The Impact of New Immigrants on
Over the 2000-2005 period, immigration levels remained very high and roughly half of new immigrant workers were illegal. This report finds that the arrival of new immigrants (legal and illegal) in a state results in a decline in employment among young native-born workers in that state. Our findings indicate that young native-born workers are being displaced in the labor market by the arrival of new immigrants.
During the last five years, new immigrants have accounted for an overwhelming share of the growth in the number of employed persons in the United States. Native-born adults and established immigrants have been unable to capture much of the new employment opportunities that have been created in the nation since 2000. The number of employed persons in the civilian working-age (16 and over) population rose by 4.835 million between 2000 and 2005. During 2005, a total of 4.134 million new immigrants were working in the United States. New immigrants who entered the United States since 2000 and were still residing here during 2005 accounted for 86 percent of the total increase in employment in the nation over the 2000 to 2005 period. Native-born and established immigrants accounted for less than one-sixth of the total rise in civilian employment that occurred in the nation over the past five years. These findings differ by gender. Among men, new immigrants accounted for all of the rise in employment, as the total number of employed men in the nation increased by only 2.665 million while the number of employed new immigrant males was 2.767 million during 2005. For the first time since the end of World War II, there has been no gain in employment among native-born men over a five-year period.
A substantial share of employed new immigrants appear to be illegal workers, often employed in off-payroll jobs that are increasingly concentrated in a newly emerging informal sector of the American labor market. The Pew Hispanic Center estimates that there were 4.4 million illegal immigrants residing in the United States in 2005 who had entered the country since 2000.1 We estimate that 2.857 million of these new illegal immigrants were actively participating in the labor force during 2005 and that about 5.5 percent of the immigrant labor force was unemployed.2 With a labor force of 2.857 million and an estimated unemployment rate of 5.5 percent, we conclude that the number of new illegal immigrants who were working in the United States during 2005 was 2.7 million. This means that about two-thirds of all employed recent immigrants in the United States were working illegally during 2005 and that more than one-half (56 percent) of the total rise in employment that occurred in the nation between 2000 and 2005 was attributable to the growth in employment among illegal immigrant workers.
The extraordinarily high share of new employment captured by new immigrants was accompanied by a powerful shift in the organization of the nationís labor markets. In a subsequent section of this report we will provide evidence that some employers have begun to re-organize work in ways that systematically exclude certain native-born workers, especially those under the age of 35, from employment and that create work that does not meet the basic labor standards that have been developed over the years by federal and state legislation, custom and tradition, and through labor-management/collective bargaining agreements.
The ability of the nationís teen and young adult (20-24) population to become employed has deteriorated badly over the past five years. Employment levels for all those aged 16 to 34 have fallen by more than 1.5 million between 2000 and 2005, even as the total number of employed persons increased by more than 4.8 million over the same period of time. Several alternate explanations might help explain this employment decline among young people in the nation. Part of the explanation could simply be associated with demographic change. Reductions in the size of the teen and young adult age cohorts can result in employment declines even though the likelihood of a member of that cohort finding work doesnít change. Alternately, changes in the likelihood of becoming employed can reduce the number of young people working. The first explanation has no validity since the number of native-born people aged 16 to 34 rose as the echo generation (baby boomersí children born between 1978 and 1996) moved into this age group in large numbers.
The number of native-born males aged 16 to 34 in the population increased by nearly 1.1 million between 2000 and 2005.3 Rather than reducing employment levels, these demographic forces would have been expected to increase overall employment levels of native-born males aged 16 to 34. Indeed, we estimate that if the proportion of native-born young males working during 2005 were the same as the share of native-born workers employed during the full employment year of 2000, 1.721 million more young native-born men would have been at work during that year. Employment among native-born young men declined not because there were fewer young men, but because their employment rates declined precipitously. The employment to population (E/P) ratio of young males has fallen sharply over the last five years. Some of these declines are quite extraordinary and, in the case of male teens, the 2005 E/P ratio was the lowest in the nation over the entire 58-year period covered by the Current Population Survey (CPS) teen employment series.
Among females, the trends in employment have been similar. While the size of the young native-born and established-immigrant female population has increased at about the same rate as males, the number who are employed has declined sharply. Similar to findings for their male counterparts, the E/P ratio of native-born female teens and young adults fell considerably over the last five years, accounting for all of the decline in employment among young native-born females. If native-born teen and young-adult females had been able to maintain their employment rate at the same level as the full employment year of 2000, then the number who were employed in 2005 would have increased by 1.382 million.
