WHEN IMMIGRANTS ARE NOT MIGRANTS - COUNTING ARRIVALS OF THE FOREIGN-BORN USING THE US CENSUS

This paper compares characteristics of recent immigrant arrivals in th e United States using two measures from the decennial U.S. census: the came-to-stay question and the migration question. We show that a litt le under 30 percent of immigrants who reported they came to stay betwe en 1985-1990 on the 1990 U.S. Census Public Use Micro Sample were resi dent in the United States on April 1, 1985. A similar analysis of the 1980 census reveals that 22 percent of immigrants who reported they ca me to stay between 1975-1980 lived in the United States on April 1, 19 75. Thus among recent arrivals, defined as those who reported they cam e to stay in the quinquennium preceding the census, a large number wer e resident in the United States five years before the census date. Fur thermore, the proportion of recent arrivals present in the United Stat es five years before the census increased between 1975-1980 and 1985-1 990. We show that the profile of recent arrivals is sensitive to their migration status. Generally, in both the 1975-1980 and 1985-1990 coho rts, those resident in the United States five years before the census have significantly less schooling and lower incomes than those who wer e abroad. Accordingly, we argue that estimates of the skill levels and hourly wages of recent arrivals to the United States vary with the wa y arrival is measured. Researchers who rely on Public Use samples of t he U.S. census for their data should be aware that the year of entry q uestion implies a broader definition of arrival than the migration que stion. We caution that immigration researchers should consider the ide a of arrival more carefully to help distinguish newcomers from the res ident foreign born.