Do Immigrants Crowd Disadvantaged American Natives Out of Higher Education?

Although the idea that higher education is an obstacle to be overcome by those who are born disadvantaged in order to attain the full range of income, professional, and social opportunities in the United States, this notion seems to motivate various policies designed to raise access to such education for the disadvantaged. Some of the known policies include Pell grants, federal tuition subsidies for low-income pupils, and affirmative action, admissions preferences for racial and ethnic minorities. Other programs include financial assistance and counseling from the institution, remedial schooling, racial and ethnic support groups financing, and state and federal student loans.

Currently, these policies are common to both disadvantaged American natives and immigrants to the United States. Policymakers should determine the extent to which immigrants force disadvantaged American native out of attaining higher education. There are two likely ways by which this crowding out could occur:

1. The two groups merely fight for scarce access resources, like grants for underprivileged pupils.

2. Through affirmative action, since its focus does not necessarily distinguish between American natives and nonnative belonging to the same racial or ethnic group.

Although there have been several studies on the competition between immigrants and disadvantaged natives for low-earning occupations and social benefit programs, there is also a need to explore the competition for access to college education. Since very little is known on whether immigrants do force disadvantaged native out of higher education, and there is no empirical evidence as of yet, one may argue that the lack of confirmation only indicates that the impact of immigrants to higher education is too insignificant to affect anyone’s college access. However, this is a mistake anchored on the perspective of middle-class, white natives – in which their viewpoint of both disadvantaged natives and immigrants in higher education is little. The fact is it is the scale of two groups in relation to one another that is important.

The lacking evidence is due to the inability of reporting colleges and universities to deal with this question.  Although institutions of higher education annually provide ample information on enrollment they do not have distinction between the natives and non-natives in terms of enrollment by race, financial support, and curriculum. The National Postsecondary Student Aid Study (NPSAS), a survey used in this study, would help deal with the question.

The little evidence available is from the case studies of fourteen institutions accomplished by the Rand Corporation and reported in Gray, Rolph, and Melamid (1996). The GRM case study method exposed some of the mechanism through which crowding out may occur, such as the Equal Opportunity Support (EOP) programs that furnish academic and financial aid to “educationally and economically disadvantaged populations” – a criteria easily met by numerous immigrants due to their poor language skills which makes them lag behind on access tests in reading and writing. Furthermore, some faculty are concerned that the remedial reading and writing classes are completely reconstructed afar the needs of natives and to the needs of foreign-born pupils, whose problems are based in their limited English proficiency. Another mechanism is the crowding out of black and Hispanic natives in affirmative action targets by black and Hispanic pupils from privileged Caribbean and Latin American families.

Moreover, many are concerned that immigrants monopolize academic and financial support programs not only through direct competition for time and resources but also through transferring interest to a set of problems different from those of the disadvantaged native students. On the contrary, some staffs have observed that due to the immigrants’ aggressiveness in pursuing aid, many spillovers are offered to the natives by drawing in more of the college’s attention and monies towards the needs of the disadvantaged. Evidence is needed to determine whether the portion of the pie for the disadvantaged increases considerably that natives’ already smaller share is not a smaller piece.

The primary empirical challenge for this research is that the fraction of pupils in a college who are foreign born might be correlated with the college’s unobservable attributes that independently impact the native-born students’ enrollment.  Although, NPSAS aids this study since it connects a remarkable collection of student surveys, institutional reports, and individual student records and forms on file with colleges and the federal government, NPSAS is not actually designed to deal with this question. Thus the empirical evidence shown in this paper is not definitive, but exploratory. The study observes considerable evidence of displacement through access resources and affirmative action to merit more exploration.

Furthermore, the evidence of crowding out does not imply whether it is better to expand college access to disadvantaged natives or immigrants, instead it aids through providing information on how the natives’ college-going behavior relies on whether immigrants can utilize the college access programs. However, unless what is the college access programs supposed to do to maximize social welfare is actually identified, it cannot be determined whether the interaction of immigrants with them is good or bad.

I. A Model of the Relationship between Immigrant and Native College-Going

The key concern of the study is on how the natives’ college-going behavior is impacted by exogenous increases in the number of foreign-born people who are likely to be pupils of American colleges. The goal of the model presented in this study is to explain with sufficient accuracy to clarify their significance for empirical evidence the mechanism by which these impacts might operate. 

