ILW.COM - the immigration portal Immigration Daily

Home Page

Advanced search

Immigration Daily


Processing times

Immigration forms

Discussion board



Twitter feed

Immigrant Nation


CLE Workshops

Immigration books

Advertise on ILW

VIP Network


Chinese Immig. Daily


Connect to us

Make us Homepage



The leading
immigration law
publisher - over
50000 pages of free

Immigration LLC.

< Back to current issue of Immigration Daily < Back to current issue of Immigrant's Weekly

Testimony Prepared for the Subcommittee on Immigration and Claims

Hearing on "The U.S. Population and Immigration"

Committee on the Judiciary

U.S. House of Representatives

August 2, 2001


Jeffrey S. Passel and Michael Fix

Immigration Studies Program

Urban Institute

Washington, D.C. 20037

The 20th century began with the country in the midst of the greatest wave of immigration in its history. The century ended in the midst of another period of high immigration, greater in numerical terms but smaller in its relative impact than the immigration of 100 years earlier. The issues raised at the turn of the 21st century parallel those of the earlier wave: Can the country accommodate the new immigrants? Who benefits from the arrival of the immigrants? Who is harmed? Can the immigrants be absorbed and integrated or are they simply too "different" from the rest of the country? Will the country change as a result of the immigrants, and how?

In the past, such questions proved difficult to answer on a contemporaneous basis and the situation is not substantially different today. A growing body of research can be brought to bear on these issues however. In this document, we draw on our own research and that of our colleagues as well as scholars at the Urban Institute and elsewhere to assess some of the impacts of immigration on the United States. In the first section, we examine some of the basic demographic trends regarding the scale of immigration, its pace and characteristics. The results from Census 2000 have called into question some of the basic information regarding immigration that until six to nine months ago was widely agreed upon. The surprise figures from the Census suggest strongly that immigration levels, particularly undocumented and temporary immigration, are substantially higher than most had suspected.

The next section addresses some of the basic socioeconomic characteristics of immigrants living in the country. Questions of integration and adaptation of the new immigrants involve not only the immigrants themselves, but their native-born offspring. The U.S.-born children of this newest wave of immigrants, which is traditionally deemed to have begun in 1965, are reaching adulthood so that we can examine how 20 to 35 year-old children of immigrants have fared to date. While the available evidence is mixed, there are some encouraging signs. Considerations of immigrant adaptation lead naturally to the consideration of the country's minimal immigrant policies, the costs and benefits of immigration, and the potential need to explicitly design integration programs and policies on a larger scale than is currently done. The final section draws heavily on work done by Fix and Wendy Zimmermann to address immigrant integration. We close with the common plea of researchers for more, and better, data. While there is a certain pro forma character to this discussion, mismatches between policy and need that we and others have observed, as well the sudden "appearance" in national data of 3 million Hispanics and 1 million Asians, probably resulting from the mismeasurement of immigration flows highlight the strong need for improvement in this area.

Immigration Trends

During the 1990s, more immigrants came to live in the United States than in any decade in the nation's history (Figure 1). In-flows have steadily increased since the 1930s and have more than tripled in the last generation. The growth in immigration has been driven in part by legislative increases in legal admission ceilings in 1965, 1976, and 1990. Further, the acceptance of political refugees from various parts of the world has contributed to the diversity of sources and rising flows.

Not only did legal immigration flows increase steadily, large-scale undocumented immigration began in the 1970s and gradually increased since, with only a slight diminution surrounding the enactment of IRCA in 1986. The uncertainty surrounding the magnitude of immigration flows during the 1990s is directly attributable to the difficulties of measuring undocumented immigration (discussed below).

A direct consequence of large and increasing entries of immigrants is that the share of the U.S. population that is foreign-born has also grown. In 1970, the foreign-born population numbered slightly less than 10 million and accounted for less than 5 percent of the population. (See Figure 2.) Although this percentage was anomalously low, it represents a significant point in the current collective memory of the country. There were relatively few immigrants and they were disproportionately elderly (and their numbers decreasing from mortality).

