethnicity are becoming less important. Why, then, should government insist not only on keeping the distinctions alive but also on reinforcing and expanding them? It is time to stop discriminating by race in our statistics. Treat people as individuals, not merely parts of groups.

Eliminating data by race isn’t some obscure, fringe idea. France, by law, does not collect or publish data by race because such data formerly enabled the betrayal of French Jews to the Nazis by the Vichy government and were used to discriminate against minorities during the late colonial period. The French have the right idea, and so did Chief Justice John Roberts, who wrote in 2007: “The way to stop discrimination on the basis of race is to stop discriminating on the basis of race.”