Americans Residing Overseas
As described at the outset of this report, our panel’s charge pointedly excludes consideration of who should or should not be counted in the census (see Section 1–A). The issue of whether noncitizens (and particularly illegal immigrants) should be included in census counts (as they currently are) is the source of great political debate and has periodically drawn attention from Congress and other policy makers, but it is one that is out of our panel’s scope. Likewise, the current practice of excluding American citizens living overseas, who are not military personnel or employees of the federal government, is one on which we do not offer advice.
However, while we do not comment on the appropriateness of counting Americans living outside the country in the census, some comment on the way this group has been handled in previous censuses and in census tests is in order. Several of the residence rules in the 2000 census deal directly with this group, and the handling of a subset of Americans living abroad was the focus of the first legal challenge to the 2000 census count. Americans living outside the United States are a natural and important part of any listing of situations for which residence standards may be difficult, and so the issues are important to consider in residence rules. We describe procedures for counting the military population in Section 3–F; in this appendix, we briefly outline the issues associated with counting American citizens living abroad in the census.
Following the same procedures used in the 1970 and 1990 censuses (as described below), the 2000 census included military and federal government personnel stationed overseas in the census apportionment counts but not in other products, including redistricting totals. These people were allocated to
their home states based on “home of record” information from the Department of Defense or the employees’ parent agency (e.g., the Department of State).1 This overseas population count tallied 576,367 military and federal personnel, just over 59,000 of whom were allocated to California and 52,000 to Texas. Other states with large military installations gained from the inclusion of these residents, such as the 22,187 added to Virginia’s total. As would be significant shortly after the 2000 census, North Carolina—with its large naval facilities—was credited with 18,360 overseas residents while Utah received 3,545.
Estimates suggest that the overseas military and federal employees (and their dependents) represent a small share of the total American citizenry living in other countries. As of July 1999, the U.S. State Department’s Bureau of Consular Affairs estimated that 4,163,810 private American citizens lived in foreign nations (not counting military and government personnel and their dependents).2 Private American citizens living in Mexico account for 25 percent of that total, and Canada another 17 percent, by far the largest single-country contingents. Regionally, 28 percent of private American citizens overseas reside in Europe, 12 percent in Asia and Oceania, 7 percent in the Middle East, and 2 percent in Africa. Other estimates put the number of Americans outside the country at up to 10 million.
TREATMENT IN PAST CENSUSES
The 1830 and 1840 censuses were the first to include the counts of some segment of the American overseas population when they included U.S. naval crews in the counts (Mills, 1993:3); otherwise, consideration of the overseas population is absent from both census tabulations and enumerator instructions until the beginning of the 20th century.3 The 1900 census counted 91,219 Americans living abroad, consisting of military and federal civilian personnel at military stations and their dependents, along with the crews of naval vessels. Census authorities noted that these were counted and printed in the census volume because, “except for conditions arising out of the Spanish-American War, they would have been found residing largely within the limits of the United States proper, and they are, therefore, included properly in the
This included 2,037 persons allocated to the District of Columbia; the District’s population is excluded from the subsequent calculation of seats in the House of Representatives.
The 4.1 million total is shown on versions of the State Department data such as that published by American Citizens Abroad (http://www.aca.ch/amabroad.pdf [10/5/05]) and cited by U.S. Government Accountability Office (2004a). An archived version of the Bureau of Consular Affairs tabulation at http://www.pueblo.gsa.gov/cic_text/state/amcit_numbers.html [10/5/05] records the total as 3,784,693.
The 1870 and 1880 census enumerator instructions did include the somewhat vague rule that “seafaring men are to be reported at their land homes, no matter how long they may have been absent, if they are supposed to be still alive,” without specifying naval vessels (Mills, 1993:10).
total for 1900 for purposes of comparison with the population returned at preceding censuses” (Mills, 1993:11). The enumerator instructions for the 1900 census include the more generic directive that “persons temporarily residing in foreign lands”—seemingly casting eligiblity wider than military and federal employees—“should be enumerated as part of their family at their home or usual place of abode” (Gauthier, 2002:35); however, the published statistics include only those on naval vessels and in territories associated with the Spanish-American War, principally the Philippines (Mills, 1993:11).
