Академический Документы
Профессиональный Документы
Культура Документы
Immigrants in America
Sam Kwainoe
C
ompared to most other industrialized cohorts are less skilled than previous cohorts, later
countries, the U.S. has high levels of legal and arrivals can expect to earn lower incomes over their
illegal immigration. Recent talks between working lives than earlier cohorts (1995). This is a
Mexico and the U.S. have stirred public debates on frightening conclusion! Also, the declining wage level
the impact of legalizing approximately 4 million illegal for immigrants increases the income inequality between
Mexicans in the U.S. as temporary workers. the skilled and unskilled workers, thereby increasing
According to Steven Camarota (2001), these illegal the dependence of poor immigrants on social
immigrants will become temporary workers and programs.
eventually attain permanent residence after a period The question about the welfare of less skilled
of indenture of about three to five years. In fact, immigrants is the topic of this paper. This paper departs
immigrants legal and illegal now make up 13% from previous work on immigrant cohorts and instead
of the nations workers, the highest percentage since examines young native and legal immigrant adults
the 1930s. They dominate job (unskilled) who came to the U.S.
categories at both ends of the ...the declining wage level with their parents. This paper will
economic spectrum. Immigrants for immigrants increases argue that for this sample,
hold 35% of unskilled jobs, the income inequality unskilled immigrants start off with
according to the Center for
between the skilled and very low wages compared to
Immigration Studies, a think tank natives, but eventually pick up after
in Washington, D.C. Out of this unskilled workers, thereby acquiring U.S.-specific skills. This
large number of immigrants, a good increasing the depen- is probably due to positive
proportion are unskilled, which dence of poor immigrants selection of immigrants and other
spells out the danger of allowing less on social programs. factors discussed later in the
skilled immigrants into the U.S. paper. Therefore, it can be argued
In 1996, foreign that unskilled immigrants are not
immigrants who entered the U.S. legally since 1980 a drain on the economy.
made up roughly seven percent of the U.S. workforce, In order to test this hypothesis, data from the
and more than 17 percent of those workers had less National Longitudinal Survey of Youth (NLSY, 1997),
than a high school diploma (Deavers, 1999). This low comprised of a select sample of native and immigrant
level of education among immigrants raises questions children, will be used. Section II gives a brief
about their survival in the new information economy background on immigration. Section III provides a
as well as their assimilation process. theoretical foundation for the topic. Section IV lays
Numerous authors have expressed concerns out the hypotheses and explains the empirical model
about the declining level of education and skill level and the data. Finally, Section V discusses the results
that recent immigrants possess and have spelled out of the analysis, and Section VI draws conclusions from
the dangers of allowing such immigrants into the the results.
country. In their research, Edward Funkhouser and
with less equal income distributions relative to the U.S. qualifications lowers entry earnings and vice versa.
(Duleep, 1997). This decline, according to Duleep, Thus, the human capital framework affects earnings
was intensified by the 1965 Immigration and positively. From the analyses, the income distribution-
Naturalization Act, which emphasized family-based immigrant ability model is ambiguous and difficult to
admissions (1997). Obviously, the admission of more measure empirically so this paper will instead
unskilled immigrants leads to a decline in the concentrate on the skills-transferability model by
education-adjusted entry earnings of immigrants. Duleep and the human capital framework described
According to Duleep, countries that have dominated above.
