Working Paper No. 12-07
Immigration Quotas and Immigrant Skill Composition: Evidence from the Pacific Northwest
Catherine Massey
October 2012
ABSTRACT
Estimation of the causal effect of immigration restriction on the size and structure of migration
flows is complicated by selection issues and the fact that contemporary migration policies
operate concurrently with other entry restrictions. This paper examines the effect of immigration
quotas on the skills of incoming migrants using implementation of the quota outlined by the
Emergency Immigration Act of 1921 as a source of exogenous variation in migrant skill. The
1921 quota restricted entry to three percent of the nationals who were in the U.S. according to
the 1910 census. Unaffected migrants, such as Canadian-born, foreign-born Canadian citizens
and Japanese migrants, are used to construct a control group against which to compare those
affected by the quota. Newly transcribed, individual-level data collected from ship passenger
lists report occupation, birthplace, and place of last residence, which are necessary for measuring
skill and constructing the control group. By studying migration quotas in the 1920s, a true
before-and-after comparison of restricted and unrestricted migrants can be made bereft the
complicated structure of contemporary immigration policy. Difference-in-difference estimates
indicate that the 1921 quota resulted in flows of higher-earning immigrants. The probability that
a migrant was low skilled decreased by fourteen percent and the probability that a migrant was
medium skilled increased by sixteen percent. Importantly, this measured increase in skill is not
solely due to changes in the national-origins mix of migrants. Only twenty-seven percent of the
change in skill resulted from changes in the national-origins mix of migrants, implying that the
majority of the change occurred due to within-country changes in the quality of migrants.