The decline in employment levels among native-born teens and young adults implies that employers have turned to alternative sources of labor supply to meet their labor requirements. One alternative source of substitute labor is, of course, the surging older worker population fueled by the baby boom age cohort entering their pre-retirement years in the past five years. These individuals represent a ready potential source of substitute workers for teens and young adults. The other potential alternative source of labor supply is the flow of new immigrants to the United States since 2000. Large numbers of new foreign workers, the majority of whom entered the United States and work here illegally, also represent a ready source of labor supply to take the place of native-born and established-immigrant teens and young adults in the nationís labor markets.
As noted below, the job deficit for native-born male teens and young adults in the nation was 1.721 million, while the number of new immigrant male workers in the same age group in 2005 was 1.859 million (Table 4). If the jobs held by new immigrant males aged 16 to 34 were made available to jobless native-born males, then the job deficit among the native-born would be completely eliminated. Among women, the substitution of jobless native-born young women for recent young female immigrants would result in the native-born female job deficit declining by more than 60 percent. Overall, nearly 90 percent of the native-born teen and young adult job deficit that has emerged over the last five years would be eliminated if native-born teens and young adults worked in jobs now held by recent immigrants of the same age. While some mismatches in the occupational composition of employment might occur between native-born and foreign-born workers, the jobs held by these groups are quite similar to jobs in all occupations simultaneously held in large numbers by both foreign-born and native-born workers.4 These findings strongly suggest that a major proportion of the native-born job deficit of teens and young adults that has developed in the United States over the past five years is the result of newly arrived, young female, and especially male immigrants displacing these potential workers from employment. Native-born older workers are a much less-likely substitute for employers who hire many fewer native-born teens and young adults. Native-born older workers have differing levels of work experience, expectations of hours and weeks of work, and are paid at considerably higher wage rates than are teen and young adult workers. Recent young immigrant workers are much closer substitutes for young native-born workers compared to the aging members of the baby boom generation.
Impacts of New Immigrants on Young Adult Employment
Most studies of the economic impacts of immigration on native-born workers have focused on wage and annual earnings impacts rather than employment impacts. There is a general tendency among labor market analysts to assume that, as a result of labor market and wage flexibility, there are few job displacement effects of immigration on native-born workers, citing older studies to back up these opinions. Several more recent statistical studies, however, indicate that less-educated native-born workers, teenagers, and black males do suffer employment declines as a result of immigrant labor inflows.5 Ethnographic research work in Boston, Chicago, Los Angeles, New York City, and other large central cities across the nation has revealed that young immigrant workers are often preferred by employers over poorly educated native-born workers, especially those from inner city neighborhoods characterized by high poverty rates.6
One might well expect the labor displacement effects of immigration to be low in periods of full employment, when job opportunities are abundant and vacancy rates are high, such as the late 1990s in the United States when employment rose across the board among both most native-born workers and new immigrant workers. However, in more slack labor market environments, such as the 2003-2004 period, one might well expect that a rise in the supply of immigrant labor could generate displacement impacts on native-born workers, especially among those in most direct competition for available jobs with newly arrived immigrant workers, such as young, native-born adults with limited formal schooling, especially those in central cities.
To test whether the influx of new immigrant workers over the 2000-2003 period had an adverse effect on the employment prospects of the nationís young adults (16-24 years old), we estimated a series of multivariate statistical employment models for young adults, including a variable representing the relative size of new immigrant inflows into the labor force of the state in which the young adult resided at the time of the 2003 American Community Survey (ACS).7 The relative sizes of these new immigrant labor force inflows varied quite considerably across states between 2000 and 2003. The size of these immigrant inflows ranged across the 50 states and the District of Columbia from a low of .2 percent to a high of 3.9 percent, with a mean of 1.63 percent.
The dependent variable in this multivariate statistical model was the employment status of a 16-24 year-old respondent at the time of the 2003 ACS. The variable was a dichotomous variable that took on the value of one if the respondent was employed (either part-time or full-time) and the value of zero if he/she was not employed at the time of the ACS. The right-hand-side predictor variables included the gender, age, race-ethnic group, and educational attainment of the respondent, the unemployment rate of the state in which he/she resided at the time of the survey in 2003, and the relative size of new immigrant labor inflows into the state since 2000. We estimated these employment probability models for all 16-24 year olds and for a variety of gender, nativity, gender and schooling, and school enrollment subgroups.8 The findings in Table 5 display the estimates of a one percentage-point increase in the state labor force due to new immigration on the probability of employment among young adults.