U.S. immigration in recent years have less equal
income distributions relative to the U.S. than C. The Skills-Transferability Model
predominantly European source countries that The skills-transferability model provides an
dominated in the past. This unequal income distribution alternative hypothesis. It shows that the decline in the
inevitably causes people at the bottom of the income education-adjusted entry earnings of immigrants
distribution to migrate to developed countries in search reflects the decline in the extent to which the country
of wealth. In fact, George Borjas (1994) theorizes of origin skills are transferable to the United States
that immigrants from countries with greater income (Duleep & Regets, 1997). This model emphasizes
inequality than the United States (like Mexico) will be the importance of human capital accumulation for
selected from the lower tail of the ability distribution success in the U.S. economy. Immigrants coming from
whereas immigrants from countries with less income less economically developed countries may have less
inequality than U.S. will be selected from the upper transferable skills because the formal education and
tail of their countrys ability distributions (Duleep 1997 work experience from their native countries may be
p.1). In other words, some of the immigrants from less applicable to the U.S. In fact, scholars have
less developed countries like Mexico have a large argued that the skills transferability is higher among
proportion of less skilled immigrants who may have economically developed countries because of the
less transferable skills. These skills may reduce similarity of their industrial structures and in the types
productivity and question their ability. In addition, of skills that are rewarded in their labor markets,
awareness of the generous government social (Duleep 1997). However, Duleep adds that the skills
programs makes economic conditions in the U.S. acquired at given levels of schooling in other less
attractive to unskilled immigrants. Since the U.S. taxes developed countries are not necessarily inferior to
the most able while subsidizing the least productive, those acquired in the U.S. and may in fact be superior
taxes are transferred in the form of welfare to the less (1997).
fortunate, who may be unqualified immigrants. On the other hand, Chiswicks hypothesis
provides a different view. Chiswick argues that
B. Human Capital Theory immigrants initially earn significantly less than native-
On the other hand, in an empirical test, Borjas born Americans because their skills are less
(1994) finds that the extent of income inequality of transferable to the U.S. economy. After acquiring
source countries is negatively associated with the U.S. - specific skills their earnings growth surpasses
relative quality of U.S. immigrants as measured by native-born Americans. The initial investment in
the wage differential between entering immigrants and acquiring U.S. - specific skills by immigrants reduces
natives of the same educational level (Duleep 1997). their initial earnings. However, after a substantial initial
This shows the role of human capital in earnings. The investment, greater earnings will be attained. Thus, it
human capital model by Gary Becker (1971) states will just be a matter of time before less-skilled
that the skills and qualifications enhance a workers immigrants climb out of poverty to a middle class
productivity and increases the value of that worker to status, if they accumulate the necessary human capital.
the employer. These skills and qualifications, which In addition, Chiswick attributes this increase
contribute to ones productivity are referred to as in immigrant earnings to positive selection of
human capital, (Van Dyke, 2000). Consequently, immigrants from source economies. Positive selection
as the worker contributes to the employer through implies that only the most motivated individuals move
high profits, the employer in return rewards the worker in response to economic opportunities. In fact, in
through high wages. That is, a decline in the skills and one of Chiswicks research papers, he finds lower
The Park Place Economist Volume X 72
The Truth About the Impact of Unskilled Immigrants in America
initial earnings but higher earnings growth for and argues that unskilled immigrants start off with very
immigrants from non-English speaking countries as low wages, but eventually pick up after acquiring U.S.-
compared with immigrants from English-speaking specific skills. I will try to find out what factors cause
countries (1978). This signals factors other than the this increased growth by looking at education variables
ability to communicate in English at play in earnings. and other factors.
In another study, Chris Minns (2000) quotes Chiswick
as saying that second-generation immigrants IV. Empirical Model and Data
outperform native -born Americans, possibly due to The purpose of this research is to analyze the
the transmission of positive characteristics from earnings of a sample of young immigrants and natives
immigrant parents to their children. Therefore, adults drawn from the National Longitudinal Survey
immigrant children also have positive characteristics of Youth (NLSY, 1997). The immigrants in the sample
that will enable them to outperform Americans. migrated to the US with their parents. For the purpose
Moreover, the fact that immigrant children are better of this study, unskilled natives and immigrants consist
integrated into American society suggests that they of NLSY sample members who had 12 or less years
will perform well in the economy. Minns (2000) also of education. Thus, the sample consists of only
finds that new immigrant earnings convergence (using unskilled natives and immigrants. Their earnings are
data from the 1900 and 1910 census files) achieved then studied over a 12-year period.
by foreign-born immigrant cohorts was attained by To test the change in earnings of unskilled
advancing from blue-collar jobs to white-collar jobs. immigrants and natives, I test the hypothesis that the
This also supports the skills-transferability model since unskilled immigrants eventually earn more than natives.
