Census Decision May Cheat Minorities
by John Barry, Policy.com Thursday, March 8, 2001
Republicans just don't like to recount. They don't like to recount election votes. And they don't like to readjust census votes. And it looks like they're getting their way -- again.
In a surprise decision, the U.S. Census Bureau recommended against altering Census results that had been slated to be adjusted to make up for potential undercounts, by using a method called "statistical sampling." The effort has had to be abandoned because, according to William J. Barron, the Census Bureau's acting director, time constraints would make it impossible to determine whether shifting the totals would arrive at more accurate figures.
The Census Bureau doesn't have the final word on the subject. Last month, in one of his first acts as Commerce Secretary, Donald Evans stripped the Census Bureau of the authority to adjust census sampling for perceived undercounts. He also announced the appointment of six experts to advise him on the issue: three of these advisers have already come out publicly in favor against adjustment. And in 1999, the Court barred using sampling adjustments for purposes of determining the number of House seats each state would be allotted, but permitted it for determining Congressional redistricting within each state. So, the Rehnquist Court, having already begun chipping away at statistical sampling, could very well have declared any readjustment unconstitutional.
In any case, it's a reasonable bet that the figures would have remained as they are. But this recommendation by department professionals will undoubtedly add weight to Mr. Evan's case that when it comes to census figures, we should stick to the first figures.
The first group of redistricting numbers based on unadjusted data should be issued next Wednesday, March 14. Current numbers and estimates are available at http://www.census.gov/main/www/cen2000.html.
Memories of the election undercount were remarkably short-lived, and it's likewise unlikely that the Census 2000, and its margin of statistical error will make the headlines for long. But this may be Bush's opportunity to place his stamp on the American political scene for at least a decade. Not only will the census results determine the new district lines for House seats and thousands of state and local election districts; they will also be the basis for disbursement of about $200 billion in federal funds.
Census undercounts are nothing new, and they have generally favored Republicans. The 1990 Census, conducted under the aegis of the elder Bush, included a net undercount of 4 million out of the total population of 248 million, or 1.6 of the population. Most of those left out are more apt to fit the profile of Democratic voters -- urban, nonwhite and generally poorer than the mean of those counted. Hispanics were roughly twice as likely not to be counted; African-Americans three times as likely not to be counted; Native Americans five times as likely not to be counted.
In 1990, Democrats in both houses mounted a battle against the Bush Administration, claiming that the Republicans were denying minorities an accurate count for political purposes. Not surprisingly, when professionals in the Census Bureau joined these Democrats in arguing for adjustment of the final figures, President Bush rejected their report.
This time, Democrats will be fighting the same battle, but without the support of the professionals at the Census Bureau. The estimated undercount for the 2000 Census -- 1.18 percent -- is smaller than it was in 1990. But its consequences could be serious for many urban areas and the Democratic Party. . Florida's federal take may, for instance, be drastically reduced, in what might to many Republicans seem as getting their just desserts for almost deserting their keeper's brother last November. Ten years ago, 250,000 Floridians were not counted in the census, costing the state about $2 billion in federal aid. This year, the undercount is feared to be much larger, given the influx of immigrants, and the practice of "doubling up" in single homes.
According to the Public Policy Institute of California, California lost $2 billion in federal funding over the last decade and one seat in Congress because of the 1990 Census figures. This year, Californians had anticipated economic and political gains from the census. No such luck, it seems: unadjusted figures will leave that beleaguered state with far less in federal funds than projected.
It comes as no surprise that a Republican administration is pushing a set of figures that favors its followers politically. But why, many Democrats are wondering, did the pros at the Census Bureau leave them in the lurch, when the experts there had been advocating statistical analysis for years? Barron, the acting director of the Census Bureau has said that there was no accurate evidence that a new count would be any closer to the truth. Critics are wondering aloud if Bureau staff are "running scared." And some are blaming the law, which they say offers no time for "recounts."
But there's no ambivalence in the reaction of Rep. Dan Miller (R-Fla.), who recently gave voice this way to Republican glee at what appears to be a fatal blow for statistical adjustment of the census:
"Match. Set. Won."
|