U . S .    C E N S U S    O F    H O U S I N G   &   P O P U L A T I O N



See the HOUSING issues page
Fixing the Census
NYTIMES
By Alan B. Krueger (Alan B. Krueger is an economics professor at Princeton).
January 26, 2009, 6:31 am

Serious problems in the planning for the 2010 census have been in the news lately. The census has fallen well behind schedule because of technology glitches, and as a result the Government Accountability Office has listed the population count as one of the 13 urgent issues requiring immediate attention in the first year of the new presidential administration, up there with homeland security and Iraq. Without urgent action to prepare and test survey procedures, the 2010 census will miss more people than the 2000 census.

Why should you care? The census provides the foundation for computing many economic statistics, including the unemployment and poverty rates, and is the basis for congressional redistricting. Census data are also used to allocate money to states and local areas. Minorities and city dwellers are more likely to be undercounted by the census. Thus, problems with the census have serious economic and political consequences.

This has me thinking about what David Freedman would say.

David Freedman was a statistician and probabilist at the University of California, Berkeley, who died of bone cancer at the age of 70 on Oct. 17. He made seminal contributions to the theory of statistics as well as to applied topics such as statistics and the law, and the efficacy of mammograms. His work is summarized and available at his Web page.

Mr. Freedman was also a leading skeptic of the view that the census could be improved by statistical adjustments. The Census Bureau has developed an elaborate post-census estimation procedure to assess the number of people who were missed and to possibly adjust the original population count. In this adjustment procedure, a second sample is surveyed after the original census, and the extent of overlap between the census and the second sample is used to adjust the original count.

This procedure is based on a technique known as the capture-recapture method. It works if the chance that an individual is counted in the second survey is unaffected by whether he or she was counted in the original census. This requirement is known as the independence assumption.

Together with his colleague Kenneth W. Wachter, Mr. Freedman criticized the Census Bureau’s methods for adjusting the census for the people it missed. They argued: “The census turns out to be remarkably good, despite the generally bad press reviews. Statistical adjustment is unlikely to improve the accuracy, because adjustment can easily put in more error than it takes out.”

Many prominent statisticians took the other side and favored adjusting the official count before the 2000 census and post-census survey were completed.

Mr. Freedman and Mr. Wachter argued, however, that the adjustment procedures could yield misleading results. For example, if people who moved between the original census and the post-census survey were not matched to their original addresses, the adjustment would be flawed.

More fundamentally, Mr. Freedman and Mr. Wachter noted problems with the key assumption that the chance of counting someone in the original census was independent of whether that person was counted in the post-census sample, and provided some evidence against the independence assumption. Mr. Freedman argued that demographic accounting — a way of counting the number of people in the country from birth, death and immigration records — provided a separate benchmark against which to judge the accuracy of the census. The Census Bureau relied heavily on demographic accounting in 2000 in deciding to forego the statistical adjustment.

I have to confess that before the 2000 census, I was skeptical of Mr. Freedman’s and Mr. Wachter’s criticisms of statistical adjustment, which sounded more like carping than fundamental objections to me. However, after the 2000 census was concluded — and it turned out to yield a more complete count than expected due to advertising, outreach and other innovative efforts by the Census Bureau — problems in the under-count adjustment were revealed. I became persuaded that Mr. Freedman and Mr. Wachter were prescient and right to highlight the practical problems inherent in implementing the under-count adjustment.

Now that the 2010 census is in need of urgent attention, I wonder what David Freedman would advise. Will the problems in the census be so severe that an under-count adjustment improves the count? Or would an adjustment make the count less accurate?

On a personal note, David Freedman was a friend and a champion. He could be a curmudgeon when it came to drawing conclusions about causality from actual data, but he always had the best intentions. His exacting standards raised the quality of research, and his advice and friendship will be missed. The day before he died, he returned an e-mail message to me with the note: “Alan, good to hear from you. Yes, I am pretty sick, life expectancy very, very short.”

Sadly, he was a good forecaster to the end.
Data Show Steady Drop in Americans on Move
NYTIMES
By SAM ROBERTS
December 21, 2008

Despite the nation’s reputation as a rootless society, only about one in 10 Americans moved in the last year — roughly half the proportion that changed residences as recently as four decades ago, census data show.

The monthly Current Population Survey found that fewer than 12 percent of Americans moved since 2007, a decline of nearly a full percentage point compared with the year before. In the 1950s and ’60s, the number of movers hovered near 20 percent.

The number has been declining steadily, and 12 percent is the lowest rate since the Census Bureau began counting people who move in 1940.

An analysis by the Pew Research Center attributes the decline to a number of factors, including the aging of the population (older people are less likely to change residences) and an increase in two-career couples.

