Empirical evidence against U.S. H1-B visa restrictions, the case of employment in the manufacturing industry

This empirical research paper provides ample evidence for policy makers to readdress the immigration policy--especially H1-B visa cap restrictions. The paper is focused on employment shifts and human capital achievement for two groups: the U.S. born and foreign-born working within the United States manufacturing industry. The manufacturing industry is the largest industry employer in the country, and it is in the brink of being at a disadvantage in the global stage due to labor shortage as the workforce ages. The paper uses data for three-year periods--2000, 2010, and 2019--from U.S. Census, The American Community Survey (ACS), Public Use Microdata Sample (PUMS) files, thus providing an overview in labor trends and the human capital needed for the industry to be competitive. The paper builds from the Mincer (1974) earnings function to determine hourly wages for the two groups and then uses the Oaxaca-Blinder (1973) method to measure potential bias between the U.S. born and foreign-born employees in the manufacturing industry. The results in this paper align with other recent research findings (Gest et al., 2021; Eckstein & Peri, 2018) that show immigration as a tool to economic competitiveness. The data trends and findings in this paper synchronize with Borjas and Edo (2021) insights indicating that the native-born may respond to supply shocks of immigration by moving to other labor markets that are not directly affected by immigration.

I.

Introduction
This research paper looks at the individual wages for two separate groups both employed in the manufacturing industry in the United States: the U.S. born and the foreign-born. After the selection of each group, the multiple linear regression is performed to determine the average hourly wages for three -year periods: 2000, 2010, and 2019. Year 2000 is the first year such data was made available and year 2019 is the last release from the U.S. Census.
The purpose of the paper is to determine whether there are earning differences between these two groups. The objective is to examine these differences for these three-year periods and then to determine whether higher rates of investment in human capital --years of education and skills--have resulted in higher wages or are the differences in earnings due to possible bias directed towards a particular group. Literature since 1982 (Borjas, 1982) and current reports (Gest et al., 2021) show that immigrants to the United States have higher incentives to improve their marketable skills. Both incentives and competitiveness results in higher efficiency and economic growth. The aging population and labor shortage in the manufacturing industry goes beyond, individual jobs, or, claims that immigration ousts domestic employment. S. Eckstein & G. Peri (2018, p. 15) concluded that "high skilled immigrant niches contribute the most to the economy, even if they leave some highly qualified U.S.-born workers on the sidelines." Also, aligning with Borjas and Edo, the U.S. born also responds to any shocks from immigration by moving to other labor markets. Therefore, efficiency and economic growth ought to drive the immigration policy, thus enabling not only the manufacturers but other industries to be globally competitive and grow.
The analysis in this paper will point out whether there is any bias towards either of the two groups--U.S. born vs. foreign-born--and assess the literature that indicates that investment in human capital continues to drive economic growth using higher wages as a surrogate for economic growth. The goal of this analysis is that the findings be used by policy makers. Recent research finds that cap restrictions significantly reduce the hiring of new H1-B employees in forprofit firms (Mayda al., 2018). Also, recent survey data from the Philadelphia Federal Reserve Bank (2021) shows that manufacturing firms remain optimistic about growth, with one in three (33%) of manufacturers expect increases (vs. 13% expecting decrease) in general business activity. Thus, the need for the labor force is essential and these restrictions are very costly for the manufacturing industry in terms of sustainability and growth.
I.1 Background U.S. Census data shows that in the manufacturing industry on average, the foreign-born employees' yearly wages are higher than the U.S. born employees by $1,565 for 2000, $2,387 for 2010, and $9,373 for 2019. The median yearly wages were higher for U.S. born by $2,037, $1,008 for the years 2000, 2010 respectively, and lower median wage of $2,020 for the year 2019. See appendix 2 for more details. These differences can be attributed to the investment in human capital as illustrated by the higher median wages in the U.S. for 2000 and 2010 correlating with the distribution of education and experience among employees in the industry.
• 24% foreign-born had no-to-less than a high school compared to 9.8% U.S. born in the year 2000 --  Census Bureau, 2000). This group continues to be the largest group for the U.S. born employees in the manufacturing industry. Table 2 data section below for details for the other years and groups all showing similar correlation.
The two education groups discussed above--no schooling and those with some years of college--continue to be the largest portion of the employees for the manufacturing industry, making 71% of the total labor force for the U.S. born employees and 58.3% of the labor force for the U.S. born employees, thus this group determines the median wage earned --  (Mincer, 1958;Mincer, 1974;Heckman et al., 2003;Lemieux, 2006;Lagakos et al., 2018) The following methodology section of this paper elaborates further. After the equation has been established, then the analysis uses the Oaxaca-Blinder (1973) methodology to determine the differences and potential bias between the U.S. born and foreign-born civilians employed in the manufacturing industry in the United States. Building on these model foundations continues to be a benchmark for economist research.
Based on literary review, this is the first empirical research paper to apply the two models to employment in the manufacturing industry. The Sana Commerce commissioned 2021 Sapio Research study of digital transformation in manufacturing shows that a significant number of manufacturers will continue their long-term investment in digital strategies. Also, 94% of manufacturers will need to change their market strategy due to the disruption faced from the pandemic in the year 2020. As a result of the evolution in the industry, the need for an expanded labor force--especially skilled labor--increases, and this paper concludes that the need for loosening visa restriction--particularly increasing the H1-B visa cap restrictions--will help the U.S. manufacturing industry's sustainability and competitiveness. II.

