Woodrow, Karen A., Jeffrey S. Passel, with the assistance of J. Gregory Robinson. 1989. “Estimates of Emigration Based on Sample Survey Data from Resident Relatives,” Prepared for the Office of Management and Budget, Office of Information and Regulatory Affairs, U.S. Census Bureau. Executive Summary.


Immigration has been important throughout the history of the United States, but in recent years it has become even more so. A full assessment of immigration’s impact on U.S. society depends on accurate measurement of the phenomenon. Net immigration consists of three major components—legal immigration, emigration, and undocumented immigration. Legal immigration is measured accurately with administrative statistics. Estimating the scale of undocumented immigration is now feasible because of recent advances principally by researchers at the Census Bureau.

Measurement of emigration has remained as the stepchild of the immigration statistical system, even though data for 1907-1980 show that emigration amounted to roughly 30 percent of the level of legal immigration. Unfortunately, the various techniques that have been used to estimate emigration in the past are no longer applicable for the 1980s and beyond. This report describes the results of the Census Bureau’s research into the use of innovative survey techniques for measuring emigration, with a particular emphasis on recent emigration (i.e., during the 1980s).

Measurement of emigration with conventional sample survey procedures faces one seemingly impossible obstacle—the emigrants are no longer in the United States to be interviewed. Network or multiplicity sampling attempts to overcome this problem by allowing each respondent to report an event occurring to another person. The reporting is based on a counting rule that specifies the eligible linkages of the survey respondent to the event or person. Multiplicity sampling has considerable potential for providing a detailed description of the size and characteristics of the emigrant population.

In the July 1987 Current Population Survey (CPS), all persons aged 14 years and older were asked if any of their immediate relatives (parents, siblings, and children) resided outside the United States. If a relative living outside the United States was identified, the respondent was asked if that relative had ever lived in the United States. The groups of immediate relatives identified who are living outside the United States, but have lived in the United States, provide a universe that includes emigrants. The universe consists of potential emigrants and Americans living overseas temporarily, including members of the Armed Forces and their dependents.

To derive estimates for emigrants and groups of Americans living overseas, it is necessary to weight these reports of potential emigrants and Americans living overseas for the multiplicity of linkages in the survey universe (that is, the number of resident relatives who could report them). The survey estimate of Americans living overseas temporarily in 1987 is 465,000. The largest group is the 303,000 estimated members of the Armed Forces overseas.

The survey estimate of emigrants is 926,000. For nearly one-half (443,000), the major activity is “working.” More than two-thirds of the emigrants (621,000) were between ages 20 and 55. According to the basic survey data, an estimated 515,000 persons left the United States to live in other countries during the period 1980 to 1987; more emigrants had departed during this period than any other corresponding period. This pattern is not unexpected as the multiplicity approach should yield better estimates for the most recent period than for earlier periods. Emigrants during earlier periods are less likely to have survived and less likely to have resident relatives to report them.

The July 1987 CPS questionnaire also included questions as to whether children accompanied the emigrant relative (or the American living overseas). The estimated number of children accompanying emigrant parents is 273,000 according to the survey. For Americans living overseas, the estimate is 157,000 secondary emigrant children. Combining the estimates of children who are secondary emigrants with the estimates of primary emigrants raises the overall survey estimate of emigration by 29 percent from 926,000 to 1,199,000. For Americans living overseas, inclusion of emigrant children substantially increases the survey estimate by 34 percent from 465,000 to 622,000.

These estimates should not necessarily be viewed as “final” estimates. Comparison of the survey estimates with administrative data for Americans living overseas and with independent estimates of the survived emigrant population based on past trends shows that the multiplicity-derived estimates understate “true” emigration levels. For example, the survey estimates of the military population and Federal affiliated civilians living overseas are each about 60 percent as large as the administrative counts. The report discusses several reasons why the multiplicity estimates of emigration are low. However, the administrative data provide a confident assessment of undercoverage in the CPS for Americans living overseas that can be used as a basis for adjusting the estimates of overall emigration.