For the entire sample of 16-24 year olds,9 a one percentage-point increase in the state labor force due to new immigration would have lowered the predicted employment rate of such youth by 1.2 percentage points. The estimated impact was highly statistically significant (.001). For a state with a recent large influx of new immigrants (a three percentage-point rise in the civilian labor force of the state), the probability of employment among 16-24 year olds in that state would have declined by a substantial 3.6 percentage points.
The estimated impacts of new immigrant workers on the employment rates of 16-24 year olds were approximately the same for the native-born as they were for all 16-24 year olds, but as expected were considerably larger for men than for women (-1.6 percentage points for men versus -.9 percentage points for women),10 and were larger for less-educated women than for women with some post-secondary schooling.11 The finding of larger adverse employment impacts for men than for women is not surprising given the relatively high share of new immigrant workers that were men (66 percent). Larger adverse impacts for less-educated workers were also expected given the above-average share of new immigrant workers who lacked a high school diploma and the weaker national labor market for less-educated native-born workers during this time period. The results in Table 6, thus, provide substantive empirical evidence that the recent influx of new immigrant workers has resulted in job losses for many subgroups of young adults in the nation, especially in those states that were more heavily impacted by new immigrant labor. Males, in-school youth, less-educated workers, and black males appear to have been more adversely affected than other demographic subgroups of young adults.
The availability of the public use micro data from the 2004 ACS allowed us to update our findings on the impacts of new immigrant worker inflows in states on the employment probabilities of very young adults. Given the continuing severe labor market problems of teens and youth in their early 20s throughout 2004, we selected 16-20 year olds for our analysis. There were observations for approximately 74,000 youth in this age group on the ACS public use files, of whom 58,600, or nearly 80 percent, were enrolled in school at the time of the ACS surveyís completion.12
The dependent variable in these models is the employment status of the respondent at the time of the survey. Those employed respondents, including persons with a job but temporarily absent due to vacation, weather-related factors, etc., were coded as a "1" and all others as "0." In these models, we control for a wide array of demographic and family income background variables, the school enrollment status and educational attainment of the respondents, the unemployment rate of the state in which they lived, and the relative size of new immigrant worker inflows since 2000.13
These regression models of young adult employment rates were estimated for all 16-20 year olds and for a variety of gender, race, and school enrollment subgroups. Estimates of the impact of new immigrant inflows on the probability of employment of young adults are displayed in Table 6.
For the entire group of 16-20 year olds, the presence of new immigrants in their stateís workforce had a strong, statistically significant, negative impact on the likelihood that they will be employed. A one percentage-point increase in the share of new immigrants in the stateís workforce will reduce the probability of employment of young adults by 2.1 percentage points. The effects of new immigrant workers are negative and statistically significant for each subgroup of young adults in Table 6, and are equally large for both men and women,14 but they are much larger for in-school youth than for out-of-school youth (2.4 percent vs. 0.6 percent). The size of the coefficient was highest for black men, implying that they are the most adversely affected by new immigrant inflows.
New Immigrant Workersí Impact on the Job Market
The rise in immigrant employment, especially among illegal workers, over the past decade has been accompanied by a number of important changes in the structure of employment relationships in U.S. labor markets. Recent years have seen the growth in contractor employment relationships and the use of independent consultants and off-the-books workers.15 These newly hired workers do not go on the formal payrolls of the firms that hire them, and they typically are not paid employee benefits such as health insurance and pension benefits or covered by the Unemployment Insurance, workers compensation, or Social Security systems.
These changing employment relationships are not simply revealed in growing media coverage of labor market developments at the local level, but also show up in the large differences between employment changes registered by the two national surveys used by the U.S. Bureau of Labor Statistics to estimate monthly employment, the Current Employment Statistics Survey (CES) payroll survey and the Current Population Survey (CPS) household survey.16 The CES generates a monthly count of wage and salary payroll jobs from a monthly sample of about 160,000 businesses and federal, state, and local government organizations covering 400,000 individual establishments that participate in the unemployment insurance system. The CES is considered by many economic and financial analysts to be the primary source of data on wage and salary job growth and decline in the nation and among states and is a primary topic of discussion and analysis in BLSí monthly Employment Situation news release, which is widely covered by the national media. One of the most important uses of the CES data at the national level is to measure the job-generating performance of the economy over the course of the business cycle.