The Pew analysis is drawn from census data and a survey, which found that 63 percent of Americans said they had moved to another community at least once in their lives, while 37 percent said they lived in the community where they were born.

According to the census’s American Community Survey, New York retained first place in the proportion of residents who were born in the state — more than 81 percent — with upstaters generally less mobile.

The top five also included Louisiana, Pennsylvania, Michigan and Ohio, generally Rust Belt states with older populations.

In contrast, fewer than 14 percent of Nevadans and 28 percent of Arizonans were born in those states.

Measuring the percentage of people born in a state who still live there, Texas ranked first, with nearly 76 percent, followed by North Carolina, Georgia, California and Wisconsin.

Alaska recorded the smallest share of people born in the state and still living there, 28 percent, followed by Wyoming, the Dakotas and Montana.

The telephone survey of 2,260 adults in October found that 57 percent had never moved outside their home state, while 15 percent had lived in four or more states.

About 23 percent say their current home is not where their heart is — typically because they were born someplace else, where they lived longer or their family still resides. About half who identify home as someplace else want to stay put; 40 percent say they would like to return.

Most people who do not move are kept close to home by family ties, the survey found, while most who do move are drawn by better jobs.

The Pew survey found that among all foreign-born adults, including recent arrivals, 38 percent describe home as their country of birth.

Among those who have lived in the United States 20 years or more, 76 percent describe America as home.
Estimate from State Data Center at UCONN...is this partially the result of lagging housing starts and sales of existing units?  What will the housing data forthcoming from Census 2010 show?
Growth stalls in the state
Stamford ADVOCATE
By Kate King, Special Correspondent
Article Launched: 07/11/2008 01:00:00 AM EDT

Fairfield County saw a small increase in population despite a drop statewide, according to census figures released yesterday.

Stamford, Greenwich, New Canaan, Darien and Westport all saw minor increases, according to the data. Norwalk posted a decline of 0.1 percent.  But the statewide picture isn't promising, experts say, pointing to a shrinking work force, loss of jobs, an aging population and a potential reduction in state representation in Washington, D.C.

"This population growth is consistent with our slow growth in the recent past," said Lisa Mercurio, director of the Business Council of Fairfield County. "New England as a whole has been growing more slowly than the rest of the U.S."

The population in Connecticut rose 0.19 percent over the last year, according to the census data.  Connecticut's population growth is the eighth lowest in the nation, according the report. Nevada had the highest growth rate since 2006 at 2.9 percent, and Rhode Island had the lowest at minus 0.36 percent.

Within Connecticut, Milford's population grew the most, by 532 people. Bridgeport showed the biggest population decline, losing 252 people over the past year.

Though 35 percent of the state's 169 towns declined in population, Fairfield County's population grew 0.1 percent. Stamford and Darien's population grew by 0.1 percent, and Greenwich and Westport grew by 0.4 percent. New Canaan's population increased by 0.5 percent.

Local leaders remained optimistic.  Despite the city's losses, Norwalk Mayor Richard Moccia said the city's redevelopment plans put Norwalk on a course toward growth.

"From a statistical point of view, in my mind it's virtually no loss whatsoever, so I'm not worried," he said. "I don't see this as a threat to our economic viability. I think with our new development projects in place, with more affordable housing going in down the road, you might see an increase in population."

Stamford Mayor Dannel Malloy painted an equally bright picture, estimating a 10-year growth rate of 6 percent for the city, coming from an increase in housing stock.  Malloy also suspected that the census counts miss some of the city's population.

"As much as I think these reports are interesting. . . . I respectfully would argue that it probably undercounts our immigrant population," Malloy said.

The slow statewide growth comes on the heels of a population boom, which lasted from 1995 to 2003, said Orlando Rodriguez, demographer and manager of the Connecticut State Data Center.  The population growth during those eight years was abnormal, a reaction to the end of a deep economic recession that took place in Connecticut from 1990 to 1995.

The population growth rate that Connecticut has experienced since 2004 is "more normal, looking forward, than what happened between 1995 and 2003," Rodriguez said.  But the return to normal of Connecticut's population growth rate isn't necessarily a good thing for the state.

"One of the things that's concerning us is that we're seeing a decline in population in urban areas, which is counter to what we had expected," he said.

A declining urban population means a smaller work force to replace the growing elderly population in the state, Rodriguez said.  Connecticut has one of the nation's oldest populations, meaning a high number of senior citizens.

Before the 2007 census, demographers had projected a job loss of 60,000 workers by the year 2030, he said. However, if urban areas continue to lose population, that worker loss will be even greater.  Also contributing to the slow growth rate in Connecticut is a net loss of population to other states.

