Commerce Secretary Robert A. Mosbacher refused this week to ''abandon a 200-year tradition of how we actually count people.''
But the U.S. census will never be the same.
The federal government spent $2.6 billion on the 1990 census and wound up with a result about as popular as an infectious disease. In fact, a top official of the Census Bureau's parent agency likened the census to cancer.
''We've known about cancer for a long time and still can't cure it. This is not a whole lot easier problem,'' said Michael Darby, assistant secretary of commerce.
The Census Bureau's director herself voted to junk the head count and replace it with a healthier statistical substitute -- Diet Census? -- to make up for large number of minorities missed by census takers.
And the nation's big-city mayors, including Baltimore's Kurt L. Schmoke, treated the census numbers as something akin to the plague and deplored Mr. Mosbacher's decision. ''The message to the American people is the poor don't count so don't count the poor,'' said Mayor Raymond Flynn of Boston, president of the U.S. Conference of Mayors.
The basic problem with the census is that it counts whites much more accurately than blacks and Hispanics. Understandably, mayors of cities with large minority populations don't like that. By the Census Bureau's own estimate, it missed 5.3 million persons in the 1990 census -- 4.8 percent of blacks and 5.2 percent of Hispanics, but only 1.7 percent of whites.
Actually, the results were even worse than they sound. The General Accounting Office, the investigative arm of Congress, found that the census missed 9.7 million persons. But because it counted 4.4 million persons twice, the net under count came to a mere 5.3 million.
In retrospect, it seems the 1990 census was doomed from the day the questionnaires went out. The census is largely taken by mail; anything else would be hideously expensive (even more hideous than $2.6 billion). Census officials expected 70 percent of U.S. residents to mail back their census forms, down from 78 percent in 1980. Only 63 percent did. The census never caught up.
Despite unprecedented promotion and appeals to their sense of civic duty, many Americans treated the bicentennial census as just another piece of junk mail. Census takers went out to find those who didn't respond by mail. But many U.S. residents, distrustful of the government, apparently went out of their way not to be counted.
After the count was completed, the Census Bureau did a survey to estimate how many people it missed. Based on the survey, it made a proposed adjustment of the count. Mr. Mosbacher rejected the cure because he said it might be worse than the disease.
The U.S. almost certainly has closer to 254 million residents (the adjusted figure) than 248.7 million (the census count). But experts disagree whether Baltimore, for example, has closer to 772,000 residents (adjusted) or 736,014 (census). As the adjusted figures are applied to smaller and smaller geographic areas -- all the way down to city blocks -- they become less and less accurate, perhaps much more inaccurate than the census itself.
Baltimore clearly would be better off with 772,000 people. Federal aid based on population would have increased to the city -- by how much no one has calculated with any precision because many different funding formulas are involved and all are complex.
Perhaps more important, the city would have retained more political power. When Maryland General Assembly districts are redrawn next year, each will have an ideal size of 101,733. But Baltimore (population 736,014) has enough residents for only about 92,000 persons in each of eight districts -- one less than the city has now.
''It makes our task much more complicated and difficult,'' said Sen. John A. Pica Jr., D-Baltimore, chairman of the Senate Committee on Redistricting and Reapportionment. For the city to prevail, Mr. Pica and his city colleagues must forge split city-county districts that Baltimore can dominate.
''It's another blow to major urban areas,'' he said. ''We are steadily losing federal money, and now this is a shot from a different angle, where we lose in the legislatures as well.''
However, if Mr. Mosbacher had adjusted the census, Maryland as a whole would have been a loser. True, the state's population would have increased by 87,000, but its share of the U.S. population would have declined slightly. Maryland presumably would have received a smaller slice of the pie of federal funds.
''Because there is a loser for every winner, we need solid ground to stand on in making any changes. I do not find solid enough ground to proceed with an adjustment,'' Mr. Mosbacher said. The commerce secretary feared that adjustment ''would open the door to political tampering with the census in the future.''