More typical of the U.S. historical experience, however, was the situation at the beginning of the 20th century. At that time, about one-seventh of the U.S. population was foreign-born, as was the case from 1870 to 1930. More recently, since the low of 4.7 percent foreign-born in 1970, the percentage foreign-born has more than doubled to almost 11 percent as the number of immigrants living in the country more than tripled to just over 30 million. This very rapid change occurred in the space of one generation as the percentage foreign-born approaches the levels experienced at the height of the 19th century wave of immigration.

Dispersal. The very large numbers of immigrants arriving and the large foreign-born population means that the impact of the new immigrants is being felt in areas outside those traditionally settled by immigrants since even a small percentage of the increasing foreign-born population can lead to noticeable local populations. However, a new settlement pattern has emerged that has resulted in sizable new immigrant communities outside the traditional receiving states.

Throughout much of the 1970s, all of the 1980s, and into the early 1990s, almost 75 percent of newly-arriving immigrants settled in just 6 states (California, New York, Texas, Florida, New Jersey, and Illinois). About one-third settled in California alone. In the late 1990s, however, this pattern changed. California still received the largest numbers of new immigrants, but only about 22 percent of the total settled there, instead of the more usual 33 percent. The immigrants who would have gone to California did not settle in the other large immigrant-receiving states as they continued to get about 40 percent of new arrivals. Instead, they went to a swath of states across the middle of the country stretching from Oregon to Arizona to Iowa and Arkansas to Georgia, North Carolina, and Virginia. (See Figure 3.) The foreign-born population in these states grew twice as fast during the 1990s as it did in the more traditional immigrant receiving states.

This dispersal appears to have been led by immigrants, especially Mexicans, initially moving out of California and other traditional settlement areas. Later, as the new communities were established, they attracted new immigrants arriving directly from outside the country. The immigrants, many of whom are undocumented, appear to have been drawn by readily-available, low-wage employment, and affordable living conditions. The results of this redistribution appear, at this point, in Census 2000 as very rapid growth of the Hispanic population (and to some extent the Asian population). More definitive analyses must await the full release of sample data from Census 2000.

Legal Status of the Immigrant Population. The new laws relating to immigrants passed in 1996, especially welfare reform, brought into focus some of the basic deficiencies of the nation's immigration data systems. This law limits access to a range of benefits for many categories of non-citizens, some of whom had formerly been eligible and receiving benefits. For example, legal permanent residents who have not worked 40 quarters in the United States and refugees who have been in the U.S. for more than 5 years are no longer eligible; naturalized citizens remain eligible. In addition, whereas eligibility rules governing aliens (legal and undocumented), had been the exclusive province of the federal government, the welfare reform law devolved responsibility for setting many of these eligibility rules, such as those covering undocumented aliens, to the states. Since the financial responsibility for providing services to the groups no longer eligible for federally-provided coverage was also devolved to states and localities, a number of new parties became interested in the numbers of naturalized citizens and aliens of various types, and in rates of naturalization; moreover, such data are needed for states and even smaller governmental units. Many of the population numbers needed are available only as rough estimates or simply do not exist.

There are five main legal statuses pertaining to immigrant populations that are of interest to most observers. In roughly decreasing order of size, they are:

Legal permanent residents;

Naturalized citizens;

Undocumented aliens;

Refugees, asylees, and parolees; and

Legal nonimmigrant residents.

In addition, for some purposes, many users want data for subgroups of these major groups -- family-sponsored immigrants, employment immigrants, foreign students (a category of non-immigrants) to name just a few. Data are needed on the numbers in each category, the inflows and outflows each year (or at least the annual net change), and characteristics of the particular individuals. Interestingly enough, the only one of the five groups for which there are "official" estimates of size and annual net change is the one most would characterize as the hardest to measure -- undocumented aliens. Using a combination of demographic estimation techniques and survey data, we at the Urban Institute have developed some estimates of the populations in the key legal statuses, shown in Figure 4.