immigrants moved from blue to white-collar jobs after To test this hypothesis, four independent regressions
obtaining the necessary skills. are run for 1986, 1990, 1994, and 1998. Due to
Chiswicks self-selection argument is, data incompleteness, the regressions can not be run
however, challenged by George Borjas. Borjas claims for the entire 12-year period. Simple OLS regressions
that self-selection may be either positive or negative, are used to capture the earnings of both immigrant
depending on the relative wage dispersion in the source children and natives in the years chosen. The
and host countries (Minns, 2000). Borjas argues that regressions are run with data from the National
evidence supporting rapid wage growth among Longitudinal Survey of Youth (NLSY, 1997). The
immigrants may be an illusion caused by declining NLSY is a database that is derived from in-person
labor-market quality between successive cohorts. interviews with 12,686 people from 1979 through
That is, comparing earnings between immigrant cohorts 1998.
without considering the level of human capital between Two empirical models are used to test the
the cohorts will make earnings seem to grow at a rapid earnings change of unskilled immigrants and natives.
rate, which is not necessarily accurate. Similarly, The first model analyzes the earnings of both skilled
Robert Higgs (1971) found that there was little and unskilled natives and immigrants for each of the
difference in earnings between native-born and four years. Even though this model does not
immigrant workers after controlling for differences in accurately measure the effect of being an immigrant
literacy and English language ability (Minns, 2000). on earnings, it gives an idea of immigrants earnings
Higgs finding parallels Chiswicks convergence relative to natives during the regression years. The
hypothesis. model is as follows:
More recently, Borjas concludes that the
economic gap between immigrants and natives does E = a+ b1 IM+ u (Model 1)
not narrow substantially during the immigrants
working lives (Borjas, 1999). That is, the wages of In the second model, the growth in immigrant
less skilled immigrants remain below native earnings. earnings (E) is the dependent variable and it is tested
This view contradicts Chiswicks analysis. One against factors like age (AGE), gender (MALE),
possible explanation may be the new Information Age, parents education (MED & FED), and a dummy
which requires at least a high school degree and some variable, immigrant (IM). This model consists of
working computer knowledge. This papers individuals with twelve or less years of education.
hypothesis parallels the work of Chiswick and Minns Please refer to Table 1 for a summary of the variables.
73 The Park Place Economist Volume X
Sam Kwainoe
All the regressions are run in each of the four years to capture this effect with this variable. Thus, the
find the variables that have significant coefficients, to coefficient for age is expected to be positive.
explain the change in earnings. The regression equation The gender variable (MALE) is also a dummy
and model take the form: variable assuming a value of one for males and zero
for females. It is expected that males will earn more
E = a + b1 IM + b2 MED + b3 FED+ + b4 MALE + than females because males normally work more and
b5 AGE + u (Model 2) can have multiple jobs. Also, since women may be
discriminated against because they are perceived to
The Earnings (E) variable refers to the total be less productive than men, women will experience
annual earnings of each individual in the sample. The less earnings with time compared to males. The error
higher the earnings, the better off the person is. It term is (u). A summary of the variables are given in
symbolizes the economic well being of the individual. Table 1.
It is hypothesized that this variable will increase with Before running the OLS regressions,
more education, experience, and age. descriptive statistics were run to get the average
The immigrant (IM) variable is a dummy earnings for the individuals, as well as the mean years
variable that assumes the value of one for immigrants of education for the individual and their parents. The
and zero for Americans. Immigrants consist of sample used in this analysis consists of both skilled
individuals born outside the U.S. and had foreign unskilled natives and immigrants. The results are
parents. This effectively removes the probability of shown in Tables 2 and 3. This is done to see the effect
selecting Americans born outside the U.S. from the of parents education on the earnings of their children.
