Earlier this month, the United States Census Bureau declared New York City to have the highest population ever, based on a new estimate. According to the bureau's newly-released annual county population estimates, New York City's population now stands at 8,085,742. The estimate of population growth since the 2000 Census is as follows:

38,000 from 2000 to 2001
17,000 from 2001 to 2002
14,000 from 2002 to 2003

In other words, according to the bureau, the city's population is still growing, though its rate of growth has substantially slowed since September 11th.

All of this was duly reported in the press. However, some demographers disagree with the bureau's method of estimating population.

Warren Brown of Cornell University says that the count is much too low. Currently he is helping New York City launch an informal challenge to the estimate. For example, the census bureau is saying that Brooklyn lost population from 2002 to 2003. But city officials point outthat the borough has gained some 4,000 housing units since the 2000 census 2000. Surely, they say, this is concrete evidence that the borough's population is on the rise.

But other demographers, myself included, believe that the new estimate for the city's population may instead be too high. I think it likely that New York City has actually been losing population since September 11th.

CENSUS ESTIMATES REQUIRED BY LAW

Federal law requires the census bureau to make estimates of the nation's population in-between the official census, which, as required by the U.S. Constitution, is conducted every ten years. These estimates are used for many purposes, including: allocating over $200 billion in federal funds; calculating such statistics as birth, death and crime rates determining the location of public and private services; computing controls for federal surveys;

To estimate the population the census bureau uses the "components of population change" method, which takes the number of births, subtracts the number of deaths and adds or subtracts the net number of people who migrated into or out of New York City. This number is then added to last year's population estimate.

If any of these figures are wrong the estimate will be wrong. Births and deaths, of course, are relatively simple numbers to acquire, since they are recorded throughout the United States. But gauging migration is much trickier. The bureau uses two estimates -- one for net domestic migration and another for net international migration. The Internal Revenue Service produces a file that reports the number of people who change the county in which they file tax returns from one year to the next. This file is considered quite reliable except for its ability to track college students, military personnel and others who may move unpredictably.

Getting truly accurate estimates is challenging, especially in places such as New York City.

TRYING TO CORRECT PAST INACCURACIES

During the 1990s, as I have discussed in earlier columns, the census estimate of the growth in New York City was too low by about 50,000 per year. It became obvious that the way the bureau estimated the growth of the foreign born population was grossly inaccurate. Since the bureau used the number of individuals who received permanent resident status by county in a given year, they missed people who entered or stayed in the United States illegally, those on student and work visas, or those who are visiting the United States for other reasons.

So beginning with the 2002 estimates, the census bureau began to use the new American Community Survey(ACS) to estimate changes to the foreign born population. This survey will replace the census long form, which asked about 50 questions of roughly one-sixth of the population in 2000. It will be a large continuous collection of social and economic data and will offer a continuously updated picture of every locale in the United States, down to each census tract (which in New York City can be a single block). But this survey is not yet implemented in every county in the United States, because it is yet to be fully funded. For this reason, the bureau cannot determine if the number of new foreign born at the national level has changed significantly since the 2003 data are not yet available.

Furthermore, they are not yet confident in using information at the county level from this data set. To estimate change in foreign born in each county, they assumed that the pattern found in the 2000 Census continued through 2003. Therefore, this method does not take into account changes in the number and destination of new foreign born caused by a reaction to 9/11 or the worsening New York City economy. According to Signe Wetrogan, the assistant division chief of the bureau's Population Estimates and Projections of the Population Division, bureau officials are reviewing methods and data, so that they can create leading indicators of population change, such as employment figures and yearly school enrollment.

THE EFFECTS OF CHANGING MIGRATION OF FOREIGN BORN

All these technical details mean that the estimated net number of new foreign born coming to New York City each year since 2000 is about 104,300. If Warren Brown is right, and the foreign born or other migration estimate is too low, then the city's rapid 1990s growth may indeed be continuing. It is certainly possible that the Census's new method has missed many new foreign born.

However, I think it is more likely that just as net out-migration to the rest of the United States increased after 9/11, the number of new foreigners coming to New York City would have decreased, as well. If the number of new foreign born actually declined by 17,000 or more after 9/11, then New York City has lost population in the last two years.

Until the census bureau refines its methods we cannot be sure if New York City's population is growing, declining or stable. However, I think it is important to be patient with the bureau and its efforts to get this right. When it has more confidence in its data and methods, it will simply re-estimate the population change in New York City, not just for one additional year, but for all of the years back to 2000.

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.