"We send more people to other states than we get from other states," Rodriguez said. "You've got a lot of elderly people, not a lot of children being born, and folks leaving - it's lucky Connecticut has any population growth at all."

The only reason Connecticut has not dipped into negative population growth is the 15,000 foreign immigrants who have been coming to the state yearly since 2004.  In addition to contributing to a decline in the work force, slow population growth could cut Connecticut's representation in Congress.

"By 2020 we will lose a congressional seat," Rodriguez said. "For 2010, we'll probably be OK unless the bottom really falls out and Connecticut goes into major population loss."

Although Connecticut will most likely remain a five-district state for the next 12 years, uneven population growth within the districts will probably force a redrawing of the districts before then, he said.
"About Town" notes that census data counts the number of  people living in a residence--but it is always an under-counted number because people tend to not declare illegal apartments, more families living in an apartment than the lease permits, or illegal alians...
2010 Census: Who Should Count?
By MICHAEL REGAN | Courant Staff Writer
September 30, 2007

Border states in America's South and West are battlegrounds in the debate over illegal immigration, but when it's time to pass out seats in Congress, they are beneficiaries as well, a new study says.

Because of their large populations of undocumented residents, Texas and Arizona will each get one extra seat in the U.S. House of Representatives after the 2010 Census, the Connecticut State Data Center projects in a report being released today. California will keep two seats it otherwise would have lost.

Overall, the South and West each stand to gain five seats in the House, the center at the University of Connecticut says. If it weren't for their populations of illegal immigrants, each of these regions would gain only three.

The big loser in the reapportionment will be the Midwest, the center says. Five states in that region are projected to lose a total of six seats, four more than they would have if illegal immigrants were not included in the census tally.

Connecticut, which lost a seat in the last reapportionment, should keep the five it now has, but the Northeast as a whole will lose four - two in New York and one each in Massachusetts and Pennsylvania.

There's more than congressional clout at stake in the reapportionment: It also helps determine the makeup of the Electoral College. And the census itself influences everything from federal aid to the makeup of state legislatures. So as the 2010 Census approaches, attention is turning to the issue of whether it's fair to continue counting illegal immigrants.

Orlando J. Rodriguez, manager of the Connecticut State Data Center and author of the new report, considered that issue when designing the study. He figured the reapportionment two ways - one in which all residents are counted, as is currently done, and one in which illegal immigrants are factored out. Although politics watchers have been handicapping the 2010 reapportionment almost since 2000 was completed, Rodriguez said this is the first study he knows of to factor in the immigration question.

In part, the shift expected in 2010 is the result of a long-term population trend that has states in the South and West growing far faster than states in the Northeast and Midwest. In the 1960s, the Northeast and Midwest had 233 seats in the House, the South and West 202. The numbers roughly reversed two decades later, and now stand at 183 to 252. The new CSDC report projects that the South and West will have 262 seats to 173 for the Northeast and Midwest after 2010.

The winners and losers don't fall strictly along regional lines. New Jersey, for example, with the highest proportion of undocumented workers in the Northeast, would lose one seat if illegal residents were not counted, according to the CSDC projection. Montana would gain a seat if they weren't counted. Louisiana, in the wake of Hurricane Katrina, is expected to lose a seat regardless.

The new report suggests that the country's illegal immigrant population is playing an increasing role in congressional apportionment. After the 2000 Census, an analysis by the Center for Immigration Studies found that illegal immigrant populations affected the apportionment of four seats. The CSDC report projects that six seats will be affected by undocumented residents after 2010.

The projections are based on the most reliable data available, Rodriguez said, but studying the undocumented residents population is imprecise at best.

"Nobody really knows for sure," he said. "The bottom line is not `Is this specifically going to happen?' What I was trying to get across is, `Look at the impact [illegal immigration] is having.'"

Steven A. Camarota, director of research at the Center for Immigration Studies, said the impact is cause for concern. "You can make a strong case that there is a fundamental unfairness about this," Camarota said. "You do raise competing questions of fairness, justice, one man-one vote."

Counting illegal immigrants gives some voters disproportionate political clout. For example, Montana, which missed out on an additional seat after 2000 because of the weight of illegal immigrants elsewhere and is projected to fall short again after 2010, had almost 650,000 registered voters last November and one representative in Congress.

By contrast, California, which would lose two of its 53 seats after 2010 if illegal immigrants weren't counted, according to the projections, has four districts each with fewer than 200,000 voters registered. One district has fewer than 170,000 voters.

"You can win election [to Congress] in California with less than 50,000 votes," Camarota said.