sample. This variable will allow the regression to test As shown in Table 2, immigrant parents in
the convergence hypothesis. By looking at the the sample have almost four years less education than
coefficients between 1986 and 1998, the earnings of native parents. Perhaps this explains why the mean
natives and immigrants can be compared to see if they earnings of immigrants are less than natives in Table
increase, decrease or remain the same. If the 3. Table 3 provides the mean earnings of individuals
coefficients are similar or the coefficient for immigrants and the difference in means between natives and
is positive, then the convergence hypothesis holds. immigrants in the sample. In all of the years analyzed
In addition, if an immigrant earns more than a native, in table C, natives earn more than immigrants. Also,
the coefficient will be positive and significant and vice immigrants have one year less education than natives.
versa. This could contribute to the lower earnings. Since
The parents level of education was included low levels of education decrease the level of human
into this research to study the impact of parents capital accumulation, it is inevitable that immigrants
education on childrens earnings. It is made up of would earn less than natives. Therefore, in order to
fathers education (FED) and mothers education analyze the effect of immigrants and their parents
(MED). It is assumed that, ceterus paribus, native education on earnings, model 2 is used to control for
parents will have more education than immigrant education. The results of the regression are discussed
parents, which will enhance their childrens economic in the ensuing section.
performance. That is, parents with high levels of
education will help their children with schoolwork, V. Results
encourage them to read, and use their work-related
connections to help their children find good jobs. A. Model 1
These things enhance the human capital of native The results from Table 4 indicate that
children and allow their earnings to increase. immigrants earned less than natives except in 1990
Therefore, the coefficient for the immigrants parents when they earned approximately $126 more than
level of education will be positive. natives. This supports Borjas argument that the wages
The older the immigrant (AGE), the higher of less skilled immigrants remain below native earnings,
his or her wage is expected to be. This is because approach native earnings, but never cross. However,
human capital increases with age, which increases the significance levels for all the regressions were low,
productivity. Therefore, an increase in age will increase which leads me to question Borjas claim. To better
earnings by a significant amount. This paper tries to
The Park Place Economist Volume X 74
The Truth About the Impact of Unskilled Immigrants in America
TAB LE 1
Variable Descriptions
Variable D e s cription Expe cte d Sign
E The Earnings variable refers to the total annual earnings of each individual in +
the sample.
IM The immigrant variable assumes the value of one for immigrants and zero for +
native- Americans.
FED Fathers education +
MED Mothers education +
AGE The age of the individual in the sample +
Male Gender of the individual in the sample. It assumes a value of 1 for males and +
0 for females.
understand the immigrant earnings question, several and hard-working than natives. In addition, immigrant
variables are added to model 1 to obtain model 2. children are as hungry as their parents, who may or
may not have a high level education. Therefore, they
are inspired by the poor economic status of their
B. Model 2 parents to work harder to earn more than their
The regression for model 2 produces some parents. Also, in some cultures children are expected
rather astonishing results as shown in Table 5. The to take care of their parents and extended family, so
results run contrary to Borjas argument that unskilled immigrant children are under more pressure to
immigrants earn less than natives and can never catch succeed. This, perhaps, explains the rapid change in
up to them. The regression produced results that the young immigrant adult earnings.
show that unskilled immigrants earn more than natives, The regressions for 1986, 1990, 1994, and
after controlling for education and age! 1998 accounted for approximately 13 percent of the
In 1986 immigrant adults earned about $458 variation in earnings and were highly significant.
more than natives, ceteris paribus. After controlling Obviously other factors like experience, language
for other factors in the regression, the high wage for skills, discrimination, and quality of education that were
the unskilled immigrants was also observed in 1990, omitted could have explained the variation in earnings
1994, and 1998. Unskilled immigrants earned more. This explains the low coefficient of variation
$3,061, $2892, and $4,999 more than unskilled values (R square).
natives respectively. In fact, immigrants in 1998 saw The male coefficient experienced tremendous
a ten times increase in earnings compared to natives growth from 1986 to 1998. The effect of being a male
in 1986. Does this mean that unskilled immigrants as compared to being a female on earnings caused
can survive in the new information economy? From earnings to double, looking at earnings in 1998 and
the significant increase in earnings between 1986 and 1986. The effect of being male on earnings increased
1998, unskilled immigrants may be complements to rapidly between 1986 and 1990 (from $4528 to
skilled labor in the new information economy. Skilled $16086). The significance level was also very high.
