Pathways to legal immigration

Abstract

In this paper we use the New Immigrant Survey Pilot Study (NISP) to describe the amount and kind of experience that immigrants accumulate in the United States before they become permanent resident aliens. The NISP surveyed a representative sample of legal immigrants who acquired residence papers during July and August of 1996, yielding a completed sample of 1,135 adults. Our analysis revealed that roughly two-thirds of these newly arrived immigrants had prior experience in the United States within one of six basic categories: illegal border-crossers, visa abusers, non-resident visitors, non-resident workers, students or exchange visitors, and refugees/asylees. Each of these pathways to legal immigration was associated with a different profile with respect to nationality, social background, and economic status. Using simple earnings regressions we demonstrate how these differences can yield misleading conclusions about the process of immigrant adaptation and assimilation, even if measured effects are reasonably accurate. We suggest that social scientists should change the way they think and ask about immigrants’ arrival in the United States.

Much of what the public believes about immigration today is drawn from a much- elaborated collective memory of America’s last period of mass immigration, from 1870 to 1920. In the public imagination, poor European immigrants stepped off the boat onto Ellis Island wide-eyed, innocent, and inexperienced in American ways, but with a gritty determination to make their way in the New World. Despite encountering prejudice and hostility along the way, they quickly got down to the business of getting ahead in their adopted land and never looked back.

This romantic notion of definitive, one-time migration suffused much of the early scholarship on European immigration (see, for example, Handlin 1951, 1954); but recent work suggests that even during the classic period, there was considerable circularity and much movement back-and-forth across the Atlantic (Nugent 1992; Wyman 1993). With the exception of Jews from the Russian Pale, virtually all European immigrant groups displayed significant rates of return migration, a tendency that grew as transportation technologies improved after the turn of the century (Hatton and Williamson 1998). From 1890 to 1914, the return flow of immigrants ran at about 30% of the inflow, and the emigration rate reached nearly 50% for some groups, such as Italians and Spaniards (Hatton and Williamson 1994). Although high rates of return do not necessarily indicate circularity, they are consistent with other evidence suggesting that a large share of the immigrants who arrived in the United States between 1870 and 1920 were not arriving for the first time (Nugent 1992; Wyman 1993).

If departure and re-entry was common among immigrants during an era of steamships and telegraphs, it must be even more prevalent in an era of jet airplanes and telecommunications. In recognition of this likely circularity, the U.S. Census and Current Population Survey presently ask foreign born respondents not simply when they “arrived” in the United States, but when they came to the United States “to stay”. Of those foreign born who said they came to stay between 1985 and 1990, however, some 30% reported that they were already in the United States in 1985, implying that sometime over the succeeding five years they had left the country and returned, leading Ellis and Wright (1998) to wonder just when an immigrant is really an immigrant.

In an era of globalism, the issue of when one arrives “to stay” is highly ambiguous. An immigrant may enter the United States with no intention of “staying” and yet find him or herself living there several years later, with no immediate plans for return. Likewise, immigrants may enter with every intention of staying forever, but later find themselves drawn home for a variety of social, economic, and cultural reasons. Undocumented migrants, for their part, can never be sure that their “stay” won’t end tomorrow through an involuntary deportation, no matter what their intentions; and even when someone holds a visa for “permanent residence”, it doesn’t necessarily mean that he or she intends “to stay” forever, or even for very long. Immigrants often use a U.S. residence visa as a simple convenience to enter and exit the country at will, while maintaining a primary home abroad (see Reichert and Massey 1979; Mines and Massey 1985).

Ambiguities about when someone has come “to stay” also results from U.S. immigration policies. For example, the amnesty and legalization programs authorized by the 1986 immigration reform and control act granted legal status to some three million former undocumented migrants who were first granted temporary legal residents and then later “adjusted” to permanent residence. As a result during the period 1988–1992 the number of immigrants to the United States swelled to values in excess of one million persons, not because immigration had suddenly increased, but because policies had changed. Likewise, the controversial 245(i) provision of U.S. immigration law permits certain undocumented migrants to apply for legal residence while remaining in the United States, rather than forcing them to return home before making their application.