The largest group in the foreign-born population is legal aliens, or aliens admitted for permanent residence (LPRs). There are about 9.3 million LPRs as of 2000, representing about 30 percent of the roughly 30 million foreign-born residents of the country. The LPRs, as defined here, include a number of different immigrant groups, but only those persons who are not U.S. citizens. The LPRs include regular family-based and employment-based immigrants; they also include aliens who acquired legal status under IRCA, the so-called SAWs and LAWs. LPRs, as defined here, do not include humanitarian admissions (since 1980), nor do they include a number of persons authorized to be in the country, but not on a permanents basis, such as asylum applicants, persons with Temporary Protected Status (TPS), and the "family fairness" immigrants. Largely as a consequence of high naturalization rates, the LPR alien population is has actually decreased in recent years.

The naturalized citizen population is a rapidly growing segment of the immigrant population. We estimate that there are approximately 9.2 million naturalized citizens who moved from the LPR alien population to become citizens and another 0.7 million citizens who entered as post-1980 refugees. The naturalized citizens also account for approximately 30 percent of the foreign-born population. This population should continue to grow rapidly through 2002 or 2003 until the INS reduces the backlog of applications for naturalization.

Humanitarian immigrants who entered the United States since 1980 account for about 2.3 million or 7 percent of the foreign-born population. This group includes refugees, asylees, Amerasians, Cuban-Haitian entrants, and certain parolees. Approximately one-third of the humanitarian entrants since 1980 have become naturalized citizens.

"Nonimmigrants" are aliens admitted to the United States for specific, temporary periods, and for specific purposes. There are solid figures on annual admissions for nonimmigrants, but virtually no population figures because nobody keeps track of how many nonimmigrants depart or change to other immigration statuses. The largest group of nonimmigrants is probably foreign students and visiting faculty (F and J visas). Temporary workers (H-1s, H-2s, and Ls) constitute another large group. Smaller nonimmigrant groups include diplomats, treaty traders, au pairs, and the like. We estimate that about 1 million nonimmigrants are living in the United States at any given time, but recognize that there could easily be more, up to about 1.5 million.

Undocumented Aliens. In Figure 4, undocumented aliens are shown as representing about 8 million immigrants or 28 percent of the foreign-born population. This total is substantially larger than previous estimates from the Urban Institute and represents new work based on the early results from Census 2000. As such, it has a substantial range (at least 8-9 million) and should be treated as preliminary. It is also quite a surprising number, at least to most analysts who had been working with empirically-based estimates of this clandestine population.

Prior to October 2000, there appeared to be a fairly broad consensus regarding both the net flow of undocumented aliens into the United States and the total number living in the country. At the heart of this consensus was the work done by Robert Warren of the INS. He had estimated that there were 5 million undocumented aliens living in the United States as of October 1996 and that the average annual increase during the preceding four years was about 275,000 per year. Extrapolating from these numbers placed the undocumented population as of April 2000 (the census date) at roughly 6 million. Passel's work using the March Current Population Surveys (CPS) of 1995 and 1996 suggested that the rate of increase might be slightly higher and that the total number could be larger by several hundred thousand (Passel 2000). Nonetheless, in "round" numbers, the estimates were virtually identical. In addition, Warren had developed some new techniques that also relied on CPS data that yielded virtually identical results.

The first indications that the consensus view might be wrong came from the March 2000 CPS which became available in late fall of 2000. The total foreign-born population in the March 2000 CPS was 28.4 million, an increase of almost 2 million over the corresponding CPS figure for March 1999. This figure should not necessarily be treated as measuring the annual change in the foreign-born population, however. Annual change as measured by the CPS does not provide a very precise measure of change in the foreign-born population (i.e., net immigration) because the CPS measures can fluctuate substantially from sampling variability. Nonetheless, many researchers, including Passel and Warren, use CPS data to measure the size of the undocumented population, with measures of change coming from a series of measurements rather than year-to-year differences.

According to initial estimates, the March 2000 CPS implied about 6.8 million undocumented aliens included in the CPS, or roughly 7 million undocumented aliens in the country, rather than the widely-used 6 million figure. This new estimate raised a number of hard-to-answer questions about the current data and previous estimates. For example, had the flow of undocumented immigrants increased substantially over 1999-2000 or were the previous estimates too low? We don't yet have answers to the questions, but there are some indications from the data that both phenomena occurred. Specifically, publicity and outreach efforts associated with Census 2000 appear to have led to substantial increases in cooperation with the Census and the CPS on the part of undocumented aliens who had been in the country for a number of years and, thus, to the larger counts. In addition, the data from the 2000 CPS also suggest that the flow of undocumented immigrants over the last 4-5 years of the 1990s was significantly higher than the extrapolated trends from earlier in the decade.