A second source of information on monthly employment trends at the national and state levels is the findings of the CPS. The CPS is a survey of approximately 60,000 households conducted each month by the Census Bureau for the U.S. Bureau of Labor Statistics. Unlike the CES, which measures only the number of private and public formal payroll jobs, the CPS provides a more comprehensive count of the number of employed persons ages 16 and older each month. The CPS employment count includes not only workers in traditional wage and salary jobs, but also workers outside the scope of the payroll employment survey, including agricultural workers, the self-employed, independent contractors, unpaid family workers, and some "under the table" or "off-the-books" workers.17 The CPS survey counts each employed person only once, regardless of the number of jobs he/she holds at the time of the survey, while persons holding multiple wage and salary jobs will be counted twice in the CES. Historically, the CPS and CES employment measures have tracked one another fairly well. However, during the past five years considerable differences have emerged between the two surveysí estimates of the overall increase in the nationís employment levels, with the CPS showing much greater growth in private sector wage and salary employment. These findings stand in sharp contrast to that observed for earlier time periods.
During both the 1980s and 1990s economic expansions, the growth in payroll employment levels in the nation was greater than that measured by the household survey. Typically, payroll employment levels in the nation grow rapidly during the early stages of recovery from an economic recession. Rising Gross Domestic Product (GDP) increases the demand for labor by employers who then add more workers to their payrolls. Over the period from the early 1960s though 2000, the nation experienced five recoveries from economic recessions. On average, the nationís wage and salary employment levels increased by 11.7 percent during the first four years of recovery for these five post-recession periods. The rates of new job creation varied from a low of 7.3 percent during the early stages of recovery from the 1990-1991 recession to a 16 percent rise in the nationís wage and salary employment levels during the recovery from the recession of 1974-75.
However, the rate of job growth during the first four years of recovery from the recession of 2001 has been much slower than the historic pace of national payroll employment growth in recovery periods. Despite robust rates of growth in real GDP, strong growth in corporate profits, and a stock market boom, the nationís rate of new payroll employment growth was just 2.5 percent between 2001 Q4 and 2005 Q4. This rate of new job creation was equal to only one-fifth of the historical average rate of new job creation over the previous five recoveries (Table 7). Why has the rate of payroll employment growth been so slow over the past four years, given the strong overall performance of the nationís economy by most key economic indicators? Increased labor productivity growth appears to be an important explanation, but part of the answer is associated with strong growth in off-payroll employment, especially among the recent-immigrant population. Since the end of the 2001 recession in the fourth quarter of 2001, payroll employment in the nation increased by just 3.23 million jobs while the number of working-age persons who were employed, according to the CPS, rose at twice that pace, increasing by 6.446 million (Table 8). Unlike the employment expansions of the 1980s and 1990s, when payroll employment growth substantially outpaced that of household employment, the current recovery is characterized by a new pattern of job growth.
Over the entire 2000 to 2005 period, the nature of the relationship between the employment growth estimates of the two surveys has changed radically. Between 2000 and 2005, wage and salary employment levels, as measured by the CES, rose by only 1.678 million or 1.3 percent while the CPS found that the number of employed workers increased by 4.672 million over the same period of time (Table 9). On an annual average basis, we find that employment as measured by the CES business establishment survey increased from 131.785 million during 2000 to 133.463 million by 2005, an increase in non-agricultural payroll jobs of only 1.678 million. In contrast, the household survey found that the number of working-age persons employed in the nation increased from 136.934 million to 141.606 million, a rise of 4.672 million over the 2000 to 2005 period, a difference of nearly three million.
The CPS household survey measured a rise in employment that was nearly three times greater than that measured by the CES over the 2000-2005 period. As we noted earlier, the CPS and the CES use somewhat different employment concepts. The CPS includes agricultural workers, the self-employed, independent contractors, unpaid family workers, and some off-the-books workers while the CES does not.