In addition to such practical matters, the concept of migrating “to stay” also has problems theoretically. Although neoclassical economics (NE) posits one-way movement in order to reap higher lifetime earnings, the new economics of labor migration (NELM) predicts circular migration and remitting to overcome market failures at home (Stark 1991). Although the two perspectives are not mutually exclusive, research has generally suggested that the behavior of contemporary international migrants is more consistent with NELM than NE (Massey et al. 1998). In Mexico, especially, critical comparisons of hypotheses derived from each theory suggests that NELM has the greater explanatory power (Massey and Espinosa 1997).

Despite the conceptual and empirical problems with concepts such as “permanence”, “settlement”, and “intention to stay”, federal statistical agencies continue to pose the arrival question in the same way, adopting an anachronistic 19th century view of immigration that was not even accurate in the 1890s. The census question, for example, forces persons of foreign birth to judge retrospectively what their subjective intentions were at some ambiguous point in the past, making the respondent not only reconstruct an earlier psychological state, but also decide whether the question refers to the first trip, last trip, or some visit in-between (the point at which they came to stay).

As a result of this methodological flaw, estimating time spent in the United States by subtracting the date of an immigrant’s arrival from the date of the census yields a result that is ambiguous at best. For immigrants with multiple U.S. trips, the residual clearly understates the true amount of U.S. experience. In a world where migrants circulate back and forth across borders accumulating entries, departures, and various periods of work and residence abroad, the estimation of migratory experience by subtraction thus yields a biased measure of experience, and, hence, biased and unreliable conclusions about fundamental processes of social, economic, and cultural assimilation (Lindstrom and Massey 1994).

The severity of these biases, moreover, is likely to vary from group to group, depending on the nature and amount of people’s earlier migratory experience. There are a variety of ways that immigrants might accumulate U.S. experience before becoming a legal resident alien: by crossing the border illegally, by overstaying a non-resident visa, by coming and going as a tourist, by entering as a temporary worker, by attending a U.S. college or university, or by becoming a refugee or asylum seeker. Because immigrants from different countries tend to display different mixes of these pre-green card statuses, each associated with a different set of background traits, the degree of unobserved heterogeneity is likely to vary considerably from country to country and group to group.

In this paper we draw on data from the New Immigrant Survey Pilot to assess the amount and kind of experience that immigrants from different countries accumulated in the United States before they received their green cards (which are now actually pink; but the name has stuck). We define seven distinct “pathways” to legalization and show that the majority of U.S. immigrants do not enter the country for the first time when they arrive as “new” legal resident aliens. We also show that the nature of prior experience varies considerably across national origin groups, and that the mix of pathways differs from country to country. Each pathway, moreover, is associated with a distinct set of socioeconomic and demographic background characteristics, yielding serious risks of selection bias and measurement error.

1. The New Immigrant Survey Pilot

The New Immigrant Survey Pilot (NISP) was funded by the National Institute of Child Health and Human Development in partnership with other federal agencies to assess the feasibility and cost of implementing a longitudinal survey of immigrants to the United States. The target population was the cohort of immigrants admitted during the months of July and August 1996, some 148,987 people in total. Because children are relatively numerous among immigrants, and because employment-based immigrants are relatively rare, the NISP employed a stratified design that under-sampled children and over-sampled employment-based immigrants. Children under the age of 18 formed one stratum, which was sampled with a selection probability of 0.003715, and adults aged 18 or older who entered with an employment visa formed a second stratum and were sampled with a selection probability of 0.047201. The third and final stratum included all other immigrants, who were chosen with a selection probability of 0.013486. Thus, employment-based immigrants were 12.7 times more likely to be selected than children, and 3.5 times more likely to be drawn than those in the residual category (see Jasso et al. 1999).