The 1910 census reversed that determination and included federal civilian employees in a separate overseas population count, an approach that continued through 2000 (McMillen, 2000a:32). Persons stationed abroad continued to be counted with the cooperation of the War and Navy Departments. Curiously, the 1910 enumerator instructions for the domestic census count included a clear directive to permit the reporting of American civilians who were living abroad “temporarily” (Mills, 1993:18):
Citizens abroad—Any citizen of the United States who is a member of a family living in your district, but abroad temporarily at the time of the enumeration, should be enumerated as of your district. It does not matter how long the absence abroad is continued, provided the person intends to return to the United States.
No inference was made on possible duplication between people included in the overseas count and reported in the stateside enumeration.
The exact means by which data were collected, and the groups covered, varied slightly over the next few decades. Both the 1900 and 1910 censuses used enumerator schedules identical to the general population schedule to gather information on military and naval personnel, federal civilian employees, and their dependents. The 1920 census saw the switch to a different basis schedule for these groups along with a modest expansion of scope—“persons in the service of the American Red Cross or U.S. consular service” were added to the target population (Mills, 1993:8). The 1920–1940 enumerator instructions repeated the “citizens abroad” advisory from 1910.
The 1950 census was the first in which the Census Bureau entered into cooperative agreements with the U.S. Departments of Defense and State, as well as the Maritime Administration, to provide information on personnel and dependents under their authority. A Bureau summary of innovations for 1950 suggested that, prior to 1950, “no systematic effort [was] made to enumerate American citizens abroad”; “since their numbers were small, this omission made little practical difference.” The situation changed after World War II, though: “now that we have hundreds of thousands of our people overseas in occupation and other military duty and on the staffs of federal civilian agencies, it has become important to have an inventory of them” (Shryock, 1950:8). The same Bureau summary judged that “it is still not feasible to
enumerate all those overseas who claim American citizenship,” and McMillen (2000a) found that the 1950 census included a limited effort to collect data on “other citizens” living abroad. Mills (1993:32) comments that “scattered” forms “were received from private U.S. citizens living abroad” who picked up a form from embassies or consulates, but the response was said to be low.
The Bureau also sought and received an opinion from the U.S. Attorney General in 1949 on the legality of its plans for enumerating Americans overseas; the Attorney General favorably approved the notion of enumerating them, but not including them in the total population figures for any state or in the total population of the continental United States. Thus, the overseas population counts were intended for use as information, but not for apportionment.
The 1960 census repeated the same procedures for the military and federal employee enumeration, including an emphasis on direct distribution of individual census report forms. The 1960 census did mark “a special effort …to obtain voluntary reports from private U.S. citizens not affiliated with the Federal Government who were abroad for an extended period” (Mills, 1993:3). The Department of State agreed to play a larger role in disseminating questionnaires from its embassies and consulates, and some of those foreign stations used local newspapers to generate word of the census operations. “Religious groups with missionaries abroad, as well as large corporations with overseas employees,” were also advised of the procedures and included notices in their periodicals and newsletters (Mills, 1993:37). Again, though, “the reporting of these private U.S. citizens was not complete” and “understated the true number of these persons abroad” (Mills, 1993:3). The effort did result in a separate census report on the social and economic characteristics of the population abroad, the first of its kind (Mills, 1993:36).