Data & Manufacturing Industry
This paper examines employed private civilians in the manufacturing industry in the United States. The data used in this study is retrieved from the one-year American Community Survey  Table 1 below gives a mean summary of the individuals included in the data set for each year period studied; for complete summary statistics tables see Appendix 1.  Data shows that the gap in higher education attainment among the two groups continues to be significantly different-with the more than 10% of foreign-born individuals pursuing bachelors' or higher degrees. This trend points out that foreign born individuals are more likely to increase their human capital via education. Given the endless and needed evolution in the industries, an educated individual will be in a better position to be employed and/or earn higher wages, another reason which demonstrates the need to ease visa restrictions. Table 3 below shows the distribution of job role categories for both U.S. born and foreign-born employees in the manufacturing industry for the three time periods from the Census data years:

II.1.B. Employment roles in the manufacturing industry
2000, 2010, and 2019.  Table 3 shows that differences in the distribution of employment roles are smaller between the two groups than the differences in the education attainment from Table 2. The U.S. born have a slightly higher percentage when it comes to management and office administration related roles, while foreigners continue to maintain the lead in engineering and computer/machine operation job roles.
II.1.C. Average hourly wages in the manufacturing industry Table 4 below shows the mean wage, standard no, and the median wage for the manufacturing industry for the three periods being studied. Data indicates that the foreign-born workers make more on average than U.S. born employees in the manufacturing industry. The data also shows the median wage is higher for the U.S. born for the years 2000 and 2010, which aligns with literature suggesting that the native born are likely to move to other labor markets that are not directly affected by immigration and where, presumably, wages could drop (Borjas and Edo, 2021). To explain why this is the case, regression analysis is needed to determine if any bias exists among the two groups. This paper presents a comprehensive model to determine the wages in the manufacturing industry while controlling for several individual characteristics and pointing out any differences among these two groups. The findings are useful validation for policy makers to ease immigration policy and/or the visa entry into the USA. The paper has three additional sections: methodology, results, and conclusion.

III. Methodology
To determine the employee wages in the manufacturing industry this paper applies and utilizes the earning equation that is primarily and widely used in labor economics literature (Mincer, 1958;Mincer, 1974;Lemieux, 2006;Weng et al., 2019). After the model is established, the paper utilizes the Oaxaca-Blinder decomposition method to point out any bias or differences between the two groups in the study, another widely used method that still works a benchmark in differences and bias across various disciplines in research (Oaxaca, 1973;Blinder, 1973;Averkamp, at al., 2020;Koh at al., 2020;Rahimi & Nazari 2021;An & Glynn, 2021).

III.1 The Equation
The equation is based on the Mincer (1974) earning function which states that years of schooling and potential experience are the determinants of wage earnings. This function continues to be a benchmark when it comes to wage earnings based on human capital-such as years of schooling and work experience.
Building on this model and given the nature of our analysis, in addition to schooling and potential work experience in this research are added several characteristics to the equation, such as marital status, gender, race, regions, and the occupation category for everyone in the dataset. The dependent variable is the natural log of the hourly wage earned. The natural log, similar to Mincer (1974) indicates that there is strong evidence to support the log earnings specification.
The wage per hour is calculated from the dataset given three variables, income in the past 12 months divided by number of weeks worked divided by number of hours worked per week.
The experience variable is based on the calculation of potential work years, Age minus years of schooling minus 6. We follow a similar approach to Mincer (1958)  The U.S. Bureau of Labor Statistics publishes average earnings by occupation and the data indicates differences among occupation categories, race, and regions in the country. To address these differences and be able to have a more comprehensive conclusion, this paper includes these variables in the equation. This paper also addresses the vast literature on gender wage discrimination, therefore the gender wage gap is included as a control for this discrimination/bias. Including the above-mentioned variables, we arrive at the equation.
• SE: Vector -These are dummy variables representing, bachelor's, master's, professional, and doctorate degrees. Important to ensure that the standard Mincer equation does not overstate the year before graduation and/or understate the year after graduation (Lemieux 2006).
• : Individuals potential work experience & 2 is the squared term of individuals' potential work experience (we expect positive sign for as more years of experience the wages rise, we expect a negative sign for 2 due to diminishing return to work experience, and 3 research points out that we add higher order polynomials to fine-tune the standard Mincer model, otherwise the e equation will understate the wage growth for younger workers (Lemieux, 2006).