This sampling strategy yielded a sample of 1,982 immigrants, of whom 1,837 were adults. Baseline interviews were conducted primarily by telephone and began in October of 1996 and continued through mid-November of 1997. In cases where respondents lacked a telephone or had a number that could not be determined, interviewers tracked down the respondent and completed the survey in person. Interviews were conducted in whatever language the respondent preferred: 43% were in English; 27% were in Spanish, and the remaining 30% were carried out in one of 16 different languages.

The present analysis is based only on the adult portion of the survey, using weights to reflect differences in selection probabilities between employment-based and other immigrants. In all, 1,134 adult interviews were completed, for a completion rate of 61.7%. The rate was much higher among respondents who were actually contacted by interviewers, however. Once contact had been established, only 15% of potential respondents refused to participate. Most of the non-response stemmed from our inability to locate respondents in the first place. There were, however, few differences between respondents and non-respondents with respect to age, sex, marital status, visa type, and state of U.S. residence (Jasso et al. 1999).

2. Pathways to Legalization

Of all the adult immigrants interviewed in the NISP, some 66% reported being in the United States at some point before receiving their visa for permanent residence. In other words, a relatively small share of immigrants actually conform to the stereotype of people who are arriving in the United States as “new” permanent residents. Some have been living in the country for a long time and are simply adjusting status; others are entering from abroad but have accumulated U.S. experience on one or more prior trips. Table 1 identifies six principal pathways to legal immigration and shows the percentage distribution of adults among them. The algorithm we used to allocate immigrants to the various pathways is summarized in Appendix A.

Table 1.

Distribution of 1996 cohort of legal immigrants by prior US experience and legalization types

Prior U.S. Experience type	Total (%)	Type of eventual legalization
		Family-based (non-quota) (%)	Family-based (quota) (%)	Employment-Based (%)	Refugee (%)	Diversity (%)
New arrival	33.8	30.6	53.9	6.0	–	80.8
Illegal border crosser	20.7	25.5	27.7	12.5	2.9	5.8
Visa abuser	10.6	12.2	9.0	13.3	10.5	1.9
Business/professional	1.2	1.1	1.6	1.1	–	–
Student/exchange	0.6	0.3	1.2	0.8	–	–
Transit	0.1	–	0.4	–	–
Temporary worker	0.3	0.6	–	0.8	–	–
Tourist	5.5	5.9	2.0	10.1	10.5	1.9
Border crosser	1.0	1.4	1.6	–	–	–
Non-resident visitor	14.7	21.5	9.0	21.2	2.9	7.7
Business/professional	1.1	0.8	0.4	5.4	–	–
Diplomat	0.1	–	–	1.1	–	–
Transit	0.1	0.3	–	–	–	–
Temporary worker	0.9	0.8	–	4.6	–	–
Tourist	11.9	19.0	7.4	10.1	2.9	7.7
Border crosser	0.3	–	1.2	–	–	–
Other	0.2	0.6	–	–	–	–
Non-resident worker	4.3	2.5	–	25.0	1.0	1.9
Business/professional	1.7	1.1	–	10.3	–	–
Diplomat	0.1	–	–	1.1	–	–
Temporary worker	2.4	1.4	–	13.6	1.0	1.9
Student/exchange	5.5	6.8	–	21.7	–	1.9
Rufugee/asylee	10.5	0.8	0.4	0.3	82.9	–
Weighted N	64,596	26,175	18,983	7,796	7,786	3,856
Sample N	1,134	353	256	368	105	52

The first pathway, new arrivals, includes the mythic immigrants of old: those who arrive in the United States without ever having set foot in the country before. At 34% they are a distinct minority of all immigrants. The remaining two-thirds of the cohort achieved their migratory experience in one of five different ways. Illegal border crossers comprise 21% of all adult immigrants; they are people who report a prior trip to the United States without any documentation who we assume to have entered the country surreptitiously. Although the vast majority probably entered across the southern border from Mexico, some must have crossed the north frontier from Canada. Visa abusers, in contrast, are persons who reported making an earlier trip using a valid visa, but who went on to violate the terms of that visa; they comprise about 10% of the entering cohort of immigrants Conceptually, this category embraces two basic forms of visa abuse: entering on a short-term visa and then staying longer than permitted or entering with a visa that does not permit work and then taking a U.S. job. The most common visa abused is the tourist visa, which at 5.5% of the entire cohort accounts for 52% of the cases of visa abuse.

