English Proficiency, Gender and the Occupations of Childhood Immigrants in the US.

This paper extends the evidence on the role of English proficiency in determining labor market outcomes of immigrants by analyzing its effects for those who arrived to the US before age 19, separately by gender. To overcome the concerns of using self-reported proficiency, it employs a set of instrumental variables that exploit the variation of proficiency driven by both the age at arrival and the linguistic proximity to English of the language of the country of origin. Higher predicted English proficiency has a limited impact on probability of employment for men and women, but it determines the nature of employment and the extent of the impact varies substantially by gender. Overall, higher predicted English proficiency is negatively associated with the level of Manual Dexterity required in an individual's current job but positively associated with Oral and Written Expression, especially for men. While the requirement of Mathematical Reasoning increases with English proficiency for the full sample, it is not significant when we exclude Central Americans and Mexicans. OLS estimates for men in the complete sample seem to suffer from attenuation bias, while larger IV estimates for women point to an ability bias. [ABSTRACT FROM AUTHOR]