But that's beside the point, said Arturo Vargas, executive director of the National Association of Latino Elected and Appointed Officials. The size of the electorate has nothing to do with representation in Congress.

Members of Congress "are elected to represent constituents. They don't just represent citizens," Vargas said. "They don't just represent the people who vote for them. They represent everybody in that congressional district."

Vargas said the framers of the Constitution drew distinctions among various classes of residents at various points. When it came to apportioning seats in Congress, he said, everyone was counted - although slaves were only counted as three-fifths of a person. "Would we go back to a time when we considered a person here to be less than human, less than a whole person?" he said.

At a time when illegal immigration in general is under heightened scrutiny, its connection to the census and reapportionment is likely to get renewed attention. One question that has already come up is how immigration enforcement might affect the count.

In 2000, the then-Immigration and Naturalization Service suspended raids before and after the census so as not to deter undocumented residents from responding. Earlier this year, when a census official raised the possibility of a similar freeze in 2010, a spokesman for Immigration and Customs Enforcement firmly ruled out the possibility.

Camarota and Vargas agreed that the question of what ICE does in 2010 depends on who is elected president in 2008.

"Under the unlikely circumstance that the Republicans win and they institute a comprehensive enforcement strategy, who knows?" Camarota said. "It could reduce the number of illegals significantly, and it could reduce the response rate."

But Vargas said the Constitution charges the government with counting everyone in the census.

"So the federal government needs to have some common sense about what its other agencies are doing that is going to compromise its constitutional duty to enumerate all persons," he said.

The other question is whether there will be renewed efforts to keep undocumented residents - or all noncitizens - out of the reapportionment count. Anti-illegal immigrant groups and states losing representation have been unsuccessful in court over the issue in the past, and Rep. Candice Miller of Michigan, which lost one seat after 2000 and is projected to lose another after 2010, has proposed a constitutional amendment to limit the reapportionment count to citizens.

Margo J. Anderson, a professor of history at the University of Wisconsin-Milwaukee and author of several books and papers on the census and reapportionment, said it's hardly a new question.

"It's an old issue. It goes back to 1790," she said. "Every time there is in some sense a political crisis in the country or a sectional dispute, the communities that think they're not going to gain from it take a hard look at it and wonder whether the rules are fair."
From the past...
Census: Connecticut Homes Got Bigger In '90s - Associated Press
September 03, 2001

HARTFORD Conn. (AP) - Soaring land prices and new zoning regulations led to a surge in the construction of large homes in Connecticut during the 1990s, according to homebuilders and a new survey by the U.S. Census Bureau.

Between 1990 and 2000, the number of large homes with nine or more rooms increased 14 percent while the number of medium-size houses with three to five rooms remained stagnant, the survey shows.

The data also showed a surprisingly rapid increase in the number of small homes with two rooms, which some housing experts attributed to the construction of homes with assisted-living services for older people.

Homebuilders say land prices and zoning regulations have forced them to build bigger houses.

"It's getting more and more difficult to basically make a reasonable profit building smaller homes," said Bill Ethier, executive vice president of the Home Builders Association of Connecticut. "If you have to spend $100,000 or $200,000 just to buy a lot, it's very difficult to put a modest-sized house on that lot."

State Rep. Jefferson B. Davis, D-Pomfret, co-chairman of the Legislature's Planning and Development Committee, agrees with homebuilders that local land regulations and local governments' heavy reliance on property taxes limit development to big houses on big lots.

With the bigger lots, people of different income levels are being prevented from sharing a sense of community, Davis said.

"It has struck me, as I listen to people around the state, that the issue of community is extremely important, whether it's a city or a suburb or a rural area," he said. "People have been expressing real hurt and a personal sense of loss as they see their view of community vanishing."

Still Mountain Estates in Bloomfield provides an example of the larger house sizes.

The 45 homes under construction will be as large as 3,200 square feet with 9-foot-high ceilings, interior balconies and oversize baseboard and trim. Prices start at $400,000, and small families are the targeted buyers.

There are plenty of other developments in Connecticut like Still Mountain Estates.

"I can go to the computer right now and I can pull off 50 listings over $1 million," said Lisa Sweeney, the William Raveis listing agent for Still Mountain Estates. "I could never have done that 10 years ago."

Although some economists say Connecticut is nearing a recession, real estate agents say it's hard to believe because of the grandeur of new homes.

"Houses are getting bigger," said Hope Firestone, a veteran agent in the West Hartford office of William Raveis. "The houses are getting nicer. Kitchens are having every appliance including convection ovens in the upper price ranges. Garages are at least two cars, with many, many threes and more."

AP-ES-09-03-01 0130EDT