The 1970 census marked a major change in the use of counts of the overseas population. Specifically, military personnel and federal civilian employees stationed overseas were included in the state-level census totals for purposes of apportionment. Two major reasons were cited for this change: the larger number of persons expected to be overseas at the time of the census because of the Vietnam War, and the then-recent court decisions on “one person, one vote.” Subcommittees in both houses of Congress held hearings on the issue in 1969, and they favored the inclusion of the military and federal civilian personnel in apportionment counts; advice was again sought from the U.S. Justice Department, and then-Assistant Attorney General William Rehnquist concluded that the addition was within the powers of the census director (Mills, 1993:4). In terms of operations, the State Department handled the enumeration of civilian employees of the federal government, their dependents, and other Americans residing abroad through contacts at embassies and consulates. Much of the overseas military count was obtained from administrative records rather than direct questionnaire enumeration. The enu-
meration of “other citizens” was again conducted on a voluntary basis and the results—as in previous tentative tests—showed low reporting. Mills (1993:46) suggests that “a post-1970 census comparison of data from the 1970 overseas census with country of birth/citizenship data from individual foreign censuses indicated that the census counts for private U.S. citizens represented a substantial undercount, particularly in Canada and Mexico, where the underenumeration probably exceeded 90 percent.”
The Census Bureau reversed itself in the 1980 census, electing to exclude all overseas population from the apportionment counts. The then substantially smaller deployment of military personnel overseas was the major reason for the reversal; the Census Bureau also expressed concerns on the reliability of data on a person’s “home state.” The 1980 census was also distinctive in that no attempt was made to directly enumerate any part of the overseas population; instead, administrative record counts (but no data on characteristics) for armed forces, civilian employees, and their overseas dependents were obtained from the Department of Defense and the Office of Personnel Management and tallied as a separate “overseas population” count. Citing low participation in the 1960 and 1970 efforts, no data or counts of “other citizens” not affiliated with the federal government were collected. Zitter (1987) argued that the Census Bureau’s attempt to count the overseas population in 1980 was minimal in comparison with earlier censuses; anything that distracted from the main census event was deemed an unnecessary risk and was not to be undertaken without a compelling reason, such as a congressional mandate.
THE 1990 CENSUS
For the 1990 census, overseas military and federal employees were again counted by administrative records. The overseas military population and their dependents living with them were counted by using Department of Defense records; federal employees and their dependents were similarly addressed using Office of Personnel Management records. No effort was made to count “other citizens” living abroad.
However, the Bureau reversed itself on including military and federal civilian employees in apportionment totals, returning to the 1970 policy of including them in their home states’ counts. The change owed a great deal to increased congressional interest in the issue, which grew as both the overseas Americans issue and the inclusion of undocumented immigrants emerged as major issues and potential sources of litigation (McMillen, 2000a).
In a background paper to inform the panel’s deliberations, Lowenthal (2005) summarized the congressional interest, despite the absence of large numbers of American troops abroad:
In April 1988, the House Subcommittee on Census and Population held a hearing to review Census Bureau policy with respect to members of the armed forces and federal employees, and their dependents, stationed overseas during the census. The Census Bureau and the Department of Defense (DoD) announced that they had agreed to administer a census on overseas military bases, but Census Bureau Director John Keane expressed doubt that his agency could allocate this population back to their states of usual residence within the legal timeframe for reporting apportionment counts to the President.
[Subcommittee chair Mervyn Dymally (D-CA)], along with the census subcommittee’s senior Republican member, Rep. Constance Morella (R-MD), subsequently sponsored legislation to include military personnel and federal civilian employees stationed overseas during the census in the state population totals used for apportionment (H.R. 4720). While the Census Bureau expressed reservations about the bill, the Committee on Post Office and Civil Service approved the measure by voice vote, and House leaders scheduled the bill for consideration under Suspension of the Rules, a procedure used almost exclusively for non-controversial legislation because it requires a two-thirds vote for passage. Surprisingly, though, in September 1988, H.R. 4720 was defeated by a recorded vote of 93–317, with debate focused on the criteria for assigning overseas military and federal personnel to a home state [Congressional Record, September 28, 1988, Roll Call No. 361.]. Many members, concerned that states with large military facilities would benefit unfairly, objected to counting overseas personnel in the state where they last resided for six months before being deployed or assigned to a foreign post. (Some Members also suggested that Congress should, at the same time, exclude undocumented residents from the apportionment counts.)