Converting these observations into estimates requires assumptions that are not especially verifiable with the data now available. The principal "checkpoints" or key assumptions are: an undocumented population of 3.3 million as of the 1990 Census (or in the range of 3.0-3.5 million; net flows of 275,000 per year or more for the early part of the 1990s; an undocumented population of about 5 million "in the middle of the decade" (i.e., as of October 1996 according to Warren or about a year earlier according to Passel's estimates); and, an undocumented population near 7 million as of 2000. Reconciling these figures implies a net increase in the undocumented population of 450,000-500,000 per year for the latter part of the 1990s. Even assuming that the estimated population figures for the early 1990s are too low still implies at least 400,000 net undocumented immigration for the latter part of the 1990s.

Even these fragile new results are called into question by the results of Census 2000. The basic census count for 2000 of 281.4 million was 5-7 million higher than the demographers at the Census Bureau expected. Of particular interest for students of immigration are the figures for the immigrant-dominated populations - Hispanics and Asians. The 2000 count of Hispanics, 35.3 million, is about 3 million higher than expected based on previous estimates. Likewise the count for the Asian and Pacific Islander population of about 12 million is 10 percent or roughly 1 million higher than expected. Both Census Bureau demographers and outside demographers have surmised from these figures that the reason for the low estimates, relative to the census count, is that the Census Bureau underestimated the amount of immigration that occurred during the 1990s. The cause of the underestimation was principally, but not entirely, underestimation of the amount of undocumented immigration that occurred during the 1990s.

In addressing the questions surrounding the higher than expected population counts from Census 2000, Passel has done a preliminary analysis using data from the March 2000 CPS but reweighted to agree with the counts from Census 2000. The initial results, noted above, point to a foreign-born population in excess of 30 million which, in turn, implies that the undocumented population as of April 2000 was about 8 million. These results are based on numerous assumptions that cannot be verified until we have more data from Census 2000. They do help to reconcile the results from Census 2000, the demographic analyses of population change during the 1990s, and the results from coverage evaluation studies of Census 2000. In addition, if Census 2000 suffers from an undercount, even the small one suggested by the evaluation studies, it would push the implied numbers of undocumented aliens somewhat higher - to about 9 million.

Until we understand better the magnitude of the undocumented population and when they entered the country, it is even more difficult to determine the size of the flow of undocumented immigrants into the country. However, if the numbers described above are "in the ballpark," i.e., if there are 8-9 million undocumented aliens in the country, the annual increase in the undocumented population must be in excess of 500,000 per year and could possibly be higher for recent years. It should be stressed repeatedly that these figures are somewhat speculative. However, they do draw on what data we have available, namely CPS data from 2000 and the complete-count data from census 2000.

In discussing these figures, it is worth noting who is included in the estimates of the undocumented population. The definitions and numbers for the undocumented population arise from a set of standard estimation techniques. Basically, recent estimates rely heavily on residual techniques whereby an estimate of legal foreign-born residents is subtracted from a survey-based estimate of the total foreign-born population (from the decennial census or the Current Population Survey). The estimates of the legally-resident foreign-born population generally use the categories discussed above - i.e., LPRs, humanitarian admissions, and nonimmigrants. The estimated undocumented population thus includes all foreign-born persons who do not fall into these legal admission categories. It mainly includes clandestine entrants who sneak into the country, usually across the Mexican border and visa abusers, or persons who enter legally with a document permitting them to be in the United States for a specific period of time, but who then fail to leave.

There are, however, some other categories of aliens who may be entitled to remain in the United States for limited periods of time and are not actually deportable, but who are not included in the legal population estimate. These groups include so-called "family fairness" immigrants, asylum applicants, persons with temporary protected status (TPS), and the new K and V visa. These groups could account for at least 1 million and possibly 2 million persons or even more who are estimated as part of the "illegal" population who are not actually deportable, and who are on the road to becoming legal immigrants.