 The interest is on who is enrolled in college for a first degree. To resolve this a reduced-form analysis is needed since for any given crowding out mechanism, the college-going behavior is an outcome of three simultaneous stages: (1) person’s choice to apply to a college, (2) the college’s acceptance and its assessment with regards to financial support for that person, and (3) the person’s decision to enroll after being accepted, wherein each stages are contingent on each other. Since a person’s decision relies on the expectation of other’s decisions, it cannot be significantly determined as to what stage at which crowding out happens.

In the crowding out model, although it is not required to differentiate among the three types of immigrants: foreign-born (1) who have become naturalized U.S. citizens, (2) who are U.S. residents, and (3) who are undocumented or on student visas, considered as “nonresident aliens”, the study includes this factor in the equation for those policymakers interested to weigh the welfare of the three groups.

A. The Basics

The basic higher education model in this study does not include crowding out and pupils are expected to be capable of estimating their possible net benefits from college attendance, given their own abilities and qualities and the college’s policies towards tuition, financial support, curriculum and so forth. The calculation is from a standard Becker-Rosen model of human capital investment, given that the colleges’ policies are known.

The calculation of the likely gross benefits from a student of a particular college includes the tuition to be paid, the federal or state monies that would be brought with him, and the net present value of any direct donations, and so on. Meanwhile, the gross costs of having the individual as a pupil are a function of that person’s academic and social necessities and any costs related with the impact his presence would have on the college’s size. However, the supply of places at a certain college is not completely elastic since additional pupils crow the campus and reduce other pupils’ benefits. Furthermore, colleges typically have disproportionately high expanding costs in relation to keeping the current size, since their place and locations are historically determined and protected from the real estate market by their nonprofit status.

Moreover, the calculation does not incorporate any benefits the pupil might provide a college merely because he is a member of a disadvantaged group whose admission to higher education has been historically inadequate.

B. “Benchmark Crowding Out”

If the group of likely applicants or a particular college expands due to an inflow of immigrant applications then some crowding out of native pupils will occur. Native pupils, particularly those having low net benefits, will be the first to be forced out, and the displacement will seem as a rise in the college’s admissions standards.

Such is a “benchmark crowding out”, which could likewise occur, for example, if social change encouraged women having previously low college-going rates. Note that in the long term, since the supply of college places can ultimately increase via accumulation of marginal, less-selective colleges, the benchmark crowding out can be explored at individual colleges, but not generally.

Meanwhile, the immigrants increase the admission standards for all and natives who are disadvantaged are potentially more marginal pupils. For a particular level of admissions standards, the fraction of immigrants in the student body has no impact on the fraction of disadvantaged natives.

However the paper does not only focus on the benchmark crowding out, but also test for other crowding-out mechanism.

C. Crowding Out Through Competition for Access Programs

A system of rewards and penalties for providing access for the disadvantaged might exist due to the view that society may benefit more than a college does from a disadvantaged pupil attending college. This would happen if some of the societal benefits are overflows that are not incorporated by the college. Some of the rewards and penalties are from governments and some are from private donations.

The model indicates that the society does not specifically prefer colleges to be completely comprised of disadvantaged pupils since one of the benefits of college attendance is the interaction with more advantaged pupils.

Since the college has more motivation, it raises the target portion of the college’s resources allotted to disadvantaged pupils greater than when there is no reward for offering access programs. Hence, a college’s expected net benefits are more positive for disadvantaged pupils – the portions of underprivileged pupils are greater and the access programs accomplish some of their aimed impacts.

Taking into account the surge of immigrants who potentially qualify as disadvantaged, the net benefits from college experienced by disadvantaged natives are disproportionately affected, since they induce influx into access programs and reduce the benefits accessible to possible native pupils. The empirical expectation is that the more underprivileged are the immigrants, the greater a given number of them forces out underprivileged native pupils – such implication is contrary to what the benchmark model suggests. 

D. Crowding Out Through Affirmative Action Targets

Consider the situation, that rather than a system of payoffs and sanctions for furnishing access programs, the college is punished for not satisfying its affirmative action targets in relation to racial or ethnic groups.

Although pupils generating negative net benefits for the college might still be accepted by that college, those pupils are likely not to enroll unless they have positive net benefits from attendance. Hence the affirmative action target accomplishes few of its objectives of having larger number of underrepresented individuals to attain positive benefits from higher education. 

Given the presence of immigrants, who are likely to be foreign-born, yielding positive net benefits for the college, and meet the qualification of being in a disadvantaged group, then with affirmative action targets not being country-of-birth specific, the college will not accept several underrepresented natives who would otherwise be accepted. Rather, its foreign peers would be the one admitted.