Despite the defeat of H.R. 4720, congressional support for including military and federal civilian personnel in the apportionment base remained high in the 100th and 101st Congresses. Additional bills were introduced in both the House and Senate in the 101st Congress, with various proposals for determining the state of residence of this population for apportionment purposes. Rep. Dymally introduced another bill (H.R. 3016) that set the place of residence for this population as “the individual’s last place of general abode within the United States for a period of 6 consecutive months or longer.” Sen. John Heinz (R-PA) sponsored a bill (S. 290) that would have left the decision on where to count overseas military and federal civilian personnel to the discretion of the Secretary of Commerce. Rep. [William Goodling (R-PA)] suggested a combination of criteria to establish usual residence, proposing to count overseas military personnel at their “home of record” and overseas government employees in the state where they had last lived for 12 consecutive months or longer (H.R. 1468).
In August 1989, with the start of the decennial census fast approaching, the Department of Commerce announced that it had decided to in-
clude military personnel and federal civilian employees assigned to overseas posts, and their dependents living with them, in the state population totals used for apportionment. At the time, the Defense and Commerce Departments were still preparing to administer a modified census on overseas military bases, to collect the necessary information on members of the armed forces (who comprised the vast majority of federal personnel stationed abroad). In December, however, DoD concluded that it could not administer an overseas census, which would have provided an opportunity for self-enumeration. Instead, it agreed to provide state-by-state counts of military personnel stationed abroad using information from administrative records. According to the Census Bureau, DoD identified three possible criteria, obtainable from administrative files, for assigning overseas military personnel to a home state: home of record; legal residence (state declared for income tax purposes); and last duty station (facility where armed forces member was assigned before deploying overseas) (Mills, 1993:5).
Concerned about the consequences of the various alternative methods, senior members from the House census authorizing committee wrote to the Census Bureau, urging that overseas military personnel be counted at their home of record. Nevertheless, with the 1990 census well underway, the Secretary of Commerce determined that these individuals would be allocated to the state where they last resided for six consecutive months or longer before deployment overseas, even though the House of Representatives had rejected just such a proposal (H.R. 4720) in the previous Congress.
Clearly displeased that the Commerce Department did not consult Congress before making this decision, Rep. [Thomas Ridge (R-PA)] and Rep. Thomas C. Sawyer, now the ranking minority member and chairman, respectively, of the Subcommittee on Census and Population, introduced legislation in May 1990, requiring the inclusion of overseas military and federal personnel, and their dependents, at their home of record (H.R. 4903) [Statement of Rep. Thomas C. Sawyer, Congressional Record, May 24, 1990, Extension of Remarks]. The House of Representatives passed the bill by voice vote in June, but the Senate did not take any further action. The clear signal sent by the House, however, with its primary interest in the outcome of apportionment, was finally heard by the Census Bureau, which concluded in July that “home of record” for military personnel “was the most consistent with the concept of ‘usual residence’ in the census (Mills, 1993:5). The Commerce Department subsequently adopted “home of record” as the basis for assigning overseas military personnel and their dependents to a state for purposes of apportionment.
When the apportionment counts and home-of-record overseas allocations were released in December 1990, the overseas figures appeared to sway the placement of a congressional seat. If the overseas counts were not included, the state of Massachusetts would have retained its eleventh seat in the House
of Representatives; instead, the state of Washington gained its ninth seat. This occasioned Massachusetts’ May 1991 legal challenge to the policy of including overseas personnel in the apportionment totals; we summarize the resulting case—Massachusetts v. Franklin—and the “enduring tie” standard it suggested in Box 2-5.
THE 2000 CENSUS
The 2000 rules for overseas populations followed those from 1990. As in 1990, the 435th and final seat in the House of Representatives was won by a small population total, and the overseas count became a focus of litigation.