By far the largest group of undocumented immigrants is persons from Mexico who probably represent half or more of the total number. Other parts of Latin America - Central America (mainly El Salvador, Guatemala, Honduras), the Caribbean (e.g., the Dominican Republic), and South America (e.g., Colombia) - probably account for another one-quarter of the undocumented population. Visa overstayers may account for one-third or more of the undocumented population.

Immigrant Families. Determining and classifying the legal status of immigrant families is even more complicated than for individuals. For example, Fix and Zimmermann (1999) found that households headed by noncitizens are more likely to contain children than those headed by citizens (55 percent versus 35 percent). Moreover, since about four-fifths of the children of immigrants are U.S. natives, most of the noncitizen families contain persons with quite different legal statuses. In particular, 85 percent of immigrant families with children are so-called "mixed status" families - that is, families where at least one parent is a noncitizen and at least one child is a citizen.

The importance of these mixed status families can be easily overlooked, but the principal findings are quite striking. Nationally, about 1 in 10 children (among all children) lives in a mixed family (Figure 5) and three-quarters of all children in families headed by noncitizens are U.S.-born citizens. The proportions vary considerably across the country. In New York City, 27 percent of all children are in mixed status families; in Los Angeles, almost half (47 percent are). In New York state, 70 percent of families with children that are headed by undocumented immigrants contain children who are U.S. citizens.

Not only are these mixed status families demographically important, they are even more prevalent among the families affected by social welfare policy. Fix and Zimmermann found that the percentages of low-income children in mixed status families are even larger than among all children; 15 percent of low-income children are in these families. Moreover, 21 percent of uninsured children are in mixed status families nationwide and fully one-half in California. They argue that these mixed status families pose special issues for policymakers. The imposition of restrictions on benefits eligibility for noncitizens tends to affect citizen children. On the other hand, policies intended to extend benefits to noncitizen children are limited in their coverage because most children in immigrant families are already citizens.

Characteristics of the Immigrant Population

Immigration obviously has impacts beyond just the demographic ones described. Immigrants add to the U.S. labor force numerically and contribute their human capital. Concerns have arisen over labor market impacts of immigrants, both for their concentration in specific, low-skilled occupations and for their overall potential impact on native employment. While not addressing these concerns directly, we present information from our CPS analyses on characteristics of immigrants of different legal/admission statuses. The differences among the immigrant groups are substantial and, in many cases, greater than native-immigrants differences.

Educational Attainment. Overall, immigrants are much more likely than natives to have low levels of education, as 32 percent of the foreign-born population aged 25-64 has not graduated from high school versus only 11 percent of natives in the same age group. The data in Figure 6 do show that the recent immigrants have generally higher levels of education than earlier immigrants. In addition, differences from natives are much less for legal immigrants and even refugees than for all immigrants, as 27 percent of these groups have less education than a high school diploma. This compositional effect occurs because the undocumented population disproportionately has a very large proportion of persons with very low levels of education (58 percent with less than a high school diploma).

At the upper end of the education distribution, i.e., those with at least a Bachelor's degree, immigrants overall are just as likely as natives to have such a degree (27 percent for natives and immigrants), notwithstanding the overrepresentation of immigrants at lower levels of education. The undocumented group, however, is less than half as likely as native s to have a college degree (11 percent). Legal immigrants are somewhat more likely than natives to be well-educated (30 percent with college degrees) as well as more poorly educated, as previously noted. Those immigrants who have naturalized have even higher percentages with college degrees. These results show clearly that use of CPS data covering all immigrants to characterize the legal immigrants exaggerates the percentage of legal immigrants at lower levels of education because of the presence of sizable numbers of undocumented immigrants in the CPS data.

Income. Income is a major determinant of life chances and life style in the United States. While immigrants are generally thought to have lower incomes than natives, the reality is somewhat more complex. Almost all immigrant groups show income increases with increasing duration of residence in the United States. Thus, to some degree, the lower incomes of immigrants overall reflect the fact that many immigrants have not been in the United States long enough for their incomes to approach those of natives. In addition, there are sizable differences between immigrants groups with legal immigrants, especially LPRs and naturalized citizens, having higher incomes than undocumented aliens and refugees. The income differences that do exist between native and immigrant families tend to be exacerbated because immigrants have somewhat larger families. Thus, the family's income must provide for more people in immigrant families than in native families.