We have adjusted the CPS employment data to more closely fit the CES employment concepts in order to obtain a more direct comparison between the CPS and CES measures of employment change over the 2000 to 2005 period.18 Our first adjustment was to exclude agricultural workers from the CPS household survey employment count since the CES measures only employment in the non-agricultural sector of the nationís economy (Table 10). After excluding agricultural workers, non-farm employment as measured by the CPS survey increased more considerably ó by 4.976 million between 2000 and 2005 ó indicating that employment among agricultural workers declined over this five-year period. It is important to note that recent immigrants are about 1.8 times more likely to work in the nationís agricultural industries than are the native-born.
The estimated decline in agricultural employment over the last five years suggests that this sector was not a major source of new employment opportunities for new immigrants.19 Excluding agricultural sector employment from the CPS totals further widens the difference between the CPS employment growth estimate and the CES job growth estimate over this period, raising the size of the gap in employment growth from 2.994 million to 3.268 million. The CPS estimate of new employment growth rises to 2.95 times that estimated from the CES payroll survey versus only 2.78 times when agricultural employment is included in the CPS totals.
Much of the new payroll job creation that occurred in the nation over the 2000 to 2005 period was concentrated in the government sector. About 60 percent of the total rise in payroll employment that was generated nationally over the last five years has been on government payrolls. Native-born workers are much more likely than immigrants, especially recent immigrants, to work in federal, state, and local government agencies. During 2005, native-born workers were three times more likely to be employed in a government job compared to employed recent immigrants (Table 11). The CES estimated that between 2000 and 2005, federal, state, and local government payroll employment increased by 1.023 million jobs while the CPS found that the number of persons who said they were employed by the government increased by 1.143 million. Thus, the CPS government employment growth estimate was nearly identical to that of the CES.
Findings from the CES, however, reveal very small increases in private sector wage and salary employment in the nation over the past five years. The CES found that non-farm, private sector payroll employment increased by just 665,000 jobs over the past five years. In contrast, the CPS household survey estimated that the number of persons employed in non-farm, private sector jobs increased by 3.026 million. The CPS estimate of non-farm, private sector employment growth between 2000 and 2005 was more than five times larger than that estimated by the monthly CES establishment survey. Recent immigrant employment is heavily concentrated in the private non-agricultural sector of the nationís labor market. While about three-quarters of all native-born workers are employed in private wage and salary jobs, 86 percent of recent immigrants report that they work for an employer in the private non-farm sector.
Over the past five years, the relationship between the CPS and CES estimates of employment growth rates has changed dramatically. Instead of observing the pattern of substantially more payroll job growth compared to increases in the number of employed persons from the CPS prevailing in the 1980s and 1990s, the employment data since 2000 reveal much higher growth in employment measured by the CPS relative to the slow growth registered by the CES. We also have analyzed the relationship between the CPS and CES estimates of job growth at the state level over the last five years. Our findings reveal that those states that had large increases in the number of employed immigrants were also those states with the largest gaps in employment growth estimates between the Local Area Unemployment Statistics (LAUS) program and the count of jobs from the state CES survey. The LAUS program is a statistical program used by states to estimate the monthly number of employed and unemployed residents.
For example, the findings in Table 12 reveal that, while the number of employed residents in the state of Texas increased by 733,000 between 2000 and 2005, total payroll employment levels in the state increased by less than half of this amount, rising by just 308,000 over the same five-year period. At the same time, the number of new working immigrants in the state increased by more than 388,000, the second largest increase in the nation. A look at the top-20 states ranked by the size of the CES-CPS employment growth gap reveals a fairly strong connection between the size of the gap and the size of the increase in the number of new immigrants employed in each state. The correlation between the CES-CPS employment gap and growth in employed immigrants is quite high. We estimate a correlation coefficient of .79 between the absolute size of the difference in employment change between the two jobs measures and the change in the number of employed immigrants in each state over the 2000 to 2005 period.
The above findings imply that large numbers of these new immigrant workers are not appearing on the formal payrolls of their employers. Instead, they are being hired as independent contractors or completely off the books and being paid in cash. Evidence from other data sets, field research by the authors, and growing media stories support this assertion that a high share of new immigrants, especially illegal immigrants, are employed in the informal or "black" economy.20 In 2003 and 2004, only one-third of new employed immigrants reported being covered by a health insurance plan at work and fewer than one in five reported that they were covered by a pension plan at work, versus nearly one-half of the native-born.21 Among less-educated workers from Mexico and Central America, the dominant sources of illegal workers, only about 15 percent reported any health insurance coverage from their employers. When unemployed, fewer than 10 percent report being covered by unemployment insurance benefits.
