Three generations in the New World: labour market outcomes of Swedish Americans in the USA, 1880–2000

Abstract

Roughly one million Swedes emigrated to the USA between 1870 and 1920. These emigrants gave rise to over four million individuals with a Swedish ancestry currently living in the USA. This study shows that Swedes have maintained the geographic concentrations chosen by their parents and grandparents. Also, Swedish migrants were on average less educated than the US-born white population, yet their descendants tended to be relatively better educated. The statistical analysis of the data shows that Swedish-born men had lower, and women much lower, earnings than the native born. However, this position was not carried over to the subsequent generations who performed similar to natives with similar characteristics.

Keywords:

• migration
• Sweden
• labour market
• integration

Notes

¹See Beijbom, Amerika (1977), Beijbom, Utvandrarna (1986), Beijbom, Svenskamerikanskt (1990), Runblom/Norman, Sweden (1976) and the pioneering work of Thomas, Social (1941) for examples.

²Although Borjas did examine certain aspects of this group. See Borjas, ‘Intergenerational Mobility’ (1993), 113–35 and Borjas, ‘Long-Run Convergence’ (1994), 553–73.

³Chiswick, ‘Effect’ (1978) ; Mincer, Schooling (1974).

⁴For our uses, we should note that the 1910 IPUMS consists of a 1.5% sample, and the 1930 IPUMS consists of a 0.5% sample. All others in this study are 1% samples. All figures presented have been corrected for these differences in sample size.

⁵For a more complete discussion of the origins and contents of the IPUMS, please see http://usa.ipums.org/usa/intro.shtml.

⁶This group is defined from the questions on place of birth (the USA) and race (White) available in all censuses.

⁷The group of all immigrants is defined from the question on place of birth (not being Sweden or the USA), while the group of all second generation is defined from the question on place of birth (the USA) and from the question on parental place of birth (not being Sweden or the USA for at least one parent).

⁸Borjas, ‘Long-Run Convergence’ (1994).

⁹Ravenstein, ‘Laws’ (1889).

¹⁰Of course, we do not actually measure ‘homogamy’ here, since we only observe couples with children, and not all marriages. But we feel that our measurement should yield a satisfactory approximation, especially during the early periods of this study.

¹¹See for instance, Rooth/Ekberg ‘Unemployment’ (2003).

¹²This high rate of homogamy is largely the result of post-migration marriage, and not emigration of previously married couples. With the exception of the earliest period of the Great Migration, Swedes tended to come as single individuals, and not as families.

¹³For instance, 26% is achieved at by adding the percentage of those with a Swedish mother and an American father to the percentage of those with an American mother and a Swedish father.

¹⁴The response rate to the mixed alternatives was not exceptionally high, with a total of only 192,000 individuals choosing these alternatives. Under the worst-case assumption where Swedes comprised one-third of the responses in each group we would lose information on 63,000 individuals. This number is not insignificant, but given that four million individuals reported ‘Swedish’ as ancestry, it is not felt that this loss impacts the results.

¹⁵Appendix Table A1 gives a listing of the states comprising the various regions.

¹⁶Compare Table 4 with Appendix Table A2 for a comparison between the residential distribution of Swedish immigrants and the general American population.

¹⁷Prior to 1940, IPUMS only records actual school enrolment and literacy.

¹⁸We make the age restriction, that is, including only economically active ages, to comply with the analysis of labour market outcomes in the following sections.

¹⁹For a discussion and application of this theory on the migration flow between Finland and Sweden see Rooth/Saarela, ‘Selection’ (2007).

²⁰The results for the other variables are available from the authors upon request.

²¹For a detailed discussion, see www.ipums.org.

²²See also Table A4 in the Appendix for the descriptive statistics.

²³The exact definition in the IPUMS variable description is: ‘OCCSCORE is a constructed variable that assigns occupational income scores to each occupation. In essence, OCCSCORE assigns each occupation in all years a value representing the median total income (in hundreds of 1950 dollars) of all persons with that particular occupation in 1950. OCCSCORE thus provides a continuous measure of occupations, according to the economic rewards enjoyed by people working at them in 1950.’

²⁴Borjas, ‘Long-Run Convergence’ (1994).