The empirical prediction is that holding the college’s admission standards constant, the portion of native pupils belonging to underrepresented groups reduces the share of colleges’ overall enrollment that is foreign born, and what’s more is that each group of natives, black and Hispanic will be disproportionately impacted by a shift in the pool of likely immigrants from its own cohort. However such implication is not a prediction of the benchmark model of displacement.

II. Empirical Challenges

A. Remedying Omitted-Variables Bias and Other Source of Endogeniety

The potential sign for the omitted variables is positive since similar pupils go to the same colleges, thereby underprivileged natives and immigrants gather in the same colleges. On the other hand, negative omitted-variables bias is less likely since it would only happen when similar native- and foreign-born pupils have varying preferences. Fortunately, this concern doesn’t arise since the data provides accurate description of individuals.

Since the sources of omitted-variable bias are mostly positioning variables as well as institutional variables, therefore several covariates are included that reflect the college’s location, such as state indicator variables, center-city campus, urban campus, metropolitan-area campus and rural campus, and also shows college’s selectivity, control type (public or private) , size of enrollment  and degree-granting programs.

In spite of these covariates, omitted-variables bias may still be present. Particularly for colleges that gain from a localized area since pupils likely to belong to the same neighborhood may have the same unobserved attributes. Therefore, omitted-variables bias will most potentially lead to observing no crowding out among colleges having low selectivity.

The best way to remove such bias and other sources of endogeneity in the independent variables is to look for an explicit cause of possibly exogenous difference in the share of foreign-born pupils in a college.

B. Remedying Division Bias Due to Measurement Error

There are two distinct biases associated to measurement error. First is the attenuation bias or the bias against observing a statistically significant impact of either sign, since the share of foreign born pupils in college is estimated with error. Such bias may prevent observing evidence for crowding out even if it were quantitatively significant.

Second is the division bias which happens when a variable that measured with error is used to estimate both the dependent and independent variable. This bias could cause the estimates to understate the true value of crowding out.

The solution for both is instrumenting each college’s 1992 share of foreign-born pupils by its 1989 share of foreign-born pupils. Such remedy will generate unbiased estimates given that the sampling error in one year is uncorrelated with the sampling error in another year.

III. Data and Descriptive Statistics

The ideal data for the study would be longitudinal which follows pupils from prior their college-going decisions. Nonnative pupils are also crucial in the sample given that the behavior of immigrants and natives influences crowding out to occur. There is also a need for curricular and financial support information on the pupils. Lastly, the survey is required to be connected to colleges’ institutional records, particularly the partition of college enrollment into immigrant, race, ability, and income groups.

Unfortunately, the chief longitudinal surveys do not satisfy these conditions since their sample is confined to pupils who attended junior high or high school in the United States. Furthermore, the amount of foreign-born pupils is far too few and not amply representative for dealing with crowding-out questions. Meanwhile the main panels of data from college and university records do not breakdown the enrollment by immigration categories. Lastly, the population census/surveys do tend to yield college enrollment rates that are perceived to be erroneous in comparison to colleges’ administrative records.

The NPSAS meets some of the criteria, as it is conducted every three years since 1986, comprising of rough 52 thousand undergraduate sample and one thousand undergraduate organizations each year.

The survey magnitude is crucial since immigrants, racial and ethnic minorities, and disadvantaged pupils represent a small portion of the college population. Furthermore, NPSAS is the only microdata sample of higher education that consists of thorough immigration status which allows breaking down of foreign born into naturalized citizens, residents, and nonresident aliens

In addition, NPSAS student information is oddly precise since pupils’ administrative data were used to validate the student’s responses.

The NPSAS data were also matched to colleges’ institutional data for the equivalent school year from the Integrated Post-Secondary Education Data System. The entire sample of colleges, the correlation among the sample estimates and the population statistics were constantly above 0.95.

The primary limitation of the NPSAS is that it is not longitudinal for individual pupils

A. Immigrants in College

The study shows that in comparison to the whole native college student population, immigrants greatly use colleges’ access resources. Moreover, breaking down the immigrants into naturalized citizens, residents, and nonresident aliens, the use of access resources increases from the first one to the last one.  As colleges become more exclusive, they are more differentially charitable towards the fairly underprivileged pupils; but there are somewhat few of those pupils since most of them do not pass the selection process. On the other side, nonselective colleges have numerous disadvantaged pupils and small endowments; hence they cannot be differentially generous.