Prior to the 2000 census, organizations representing Americans living abroad pressed for congressional interest in the issue. The House Subcommittee on the Census held a 1999 hearing to review various legislative proposals to include nongovernmental overseas Americans in the count that had been introduced in both chambers of Congress.4 At the hearing, which also considered proposals to change the Bureau’s policy on counting prisoners at the prison location (see Chapter 3), Census Bureau director Kenneth Prewitt argued that it was far too late to consider enumerating private American citizens in the 2000 census and reiterated basic concerns about conducting the count (see below).
Lowenthal (2005) writes that “while they did not seriously consider last minute proposals to extend Census 2000 coverage abroad, legislators did pursue the idea with an eye toward the future”:
In 1999, in their reports on the Fiscal Year 2000 Commerce Department funding bill, both House and Senate appropriators directed the Census Bureau to develop a plan for counting overseas Americans in the census and to inform Congress of its progress.5 The following year, appropriators issued a more specific directive to the Census Bureau for a report on methodological, operational, and policy issues associated with including all American citizens living abroad in the census, as well as an estimate of the number of Americans living or working overseas, both for the federal
government and for private organizations.6 Rep. Carolyn Maloney continued to push the issue forward, as well, sponsoring bills to authorize and pay for an interim census of Americans abroad in 2003 and to begin planning for the inclusion of all Americans living overseas in the 2010 census (H.R. 3649 and H.R. 4568).
When the apportionment counts were released in December 2000, North Carolina edged Utah for the 435th House seat by fewer than 1,000 people. Utah brought suit, charging that the Census Bureau’s treatment of Americans living abroad was unfair in that military personnel and other federal employees stationed overseas are included in the apportionment counts but other civilians are not. Specifically, Utah challenged the exclusion of approximately 11,000 missionaries of the Church of Jesus Christ of Latter-Day Saints: these missionaries may live abroad temporarily, and many of them would ordinarily be counted in Utah. For uniformity, Utah held that federal employees stationed overseas should be dropped from the apportionment counts. The U.S. District Court for Utah ruled against the state in April 2001, a ruling affirmed without comment by the U.S. Supreme Court later that year.7
THE 2004 OVERSEAS CENSUS TEST
Following the Utah challenge, as well as the Bureau’s required report to congressional appropriators on issues of counting Americans overseas, plans developed for an overseas enumeration test in July 2004. For the 2004 test, the Bureau chose France, Kuwait, and Mexico as test sites. Following procedures like those used in the 1960 and 1970 censuses, the test relied on citizens who lived abroad to obtain a form from a U.S. embassy or consular office. The Bureau mounted some publicity efforts in English-language newspapers and media in the test countries. In all, the Bureau estimated that the planning, execution, and analysis of the test cost $7.8 million.
The Government Accountability Office issued two reports on the overseas test, one noting general concern about the direction and utility of the test (U.S. General Accounting Office, 2004) and the other analyzing the test’s results (U.S. Government Accountability Office, 2004a). The reports concluded that the test suffered from design shortcomings that fundamentally
limited its capacity to answer key research questions about the feasibility and cost of gathering data on Americans living abroad; they also questioned the resulting data quality. The response rates on the test were extremely poor: the Bureau had printed 520,000 questionnaires for the test, yet only 5,390 questionnaires (1,783 paper and 3,607 on the Internet) were returned from all three sites. By comparison, the July 1999 State Department estimates suggested that 1,036,300 American citizens reside in Mexico, 101,750 in France, and 7,710 in Kuwait.
In carrying out the test, the Bureau experienced some country-specific problems that would likely be more significant if the enumeration were conducted in additional countries. Perhaps most significant was the problem experienced with collecting even short-form census information in France; under French privacy laws, collection of data on race and ethnicity is generally prohibited. The Bureau also had difficulty overseeing the contractor responsible for raising public awareness of the test. The U.S. Government Accountability Office (2004a) commented that the approach used to test the viability of counting this group by means of a voluntary survey that relied on marketing to ensure a complete count would be very costly and yield poor results.
At the outset of planning efforts for the 2004 overseas test, a follow-up test in 2006 had been scheduled. However, on the basis of the 2004 test results, funds for the 2006 test were not provided in the Bureau’s appropriation.