immigrants need unskilled immigrants to perform jobs Some explanations for this observation may be that
that they would not necessarily perform and would males tend to work more because of family
rather provide supervision for. Furthermore, the responsibilities and perhaps employers view males as
dramatic increase in earnings could also be due to the being more productive than females.
booming economy in the 1990s. The parents education on the other hand
The literature explains the growth in earnings showed the impact of parents education on the
with different reasons. Chiswicks argument that individuals earnings. A one-year increase in parents
immigrants are positively selected is one explanation. education (MED & FED) in 1986 increased earnings
What this means is that immigrants are more motivated by $423. However, in 1998 an increase in education
75 The Park Place Economist Volume X
Sam Kwainoe
TAB LE 2
Descriptive Statistics for Mean Years of Education
N (s ize ) M EAN M EAN M EAN M EAN
(1986) (1990) (1994) (1998)
M othe rs e ducation
Immigrant 597 7.2 7 .2 7 .2 7 .2
N ative 8526 11 11 11 11
Fathe rs e ducation
Immigrant 555 7.9 7 .9 7 .9 7 .9
N ative 7739 11 11 11 11
R e s ponde nt's Education
Immigrant 577 11.5 11.8 12 12 .2
N ative 8343 12 .7 12 .9 13 13 .2
Source: Nat ional Longit udinal Surv ey of Yout h (NLSY, 1997)
TAB LE 3
Mean Earnings for N atives and Immigrants and t- tests for equality of means of samples
Sample M e an Sig. Sample M e an Sig.
s ize e arnings s ize e arnings
1986 1994
Immigrant 8856 886 0 .2 4 5 583 18 5 2 3 0 .18 1
N ative 8273 9329 7777 19 7 2 0
Difference in means 443 1197
1990 1998
Immigrant 553 15 7 5 1 0 .8 5 1 503 24847 0.6
N ative 8 0 18 18 6 2 4 7252 25488
Difference in means 2873 641
TAB LE 4
Effect of Being an Immigrant on Earnings (Regression Results for Model 1)
19 8 6 19 9 0 19 9 4 19 9 8
IM - 462 12 6 - 1197 - 641
Sig 0 .2 5 0 .8 5 0 .18 0 .6
Sample size 8859 8571 6737 6420
Source: Nat ional Longit udinal Surv ey of Yout h (NLSY, 1997)
TAB LE 5
Effect on earnings after controlling for certain variables (Regression results for model 2)
Variable 1986 s ig 1990 s ig 1994 s ig 1998 s ig
IM 4 5 8 .7 3 * 0 .2 98 3 0 6 1.8 * * * 0 .0 0 0 2 8 9 2 .2 7 * * * 0 .0 00 4 9 9 9 .3 8 * * 0 .0 0 4
M ALE 4 5 2 8 .0 1* * * 0 .0 00 8 6 7 3 .8 * * * 0 .0 0 0 11231.3*** 0.000 16 0 8 6 .3 5 * * * 0 .0 0 0
M ED 2 19 .2 * * * 0 .0 00 4 8 6 .7 * * * 0 .0 0 0 6 5 3 .6 4 * * * 0 .0 00 10 19 .4 2 * * * 0 .0 0 0
FED 2 0 4 .7 1* * * 0 .0 00 5 3 8 .7 7 * * * 0 .0 0 0 7 6 7 .4 9 * * * 0 .0 00 9 2 0 .6 7 * * * 0 .0 0 0
AGE 10 4 2 .0 3 * * * 0 .0 00 7 6 6 .6 6 * * * 0 .0 0 0 6 6 8 .0 1* * * 0 .0 00 6 16 .7 3 * * * 0 .0 0 0
Sample s ize 7293 7058 6842 6441
R2 0 .13 4 0 .13 0 .13 0 . 14
NOTES: * signif icant at t he .1 lev el ** signif icant at t he .01 lev el *** signif icant at t he .001 lev el
Source: Nat ional Longit udianl Surv ey of Yout h (NLSY, 1997)
References