GOP Census Politics

This op-ed argues that the Census Bureau’s attempt to use sampling to increase the accuracy of the 2000 census is a laudable effort.

December 7, 1998

The Nation

December 7, 1998

GOP Census Politics

By Richard R. Buery Jr.

At the beginning of his classic novel of racism and alienation, Ralph Ellison’s African-American protagonist tells the reader, “I am an invisible man . . . I am invisible, understand, simply because people refuse to see me.” He was right, of course, at least as far as the census has been concerned.

The census has always undercounted the population. But some Americans are missed more often than others. The Census Bureau calls this phenomenon the “differential undercount”: In 1990, the census missed 12 percent of Native Americans living on reservations, 5 percent of Latinos, and 4.4 percent of African-Americans. In contrast, only 0.7% of non-Latino whites went uncounted. Renters were more likely to be undercounted than were homeowners, and children were more likely to be missed than were adults. The worst undercount-40,000 people-took place in New York’s 16th Congressional District in the South Bronx, the poorest district in the country and one whose population is 93 percent African-American and Latino. The highest overcount-70,000 people-took place in New York’s 3rd Congressional District in Nassau County, the country’s third-wealthiest district and one whose population is 91 percent white. And despite the Bureau’s exhaustive and expensive efforts to catalogue the population, the 1990 census has the dubious distinction of being the first census that was less accurate than its predecessor. The Bureau predicts that, unless changes are made, both the total undercount and the differential undercount will be worse in 2000.

The Bureau wants to do better. By using sophisticated data-gathering and statistical-sampling techniques to augment the direct count, it believes it can reduce the total and differential undercounts and save money to boot. The National Academy of Sciences agrees, as does about every statistician worth his or her salt. In 1990, the Census Bureau thought such sampling were the way to go, but Republican officials in the Bush Administration overruled the Bureau’s experts. The courts refused to intervene.

The Clinton Administration seeks to improve the 2000 census by using sampling techniques. But the Republican majority in the House, led by Speaker Newt Gingrich and Rep. Bob Barr of Georgia, sued the Administration, claiming that the use of sampling would violate the Constitution as well as the Census Act, the law authorizing the Bureau to conduct the census. Two lower federal courts have agreed that the Census Act prohibits sampling (neither court considered the constitutional question). The Supreme Court will hear the Administration’s appeal on November 30th.

Why would the House Republicans sue to block a cheaper and more accurate census, one that promises to include, rather than exclude, the most invisible elements of society? The answers, of course, are politics and money. Census data are the basis for appointment of seats in the House, as well as for drawing new Congressional districts within the states for state and local government. In addition, billions of dollars in federal funding are distributed each year on the basis of census data. Republicans know that people of color, renters, recent immigrants, linguistic minorities and urban residents-the very folks the census keeps missing-tend to vote for Democrats. They fear that if the census starts counting people accurately, Democratic urban areas will gain power and federal funds at the expense of suburban GOP strongholds. Throughout the country, town councils, state legislatures, perhaps even the House itself, might actually begin to look like the people who actually live in America, rather than just the people who get counted every ten years.

By opposing sampling, the House Republicans are engaging in gerrymandering of the worst kind. By maintaining the differential undercount, the GOP ensures that a white middle class voter’s ballot has greater weight than that of a poor African-American or Latino voter. For example, according to the 1990 census, New York’s 3rd and 16th Congressional Districts both had populations of about 580,000 people. However, the differential undercount overestimated the 3rd’s population and underestimated the 16th’s. As a result, in the last election, poor African-American and Latino voters in the South Bronx elected one congressman, Democrat Jose Serrano, to represent about 620,000 people, while middle class voters in Nassau County elected Republican Peter King to represent about 573,000 people. Talk about getting more bang for your buck.

The GOP’s constitutional and statutory arguments against sampling are meritless. The central theme of the Constitution, and of laws like the Census Act and Voting Rights Act of 1965, is, in the words of the Supreme Court, to grant all people an “equally effective voice” in government by promoting the principle of “equal representation for equal numbers of people.” Thus, the Supreme Court has held that population variation between Congressional districts of as little as 0.7 percent may violate the Constitution’s sacred “one person, one vote” principle. Just as the Republicans may not simply pack people of color into larger districts, they may not pretend that predominantly African-American and Latino districts are less populous than they truly are. To argue that either the Census Act or the Constitution mandates this result is to pervert their very essence.

The sad truth revealed by the census litigation is that the GOP is not interested in appealing to people of color as voters. Instead, they prefer to pretend that we are not here. But we are here, of course. And our numbers are growing. According to Census Bureau projections, by 2050 more than 47 percent of Americans will be Latino, African-American, Native American or Asian-American/Pacific Islander (and that doesn’t include Puerto Rico). The GOP is afraid. And it probably should be, given the callous and cynical disrespect that its litigation displays towards the undercounted.

Purposefully perpetuating the differential undercount can only further alienate the poor and people of color from American democracy. What surer signal could there be that we are, like Ellison’s hero, invisible to America? We must not forget that before the abolition of slavery, the Constitution required that only three-fifths of African-Americans would be counted for apportionment purposes-the ultimate differential undercount. Today, while more than three-fifths of Africans-Americans are counted, the “differential undercount,” and its message of exclusion, remains. The Census Bureau’s attempt to use sampling to increase the accuracy of the 2000 census is a laudable effort to extend equal citizenship to all Americans, regardless of who we are or where we live. One would hope that the Supreme Court will not stand in the Bureau’s way.

------------------------------------------------

ABOUT THE AUTHOR

Richard R. Buery, Jr. is a staff attorney at the Brennan Center for Justice at New York University School of Law. He co-authored an amicus brief filed in the Supreme Court’s upcoming census case.