However considering the case “within” black or Hispanic pupils, and in comparison to these disadvantaged natives, immigrants do not seem that needy. In general, nonresident aliens appear to have the highest SAT scores, with parents having the highest educational attainment, also naturalized citizens seem to be the most economically and educationally disadvantaged group.  However, the education differences observed by the study seem to understate the degree to which the foreign born residents’ and nonresident aliens’ parents’ education surpass the natives’ parents’ education.

Finally, the study suggests that if the two identified medium of crowding out exist, displacement through affirmative action is expected to be more prominent in more selective colleges, while displacement through competition for access would be more observed in the center of the selectivity spectrum where there is institutional resources to be differentially allocated among pupils and there are numerous underprivileged pupils.

IV. Results

Note that the estimates are potentially biased in a positive direction merely due to the fact that similar pupils are likely to go the same school. Thus the result would possibly understate the true value of crowding out.

A. The Baseline: Results Using OLS Estimation on Cross-Section Data

The findings indicate that a percentage point rise in the fraction of nonblack immigrant college pupils reduces the fraction of black native pupils by some percentage points, wherein the size of the decrease grows as the college becomes more selective; while, for the least selective colleges, there is no evidence of crowding out.

Meanwhile for the Hispanic, there is only a very small degree of crowding out observed across ethnic groups, in which most of the displacement happens through nonresident aliens. Moreover, for the most selective colleges, a percentage point increase in the share of foreign-born Hispanic decreases the share of native Hispanic by some percentage point.

Overall, the findings imply that displacement via affirmative action targets happens at colleges that are (very) selective, and seems to be mostly within racial/ethnic groups indicating that there is little substitution for such targets.

B. Basic Tests of Crowding Out Through Competition for Access Resources

In testing the affirmative action medium, the measure of the pupil’s underprivileged background is race and ethnicity. The study observed that the very selective colleges showed the most displacement, while the somewhat selective colleges showed lesser but still consistent crowding out. But for the two extreme categories, (Not-very- and Extremely-Selective), there is no evidence that foreign-born students crowd out disadvantaged pupils.

On another note, if foreign-born pupil transform remedial writing classes to address their own needs and force out disadvantaged native pupils, the classes would potentially be more appealing to foreign-born pupils. If such method, by which natives and immigrants contend with, does occur, then the initial rise in the share of foreign-born pupils taking remedial classes reduces the share of disadvantaged native pupils who have low-earning parents.

In summary, the more the immigrants appear similar to those who would need access resources, the more the disadvantaged natives are displaced. However, such crowding out seems to be irrelevant at colleges that are not selective. This may be due to omitted-variables bias, or due to the fact that these colleges have little access resources that come from the institution.

Most displacement through competition for access resources happens at colleges that are selective, (but not extremely selective) such as state colleges and universities, and good private colleges, but does not include community colleges.

C. Correcting for Attenuation and Division Bias with Instrumental Variables

The result of instrumenting for a colleges’ 1992 immigrant share by its 1989 immigrant share, indicates that crowding out through access resources happens at colleges with middle-level of selectivity, while crowding out through affirmative action happens at colleges with high-level of selectivity. 

V. Conclusion

The study furnishes some for the first empirical evidence on whether immigrants force disadvantaged American natives out of higher education. Although neither immigrants nor disadvantaged natives form a huge chunk of the college-attending population, college education is significant for both groups to eventually change themselves into higher-skilled and better-integrated individuals. Both groups are considerably large compared to each other and relative to the resources provided for enhancing their access to college that immigrants do force disadvantaged native out of higher education.

In very selective colleges, crowding out occurs, for instance, black and Hispanic natives seem to be displaced by less disadvantaged foreign-born pupils who are still included in the same racial/ethnic affirmative action targets. Crowding out through affirmative action targets seems to happen mostly in colleges that are very selective.

On the other hand, crowding out through competition for limited access resources occurs mainly in selective colleges, but not extremely selective. Since such colleges’ discretionary funds are distributed over and stretched by the disadvantaged and immigrants.  The displacement does not occur much at nonselective colleges since their main access resources are not discretionary but are from federal government programs.

Despite the fact that the empirical evidence on crowding out cannot tell policymakers whether access resource or affirmative action must or must not differentiate between natives and immigrants, still it can offer policymakers an idea as to whether their choices have trade-offs between the groups. Lastly, the empirical evidence should also aid the federal government in deciding on how colleges could meet affirmative action targets. 

Source:

Caroline M. Hoxby, “Do Immigrants Crowd Disadvantaged American Natives Out of Higher Education?” In Daniel S. Hamermesh and Frank D. Bean (Eds.) Help or Hindrance, New York: Russell Sage Foundation, 1998, pp. 282-321.