CONCEPTS IN COUNTING AMERICAN CIVILIANS OVERSEAS
The U.S. General Accounting Office (2004:8) usefully summarizes the basic “logistical, conceptual, policy, and other questions that surround the counting of overseas Americans.” Similar issues were also raised by then-Census director Kenneth Prewitt in his testimony on the matter prior to the 2000 census (U.S. House of Representatives, Committee on Government Reform, 2000:44–46):
Who should be counted? U.S. citizens only? Foreign-born spouses? Children born overseas? Dual citizens? American citizens who have no intention of ever returning to the United States? Naturalized citizens?
How should overseas Americans be assigned to individual states? For certain purposes, such as apportioning Congress, the Bureau would need to assign overseas Americans to a particular state. Should one’s state be determined by the state claimed for income tax purposes? Where one is registered to vote? Last state of residence before going overseas? These and other options all have limitations that would need to be addressed.
How should the population data be used? To apportion Congress? To redistrict Congress? To allocate federal funds? To provide a count of overseas Americans only for general informational purposes? The answers to these questions have significant implications for the level of precision needed for the data and ultimately, the enumeration methodology.
How can the Bureau verify U.S. citizenship? Administrative records such as passports and Social Security data have limitations. For example, Americans can reside in Mexico and Canada without a passport and many Americans overseas do not have Social Security numbers, especially dependents.
How can the Bureau ensure a complete count without a master address list? The foundation of the stateside decennial census is a master address list. Because the list is essentially the universe of all known living quarters in the United States, the Bureau uses it to deliver questionnaires, follow up with nonrespondents, determine vacancies, and determine individuals the Bureau may have missed or counted more than once. The Bureau lacks a complete and accurate address list of overseas Americans. Consequently, these operations would be impossible and the quality of the data would suffer as a result.
Can administrative records be used to help locate and count overseas Americans? Administrative records such as passport and visa files, voter registration forms, as well as records held by private companies and organizations have the potential to help the Bureau enumerate Americans abroad. However, the accuracy of these records, the Bureau’s ability to access them, confidentiality issues, and the possibility of duplication all remain open questions.
Do certain countries have requirements that could restrict the Bureau’s ability to conduct a count? According to the Bureau, in planning the overseas test, the Bureau was informed that French privacy laws prohibit asking about race and ethnicity, two questions that are included on the U.S. census questionnaire. Although the Bureau worked with French officials to address this problem, the extent to which the Bureau will encounter restrictions in other countries, or whether other countries will cooperate with the Bureau at all, is unknown.
Still, the issue of counting all Americans overseas has been contentious in the past two decennial censuses, and the directions of modern business—and the degree to which advances in transportation and technology continue to make the world seem a smaller place—suggest that the issue will endure. Students continue to take advantage of opportunities to study abroad; jobs in foreign countries remain attractive possibilities, and the nature of modern
businesses can lead to employee’s living in foreign sites for weeks or more. Generally, questions may be raised as to why federal government employees should be counted but not American employees of multinational companies.
Mills (1993:1) soundly notes a “major observation” that arises from reviewing the history of overseas Americans in the census—namely, “the lack of a single conceptual thread running through the censuses concerning how Americans abroad fit into the overall decennial enumeration. It was partly this absence that led to the inconsistencies—evident in this report—in census treatment of Americans overseas.” The voluntary survey approach tried by the Bureau in 1960 and 1970—and retested in 2004—seems unlikely to produce satisfactory results. Much as is the case for prisoners and other domestic institutional populations, the quality of enumeration of overseas Americans will likely depend on the use and reliability of administrative records. In addition to the records maintained by the Defense and State Departments, one might also need to consider the employee rosters maintained by companies and religious organizations, as well as other government sources (e.g., Social Security Administration rosters of American retirees receiving checks in other nations). A useful test would involve assembling these data sources (under appropriate confidentiality agreements) and studying the resulting picture of overseas Americans, possibly comparing the results with survey measures.