III.2 Oaxaca-Blinder Decomposition Methodology
Oaxaca -Blinder (1973) decomposition is a method to analyze the wage differentials between two groups and determine the portion of the wage differential that is attributable to differences in education and experience and the portion that is potentially attributed to bias or discrimination toward one group (Oaxaca and Blinder 1973). We use this method to measure the differences between our two groups and the results from the above equation.
To best understand the model, we will consider the base of the equation, which is a white single male with less than a college degree living in the northeast and working in the production line in the manufacturing industry. This is the base for the equation Foreign born (f): ( ) = + 1 ( ) + 2 ( ) + 3 ( 2 ) + 4 ( 3 )

Potential bias differences equation (Ω)
: The results in the above equation will show the potential bias towards the subgroup mentioned above. To compare the same subgroup with a college or higher degree the equation takes the following form. We only need to take the differences in coefficients for each attribute since the dummy variable only takes two values of 1 or 0. To expand the subgroup to include other races, regions, occupation categories, we just subtract the coefficients from each other. The next section shows the results of the equation and the potential bias differences.

IV. Results
The regression equation analyzes the data for the three periods for each group, U.S. born vs.

IV.1 Oaxaca-Blinder decomposition in the manufacturing industry
The base of the model is -white male employee -located in the northeast -working in the production occupation--does not have a bachelor's degree -and is not married.

V. Conclusion
The manufacturing industry, being the largest employer and having a large share of foreign-born employees in their payroll, is in the brink of a disadvantage globally, due to aging current employees and the demand for highly skilled labor force. The key finding in this research shows that there is no bias toward either group, U.S born vs. foreign-born, a finding that demonstrates that the manufacturing industry in the U.S. does not favor any group based on birth origin. This finding is both attractive and provides competitiveness for skilled employees within and outside the country.
Data shows those holding a bachelor's degree or higher make between 28-37% of the employment in the manufacturing industry. This data is a clear indicator that a qualified educated labor force is required for the industry to be competitive on the global stage and such a labor force that the manufacturers will have to look beyond U.S. borders.
The results confirm that the higher earnings from foreign born employees (Table 4) in the manufacturing industry is due to human capital advantage, on average more years of schooling and more years of work experience. The findings align with previous research stating that foreign born have higher incentives to complete more years of education (Borjas, 1982). The results also indicate that, in recent years, the bias towards the two groups has nearly diminished. Therefore, for the U.S. to continue to have a competitive manufacturing industry, it must simplify and ease the restrictions in labor hirings. The industry needs a qualified labor force as the U.S. remains a desired location for educated skilled labor. The findings show no bias, so the desire is supported by evidence that the U.S. has the tools to attract the needed talent. Loosening the H1-B visa restrictions will allow the manufacturing industry to attract talent across the globe, and these individuals will have well paid jobs thus resulting in higher tax revenue as well.
The results demonstrate that by 2019 the industry posed little to no bias towards either of the groups, U.S. born vs. foreign-born. The data trends show shifts towards higher demand for skilled labor force in the manufacturing industry. The evidence demonstrates the need for policy makers to loosen the immigration policy and attract talent beyond U.S. borders, especially increasing/or eliminating the cap restrictions on H1-B visas.

Declaration Ethics approval and consent to participate
This study uses individual data made available by the U.S. Census Bureau to all researchers. The ACS PUMS files are a set of records from individual people or housing units, with disclosure protection enabled so that individuals or housing units cannot be identified.
"The Census Bureau has created these data to exclude information that would directly identify respondents and characteristics that may lead to the identification of respondents. The Census Bureau provides these data to facilitate statistical research and analysis." https://www.census.gov/data/developers/about/terms-of-service.html Consent for publication I, Erjon Gjoci, give my consent for the publication of "Empirical evidence against U.S. H1-B visa restrictions, the case of employment in the manufacturing industry" research paper to the Journal for Labour Market Research.

Availability of data and materials
Data is available to all researchers. PUMS files on the File Transfer Protocol (FTP) site are available in CSV and SAS formats. PUMS data prior to 2005 can also be found on our FTP site. Microdata access from 2005-current is available on data.census.gov.