The education differentials noted above translate rather directly into income differences across native and immigrant populations. Thus, the mean income for foreign-born households ($45,400 in 1997) is 10 percent lower than the mean for native households ($50,200). Differences among immigrant groups are more substantial than this native-immigrant difference, however. LPR entrants have household incomes ($49,100) only 2 percent lower than native households, whereas refugee incomes ($37,100) are 26 percent lower. Undocumented immigrants have the lowest incomes of any group ($31,500) - more than one-third below those of natives.

The growth in income between newly-arrived immigrants (i.e., those in the country of 10 years or less) and those in the country longer (i.e., more than 10 years) is impressive. Among LPRs, the longer-term residents have incomes that are roughly the same as those of natives and 14 percent higher than their newly-arrived counterparts ($50,400 versus $44,000, in Figure 7). Refugees show a larger differential, 21 percent growth, but even the longer-term residents have incomes ($41,000) that are 18 percent lower than natives. The differences between LPRs and refugees reflect not only the education differentials between the groups, but undoubtedly also are a function of the more traumatic situation facing refugees in their home countries and the likely greater presence of support networks among LPRs in the United States.

Because immigrants can change status (for instance, LPR aliens naturalizing or undocumented aliens becoming LPR aliens), some income differences observed between groups is attributable to selectivity in the transition. For example, long-term naturalized citizens have higher incomes than long-term LPR aliens in part because the more successful chose to naturalize. As a result, the average income of long-term LPR aliens is not substantially higher than the short-term LPR aliens'. In a similar vein, there is little difference in incomes between short and long-term undocumented aliens. Part of the reason for the lower incomes of this group (beyond their low levels of education) is the necessity of working without valid documents. Any such wage penalty is not obviated by longer residence in the United States. Further, to the extent that undocumented aliens stay in the country for long periods, many eventually acquire legal status and thus pass into the LPR groups. As a result, those longer-term residents left in the undocumented category are likely to be among the least successful economically.

Change Across Generations. A full assessment of integration of the children of today's immigrants will require the passage of decades. However, The U.S.-born children of the earliest post-1965 immigrants are beginning to enter adulthood. Jennifer Van Hook and Passel have analyzed cross-generational trends using CPS data from 1995 through 1998 to examine generational differences among persons in their twenties (i.e., born after 1965). We compare the following groups:

1st generation - immigrants entering the U.S. after age 10;

1.5 generation -immigrants entering the U.S. before age 10;

2nd generation - persons born in the United States with one or two immigrant parents;

3rd- and-higher generations - natives with native parents.

By the 2nd generation, immigrants overall have reached (or exceeded in some instances) the levels attained by 3rd generation non-Hispanic native whites in terms of educational attainment (i.e., percent with high school diploma and percent with bachelor's degree or more), labor force participation, wages, and household income. There are substantial differences across racial and ethnic groups - Asian and Pacific Islanders generally are doing better than whites , Hispanics doing worse, and 2nd generation blacks show mixed outcomes. For wages, virtually all of the differences across generations and across racial/ethnic groups can be explained by educational differences; when the groups are standardized by educational attainment the wage differences are not statistically significant.

While generally positive integrative trends are found in economic areas, a more mixed picture emerges in the social statistics. First, we do find that intermarriage across racial and ethnic lines increases with generational duration in the United States. On the other hand, our results show convergence across generations to native patterns of family disintegration. Among these young adult cohorts, the first generation is considerably less likely than natives to be divorced. By the 2nd generation, the proportion divorced or separated is double that of the 1st generation and at the level of 3rd generation non-Hispanic whites. A similar pattern occurs for the proportion of parents who are unmarried.

Fiscal Costs and Benefits of Immigration. Measuring the fiscal impact of immigration has proved to be very difficult for a number reasons, many of which are methodological in nature. First, there is no general agreement and no clear rationale for deciding which costs and impacts to include, not on how to measure them. Some of the considerations include: Should the accounting be done on a cross-sectional or longitudinal basis? How are public goods accounted for in the analysis? How are costs and revenues attributed to generations? (For example, if the cost of educating native-born children of immigrants is attributed to the immigrant parents, what happens to the taxes paid by those children when they mature. And, what about their children?) More broadly, how are cross-generational transfers treated? Should education be considered a "cost" or an "investment"? What assumptions are to be made about the overall fiscal picture?

