Funding Levels May Be Altered Before Census

Defending his decision not to adjust the 1990 census, Commerce Secretary Robert A. Mosbacher assured Congress Tuesday that statistical surveys may be used to adjust funding levels for federal programs before the next census in the year 2000.

An adjustment would mitigate the effects of a low population count on cities and states with large minority populations. Such jurisdictions would be most affected by Mosbacher’s decision not to add to census figures about 5.4 million people tabulated by a statistical survey but not counted originally. Because census figures are used to distribute federal funds as well as to reapportion Congress and state legislatures, adjustment would benefit cities and states with large minority populations but harm cities and states whose count was more accurate.

“We may be able to do it in a couple of years,” Mosbacher told the House Post Office and Civil Service subcommittee on census and population. “I have no bias against adjustment.” The panel includes several big city Democrats who harshly criticized the secretary’s refusal to adjust the census as a deliberate slight to minorities.

The Post Enumeration Survey, which Mosbacher refused to use Monday, found that about 5% of blacks and Latinos, and 2% of whites, were missed by the traditional census head count. But Mosbacher said that he would need convincing evidence of the survey’s superiority to break a 200-year tradition of simply counting heads in the best way the government knows how. The census has been done in much the same way since 1790.

However, Mosbacher, seconded by Census Bureau Director Barbara Everitt Bryant, left open the possibility that the next census would use an improved technique and include some adjustment in its final result. He said that he would accept an adjustment if the government and outside statisticians could agree that it would improve the count by 80% or 90%. By comparison, the 1990 Post Enumeration Survey would have improved the census by 50% to 60%, he said.

The adjustment would have falsely inflated the head count in 21 of 50 states and in 11 of the 23 largest cities, Mosbacher said. Those 11 cities, he said, include New York, Chicago, Baltimore and Washington--all with large minority populations, and all of which would have gained up to 5% more people if the adjustment had been made. The cities, along with Los Angeles, are among others who have gone to court or threatened litigation to force an adjustment.

But Mosbacher made it clear that he would draw the line at applying any such adjustment to the more fundamental business of electoral reapportionment and congressional redistricting.

The methodology used in the Post Enumeration Survey, based on a statistical sample of 171,390 households nationwide, is so subjective that it could easily be “skewed to political advantage,” he told subcommittee Chairman Thomas C. Sawyer (D-Ohio). Sawyer had complained that refusing to use the post enumeration figures to change the census count “has the effect of a gerrymander on a national scale.”

If the adjustment had been made, California and Arizona would have been expected to gain another congressional seat and Wisconsin and Pennsylvania would have been in line to lose one each.

Census Director Bryant, who had favored the adjustment, testified that a straight head count of the nation’s population would never include about 2% of the population--people “so disconnected from society that we will never get them through enumeration.”

Undercount in Largest U.S. Cities

Here is a list of the most populous U.S. cities and proposed corrections to their 1990 census counts.

City Census Corrected % New York 7,322,564 7,552,000 3.0 Los Angeles 3,485,398 3,671,000 5.1 Chicago 2,783,726 2,857,000 2.6 Houston 1,630,553 1,716,000 5.0 Philadelphia 1,585,577 1,606,000 1.3 San Diego 1,110,549 1,156,000 3.9 Detroit 1,027,974 1,065,000 3.5 Dallas 1,006,877 1,058,000 4.8 Phoenix 983,403 1,014,000 3.0 San Antonio 935,933 964,000 2.9

Source: Associated Press