The employment growth gaps between the CPS and the CES at the national level were systematic in nature and specific to particular classes of workers. The size of the employment growth gap for the government sector of the labor market was quite small. Government was among the least important sources of jobs for employed new immigrants and access to government jobs is largely confined to formal wage and salary positions. Few illegal workers have the opportunity to find work in most government organizations. Strict hiring protocols dramatically limit the potential use of off-the-books work for many government positions. The comparatively small employment growth gap between the household and payroll survey for the government sector appears to be the result of increasing use of workers as independent consultants by some state and local government agencies, a common practice in states such as Massachusetts.
In contrast, the CES data reveal little growth in the nationís non-farm private sector wage and salary jobs over the past five years. These positions are ones in which the overwhelming majority of employed Americans work. They are characterized by a formal employer-employee relationship such as that defined in the Social Security Act. Indeed, a hallmark of formal payroll jobs is the automatic payroll deductions made for employee contributions to the Old Age, Survivors, Disability, and Health Insurance trust fund. Yet, in contrast to the very slow private sector wage and salary job growth as measured by the CES, the CPS reveals non-farm, private wage and salary growth that was 5.5 times higher.
We find that the unprecedented gap between the household and payroll surveysí estimates of employment growth over the past five years is primarily the result of concentrating new employment growth in independent contractor and off-the-books jobs. Employers in many sectors, especially construction, landscaping, retail trade, office cleaning, and leisure and hospitality industries as well as in private households where strong job growth also has been reported in recent years, are increasingly re-organizing work to take advantage of the substantial influx of new illegal immigrants into the United States since 2000. Many of these jobs are filled by illegal immigrants who arrive on street corners, informal shape-ups, and convenience store parking lots waiting for any of a number of potential employers to come by and pick them up for a dayís work.
Increasingly, the nationís employers seem to be operating outside of the legal framework that has defined U.S. labor markets since the New Deal. Expansion of contract employment, off-the-books workers, and black labor markets in an increasing number of communities throughout the nation has meant that a growing fraction of workers now provide their labor outside of the fundamental worker protections that the nation had previously taken for granted, including wage and hour laws, worker safety and health mandates, and minimum wage protections established over the past 70 years. These changes in labor relationships also have reduced rates of unionization, lowered the share of workers receiving key employee benefits, such as health insurance, paid vacations, and pensions and have decreased unemployment insurance, Social Security, and workersí compensation tax receipts.
The growing inflow of illegal-immigrant workers has contributed to a fundamental breakdown in the nationís labor laws and labor standards as the sheer volume of illegal hiring activity overwhelms what has amounted to meager enforcement levels of basic labor standards across the nation by federal and state officials from both political parties.22 Absent renewed efforts to strengthen enforcement of both border security and federal and state labor laws, these new forms of work organization will continue to grow in the future. The past formal relationships between workers and employers will continue to unravel, undermining the unemployment insurance and social security systems and basic worker protections that have evolved in the nation over the last century. These adverse effects on employer-worker relationships have to be taken into account in any benefit-cost calculus of the impacts of new immigration. Advocates of guestworker programs have been derelict in addressing these key economic concerns.
2 Our estimates of the size of the immigrant labor force are based on applying population shares by age/sex group and labor force participation rates for key age/sex groups in the new immigrant population to Pew estimates of the number of illegal immigrants for each of these age/sex groups.
3 All of this increase was among native-born males aged 16 to 29. This age cohort increased in size by more than 1.6 million. However, this was partially offset by a decline of 537,000 in the number of native-born persons aged 30 to 34 in the nation.
4 Steven Camarota found this was the case between foreign-born and native-born workers in general: See Steven Camarota, Dropping Out: Immigrant Entry and Native Exit from the Labor Market, 2000-2005, Center for Immigration Studies, Washington DC, March 2006. http://www.cis.org/articles/2006/back206.html
5 For recent statistical evidence on the links between immigrant worker inflows and the employment of native- born workers, See: (i) George Borjas, "The Labor Demand Curve is Downward Sloping: Reexamining the Impact of Immigration on the Labor Market," Quarterly Journal of Economics, November 2003, pp. 1335-1374. (ii) Paulo Tobar, The Employment Experiences of Teens in Central City Labor Markets: The Influence of Demographic/Human Capital Traits, Family Background, and Environmental Factors, M.A. Workshop Paper, Department of Economics, Northeastern University, Boston, 2004; (iii) Ishwar Khatiwada, Andrew Sum, and Tim Barnicle, New Foreign Immigrant Workers and the Labor Market in the United States, February 2006.
