Gotham Gazette

The Place for New York Policy and politics

The U.S. Census Bureau released the first results from Census 2000 recently, and the numbers are surprising. Between 1990 and 2000 New York State appears to have grown by almost one million people. Yet, at the same time, as a result of these new numbers, the state will lose two congressional seats: The delegation will have 29 members, the smallest number since 1820.

How could New York have gained so many people and still lost political representation?

Throughout the United States, the Census found 6.8 million extra people, putting the U.S population at 281.4 million people. This was an increase of over 13 percent from 1990. For all the increase in numbers in New York (whose population as of April 1, 2000 was found to be officially 18,976,457), the percentage of increase here was actually below the national average. The Census found a greater percentage in other states. There were 1.3 million more people in California and over 700,000 in both Texas and Florida. New York has sunk to third most populous state, after California and Texas.

Still, the numbers that the Census Bureau found in New York were far larger than the estimates the Bureau had made. Comparing its estimate with the population the Census Bureau actually found, New York State has about 830,000 "extra people."

New Yorkers and others will ask: "Who are these 'extra' 830,000? Where do they live? Why were they missing until Census 2000?" The answers to these questions could have effects on the balance of political power in New York State, as well as on the distribution of federal and state funds.

They also raise questions about the accuracy of the Census count, both in 1990 and 2000.

There are very good reasons to expect many of the "extra 830,000" live in New York City and its environs. The 1980 and 1990 Censuses had both an undercount and an overcount. The undercount missed many in the urban areas, especially those who were poor, immigrant, or lived in unconventional housing arrangements (e.g. female headed, roommate situations, etc.) The Census historically has been very good at counting married couples with children living in a house they own. In fact, some households, especially those which are wealthier, are counted more than once. Sometimes this is because these families have two or more homes.

Because of these problems, in 1990, the Census selected a large sample of households to revisit in person. Extraordinary means were used to get an accurate count in each of these households a few weeks after Census Day. This count was then compared to the Census Day enumeration. From this effort the Bureau was able to estimate the undercount and overcount, and create adjusted totals for each block, tract, city, county and state in the United States based upon various characteristics, such as race, owner or renter, and so forth. This method increased the count in large cities, especially in minority and low-income areas. It also reduced the count in suburbs and other wealthier areas. New York City was undercounted by a total of roughly 250,000 in 1990.

According to a federal government study in 1990, there was a nationwide undercount of 10 million and an overcount of 4.4 million. The efforts to estimate the undercount and overcount were spurred by a court case originally filed in 1980 in New York City. The settlement of the case included plans to do a scientific adjustment in 1990 based upon the special sample. A panel of the National Academy of Sciences endorsed this approach. Nonetheless, the Secretary of Commerce in the first Bush Administration rejected the adjusted numbers, so they were not used for redistricting. This had the effect of taking political power and federal and state funds away from the urban centers and shifting them to the suburbs and less populated areas.

The 2000 Census carried out an even larger survey for similar purposes. It remains to be seen whether the younger Bush's Census Bureau will allow an adjustment.

Before the new numbers were released and most people thought New York State was just maintaining its 18.1 million population, the New York Public Interest Research Group (NYPIRG) and others noted that New York was losing population upstate and gaining population downstate. It is not that the new census numbers mean this analysis was wrong, but rather that the Census Bureau consistently understated the size of upstate and downstate. If most of the "extra 830,000" are in New York City, then the political geography of New York State may change decisively. In 2002, each Congressional District must contain 654,361 people, up from 585,000 in 1990. If most of the extra New Yorkers live downstate, when they are combined with population shifts it is possible that most or all of the loss of Congressional representation will be borne by upstate areas, particularly those not in cities. Furthermore, when the State Senate and Assembly are redistricted, New York City might gain as many as five assembly seats and two senate seats.

Originally such a shift would seem to have required an adjustment to the Census 2000 count. But now it is not obvious how important such an adjustment may turn out to be. The Bureau implemented many methods to improve the coverage and results for the 2000 Census. These included efforts to reach more households and get a higher mail back response rate, as well as many efforts to reach those who did not return the questionnaire by mail. We do not yet know how effective they were. However, since the Census found an extra 6.8 million Americans, roughly 830,000 in New York State, one might expect a much lower undercount in 2000 than 1990. Even without any adjustment, the next redistricting may shift power to cities and urban areas throughout New York State, and many other states. With an adjustment the shift would be even more pronounced. At this point, all involved are waiting for the first detailed results, which will be released by March 31, 2001. Then we will begin to see some of the important consequences of the 2000 Census in dividing up the votes and money.

Andrew A. Beveridge has taught sociology at Queens College since 1981, done demographic analyses for the New York Times since 1993, and provides expert testimony on a range of cases, including housing discrimination. The opinions expressed are his alone.