In the past decade, a number of studies have been attempted to address the question of fiscal impacts. Several conclusions have emerged upon which there appears to be fairly widespread agreement. First, immigrants (and immigrant households) pay a considerable amount in taxes to all levels of government. However, because immigrant incomes are generally lower than native incomes when considered on a cross-sectional basis, the taxes from immigrants on a per capita or per household basis are lower than for natives. Similarly, the net balance of taxes and social spending directed toward families is more positive for native families than for immigrant families. This result derives principally from three factors: the previously-mentioned income differences; the biggest cost associated with immigrant families in general is the cost of educating children; and immigrant families have more children than native families.

The National Academy of Sciences (1997) attempted the most extensive study of this issue to date. In their study, the Academy attempted to model costs and taxes on a longitudinal basis and take into account the future generations derived from immigrants. Their main conclusion was that, on average, an additional immigrant generated a positive net contribution to the country. This varied considerably according to a number of factors. In general, the younger the immigrant, the greater the net contribution because younger immigrants have longer working times in the U.S. when they pay taxes. The more highly educated the immigrant the greater the net contribution. Again, this result is related to income. More highly educated immigrants tend to have higher incomes and pay higher taxes.

The balance of taxes versus costs tends to favor the federal government. More taxes are directed to the federal government than to state and local governments. On the other hand, the highest "costs" associated with immigrants tend to be for educating children and most of these costs are incurred by state and local governments. This particular result points out some of the major problems with these analyses. Most of the costs of educating immigrant children are spent on natives (the U.S.-born children). Yet, the research shows clearly the payoffs to education. Moreover, since this is the most critical factor for the integration of immigrants and their offspring, it is the most critical for the long-term health of the U.S. economy.


There is no doubt that there are a very large number of immigrants living in the United States and that their numbers will continue to grow. It is generally accepted that previous groups of immigrants have been integrated into the country. Moreover, the results presented suggest that the current group of immigrants is integrating, but that there may be some issues that need to be addressed. Elsewhere, we have argued that there is a basic mismatch "between the nation's essentially liberal, if highly-regulated immigration policies and its historically laissez faire immigrant policies. That is, even though the nation admits more immigrants who are on a track for citizenship than any other country, U.S. immigration policies have been essentially ad hoc and small scale." (Fix, Zimmermann, and Passel 2001.) Consequently, given the magnitude of the immigrant population, the time may be ripe to explicitly consider immigrant integration issues and programs.

The issues involved are complex, but there is no doubt that successful integration of immigrants and their children will be a crucial factor in the country's future. We have raised a number of principles to be considered in an integration agenda. First, how do we best promote the social and economic mobility of immigrants and their families? These include some difficult cases such a refugees and limited English speakers. How should legal immigrants and citizen children be treated? How do we deal with undocumented populations? What is the proper role for sponsors? How do we promote intergovernmental fiscal equity regarding immigrant taxes and spending? What is the best role for the private sector?

Data Needs. The successful design of integration strategies and monitoring of new programs and policies will require more and better data. There have been great strides made in producing data for immigrants. Without the collection of nativity data in the CPS, we would know considerably less than we do and would probably still be basing assessments on data from the 1990 Census. Both the Census Bureau and INS are to be commended for instituting the CPS data collection in 1994.

However, more improvements are needed. Sample sizes are still small, especially for some populations. Information on legal status is limited, but critical for distinguishing differential impacts and consequences. Along these lines, improvements in data collection will be required to gather accurate information. There should be up-to-date, accurate estimates for the immigrant population is a variety of legal statuses, not just undocumented immigrants. To achieve this end, more data on immigration flows into (and out of) the United States needs to be collected in ways that conform to demographic and commonsensical notions of immigration. Finally, there needs to be better definition of the proper roles for government agencies (notably the Census Bureau and INS) in the collection, production, and analysis of immigration data .