Census Bureau finds it can't count Americans abroad

The process of counting all the people who live in a democracy is not simply a math problem — it’s downright political.

The process of counting all the people who live in a democracy is not simply a math problem — it’s downright political.

The 1990 and 2000 U.S. census prompted court challenges by states that lost population, and therefore representatives in Congress. New census figures, released every 10 years, routinely prompt battles in state legislatures over reapportionment, like the struggle in Texas that eventually led to the indictment of Rep. Tom DeLay (R-Texas), which cost him his post as majority leader.

Among the thorniest but least-known issues in the debate is whether it is feasible to include Americans who live overseas, a demographic represented by a surprising number of lobbying groups.


After spending nearly $8 million and three years to study that question, the U.S. Census Bureau has an answer: No. 

The bureau, under pressure from Congress, tested its ability to find expatriates in three countries — Mexico, France and Kuwait — chosen because they were thought to have sizable American expatriate communities.

Federal workers and members of the military are already included in census numbers because it is relatively simple to figure out where they are. It’s trickier though to get at the people without ties to the government. An estimated 4 million Americans live overseas, although experts believe that to be a very rough estimate.

Stakeholder groups such as American Citizens Abroad, the Association of American Residents Overseas, Democrats Abroad and Republicans Abroad have estimated that more than 1 million Americans live in the three countries chosen by the bureau for the test. The groups have pressed Congress to force the bureau to change its policy of not including in the census numbers those Americans who live overseas but do not work for the government.

“In an increasingly global economy, we have a lot of Americans who are living overseas,” said Kathleen Styles, project manager for the Census Bureau overseas test.

“These people feel like they are ambassadors for their country.”

Official State Department figures of expatriate populations in the three countries are much lower than the stakeholder groups’ figures. But even small shifts in population figures can change the size of a state’s congressional delegation.

In January 2001, Utah sued the Commerce Department, where the Census Bureau is housed, saying that it lost a congressional seat because the 2000 census failed to account for 11,000 Mormon missionaries living overseas.

The state lost its case, despite a Congressional Research Service finding that it would have retained the seat if a mere 855 more people had been counted in its column, according to a recent Government Accountability Office (GAO) report on the debate.

The Census Bureau “traditionally gets sued over this,” Styles said, given the small margin for error. In 1991, it was Massachusetts mounting a court challenge. It also lost.

Census officials knew they faced challenges in trying to accommodate the push to count overseas Americans.

In counting citizens who live in America, the bureau has the benefit of a massive address file that includes all known housing units. Styles calls the database a “control” by which to judge the census responses that the bureau receives. The bureau is now using the global positioning system to map the information in the database for even more accurate estimates.

“There is no central registry” of expats, Styles said, “and we don’t have any good way to reach out to them.”

Privacy laws in France and delivery troubles in Kuwait added to the difficulties bureau testers faced.

To make up for a lack of a database, the bureau spent $200,000 on an ad campaigns in Mexico and France to notify Americans living in those countries of its intentions.

But as the GAO noted in its report to Congress, “the response levels fell far short of what the Bureau planned for.”

The Census Bureau printed 520,000 census forms, but received only 1,783 back by the end of the test. Another 3,607 responses were received over the Internet.

It cost the government $1,450 per response, far more than the $56 it costs to count an American living domestically, the GAO reported.

Styles said it would cost so much in both time and money to overcome the challenges of counting overseas Americans that the American count would inevitably suffer. The GAO told Congress that it agreed and advised lawmakers to drop the idea.