6 See: (i) William Julius Wilson, When Work Disappears, Alfred Knopf, New York, 1996; (ii) Katherine S. Newman, No Shame in My Game: The Working Poor in the Inner City, Russell Sage Foundation, New York, 1999.
7 The immigration variable is defined as the ratio of the number of new immigrant labor force participants in the state between 2000 and 2003 to the size of the resident civilian labor force of the state in 2003.
8 The models are linear probability models estimated by ordinary least squares regression techniques. The coefficient on the foreign immigrant labor force variable indicates the percentage point change in the likelihood of employment among the designated group from a 1 percentage point increase in the stateís civilian labor force due to new immigration.
9 There were 127,151 16-24 year old youth in the ACS sample.
10 The difference between the coefficients of the new immigrant labor force variable in the male and female employment models was large enough to be statistically significant at the .01 level.
11 In fact, the coefficient on the new immigrant labor force variable was not statistically significant at the .05 level in the model for women with 13 or more years of schooling.
12 The ACS questionnaire asks respondents whether they had been enrolled in school at any time in the prior three months. If they answer "yes" to this question, they are classified as enrolled in school. Persons must be attending a school or college that will lead to the attainment of a regular diploma or a college degree.
13 The new immigrant worker variable is measured similarly to that for the previous models based on the ACS 2003 data. It is the ratio of the number of new immigrant labor force participants in 2004 as a percentage of the stateís resident labor force in 2004.
14 The modestly larger coefficient of the immigrant variable in the male equation (.022 vs. .019 for women) is not significantly different from that of women.
15 For a review of these changing job market operations in Massachusetts and the United States, See: Paul E. Harrington and Andrew Sum, "As Jobs Go Off the Books, Immigrants Edge Out Some Native-Born Workers," Commonwealth, Volume 11, Number 2, 2006, pp. 83-90.
16 For a recent review of conceptual differences between the two surveys, See: Mary Bowles and Teresa L. Morisi, "Understanding the Employment Measures from the CPS and CES Surveys," Monthly Labor Review, February 2006, pp. 23-38.
17 It is not clear that all off-the-books workers will report their employment to the CPS interviewer despite guarantees of confidentiality. Besides, immigrants have been historically undercounted in the CPS survey.
18 Changes in multiple job holding can also be a source of divergent growth in employment levels between the two surveys. While important in the past, this factor appears to have had little impact on the employment estimates of the two surveys during the first half of this decade. The number of persons who held multiple jobs remained virtually unchanged between 2000 (7.556 million) and 2005 (7.546 million). See: U.S. Bureau of Labor Statistics, Employment and Earnings, January 2001 and January 2006, Washington, D.C.
19 Jeffrey Passel estimates that fewer than 4 percent of all illegal workers in the nation worked as agricultural workers in 2005.
20 See: (i) Naomi R. Kooker, "Hospitality Immigrant Quandary," Boston Business Journal, April 2006; (ii) Josh McHugh, "Notes from the Underground Economy," www.cnn.money.com , May 30, 2005; (iii) Casey Ross, "Contractors: Stop the Illegal Insanity," The Boston Herald, May 5, 2006; (iv) Peter Reull, "Shadow Workers: Towns Take Aim at Illegal Restaurant Help," The Boston Herald, May 4, 2006; (v) Shawn Sutner, "Illegal Immigrants: These Workers Are Often Anxious and in a Constant State of Fear," Worcester Telegram, April 16, 2006.
21 These estimates are based on our analysis of the March 2004 and March 2005 CPS work experience supplements, which capture information on health insurance and pension coverage.
22 The Washington Post recently reported that, during 1999, only 182 employers were prosecuted for unlawfully employing immigrants. Remarkably this figure fell to just four prosecutions during 2003. See: Spencer S Hsu and Kari Lydersen, "Illegal Hiring Is Rarely Penalized," The Washington Post, June 19, 2006.