F. Selective Migration
V. Occupational Representation in Official English States
Though I find an impact of Official English laws on male LEP worker wages, it is unclear through which mechanism these effects are occurring. I supplement my analysis with detailed data on occupational requirements to determine if Official English laws affect occupational distribution. If I do not find evidence of worker behavior changes in response to these laws, other mechanisms—such as differences in productivity or employer discrimination—will underlie my results.
Past work has examined the differences in immigrant earnings among
occupational classifications (Kossoudji 1998; Chiswick et al. 2005). These studies do not take into account the substantial variation in job requirements among these
classifications. For example, the occupational category “managers” encompasses both Chief Executive Officers and cafeteria directors, which have differing levels of English level requirements. I use the Occupational Information Network (O*NET), a detailed
dictionary of job descriptors for 798 occupations to capture this variation in English language requirements across and within occupational classifications.
Researchers have used O*NET to examine immigrant workers’ specialization and job matching, but have not examined how legal changes affect occupational matching. Peri and Sparber (2009) use O*NET measures of physical tasks and language ability
requirements and find that immigrants and native-born workers specialize in jobs with different tasks. Chiswick and Miller (2010) use O*NET language requirement measures coupled with the 2000 Census to show that there is an earnings premium for workers that match their language skills with job requirements. This section extends this work by examining occupation matching not only by occupational language requirements, but also within the context of legal changes regarding language. This Chapter is the first to
examine how occupational choice is affected when Official English laws are adopted.
A. Data
O*NET is a comprehensive data source on job characteristics and worker attributes within occupations. O*NET is administered by the Department of Labor as a replacement for the Dictionary of Occupational Titles (DOT). O*NET provides ratings of occupational characteristics and skill requirements, including communication and
interpersonal contact requirements. Since the inception of the O*NET 1.0 in 1998, O*NET has been extended and updated based on input from job analysts and workers. I use O*NET Version 21.3, which was released in May 2017.
O*NET contains two questions of particular relevance to this analysis: “how important is knowledge of the English language to the performance of your current job?”
and “how important is communicating with others outside of your organization to the
performance of your current job?”. I use both of these questions to determine the importance of communication in English to individual’s occupations. O*NET rates the importance of both speaking English and communicating with the public on a five-point scale: (1) not important, (2) somewhat important, (3) important, (4) very important, and (5) extremely important.
To examine how pay varies with job attributes, I merge O*NET with Census and ACS data. O*NET identifies 798 occupations using detailed Standard Occupational Classification (SOC) codes, while the Census and ACS identify workers based on Census occupation codes, which is comprised of 533 occupations. To bridge the gap between these two occupation codes, I created a crosswalk between O*NET and Census
occupations using the methodology of Hirsch and Schumacher (2012). There is a one to one match from O*NET occupation codes to Census occupation codes for 491
occupations. Many of the remaining involve mapping two or more O*NET occupations to the Census category. To map these occupations onto the Census, I weight the O*NET descriptor scores of each O*NET occupation using the employment from the
corresponding year reported in the Bureau of Labor Statistic’s Occupational Employment Statistics (OES) as weights. If the OES employment is unavailable, I equally weight the O*NET descriptor scores. The SOC also contains codes that encompass “all other”
categories, such as “sales and related workers, all others.” O*NET does not have ratings for these occupation codes. Therefore, I assign O*NET values based on average ratings (using employment weights) among similar occupations.
The relevant O*NET variables for my analysis are “speaking,” or how important speaking English is in the occupation, and the importance of communicating with persons
outside of the individual’s organization. On average, LEP workers are represented in occupations that have lower importance ratings for both speaking English and
communicating with the public. The average importance of speaking English score for LEP worker occupations is 3.12 on a scale of 1–5, with a standard deviation of 0.44. This average importance is slightly lower than the mean score of English proficient workers, who have a mean importance score of 3.54 and a standard deviation of 0.45. The importance scores of communicating with the public tell a similar story: the average for LEP workers is 2.76 with a standard deviation of 0.64, while the average for English proficient workers is 3.28 with a standard deviation of 0.69. These statistics accord with Chiswick and Miller’s findings (2010) that LEP workers are likely to be in occupations that place a lower importance on speaking English or communicating with the public. I next determine if these differences are also apparent in states with or without Official English laws.
B. Are LEP Workers Less Represented in Occupations that Require English Skills