Non-resident visitors comprise another 15% of all immigrants and include persons who entered the U.S. on a visa that did not authorize work (mainly holders of B2 and C visas and their dependents) who never violated the terms of that visa (at least as far as we can judge). More than 80% of these people are simple tourists. The next pathway group, non-resident workers, comprises about 4% of the cohort and includes people admitted on a document that permits temporary work in the United States, such as a business or professional visa (B1, E, I, L, or T visa), a temporary worker visa (H visa), or a diplomatic visa (A, G, or N visa). Among non- resident workers, a little more than half (56%) were temporary workers and around 40% were business persons or professionals; few were diplomats.

The last two pathway categories, students/exchange visitors and refugees/asylees, comprise 5.5% and 10.5% of the entering cohort, respectively. The former includes students in academic institutions (F visas), vocational students (M visas), as well as persons present in the United States for purposes of exchange or practical training (J visas). The latter category comprises persons who were granted provisional entry to the United States as a refugee or asylum seeker, either under the parole authority of the U.S. Attorney General or under a specific act of Congress declaring their nationality eligible for refuge in the United States.

The various columns of Table 1 show the relative mix of pathways according to the category under which permanent resident status was ultimately achieved. Only one visa category (the smallest) conforms in any meaningful way to mythic picture of inexperienced immigrants entering the country for the first time: diversity immigrants, 81% of whom had never been to the United States before. This visa category was established by Congress specifically to reduce the dominance of large immigrant-sending nations (mainly in Asia and Latin America) and “diversify” the stream of immigrants with respect to national origins (by incorporating more Europeans). To obtain a diversity visa, people from eligible countries enter a lottery that randomly distributes a fixed annual number of such visas to aspiring immigrants.

The only other category in which newly arriving immigrants even comprise a bare majority (54%) are those entering under numerically limited, family-based classifications, typically spouses and children of legal resident aliens. Yet close to half of these people have prior experience in the United States: 28% as illegal border crossers, 9% as visa abusers, and 9% as non-resident visitors, mostly tourists. Among those entering in unlimited family-based categories, the prevalence of prior U.S. experience is even greater at 69%, with 26% being illegal border crossers, 12% visa abusers, 22% non-resident visitors (mostly tourists), 3% non- resident workers, and 7% students or exchange visitors, leaving just 31% of immigrants in this category without any U.S. experience whatsoever.

Considering patterns of prior experience among the two categories of family immigrants, and drawing on ethnographic reports from the field (see Samora 1971; López Castro 1986; Conover 1987; Massey et al 1987; Pérez 1991; Siems 1992; Chavez 1992; Durand 1994, 1996; Durand and Massey 1995), we surmise that a typical practice is for undocumented relatives to spend time with their documented family members in the United States while awaiting their green cards to be approved, whereupon they return home to pick up the necessary documents and re-enter the United States as “newly arriving” immigrants. Most of these family members gain access to the United States initially as illegal border crossers or visa abusers and simply wait for word that their papers are ready at a U.S. consulate in their home country.

Immigrants entering through employment-based categories are drawn predominantly from three pathways. Very few (only 6%) are newly arriving immigrants. Not surprisingly, a large share (25%) come from the ranks of non-resident workers, with a little more than half of them (54%) being comprised of persons admitted with a temporary worker (H) visa and the remainder comprised of business or professional migrants (41%) or diplomats (4%). A slightly smaller percentage (22%) enter as students or exchange visitors and almost the same share (21%) entered as non-resident visitors without work authorization (about half as tourists). Together these categories account for almost three-quarters of employment-based immigrants. Relatively few employment-based immigrants entered initially as illegal border crossers or visa abusers (about 13%), and virtually none enter as refugees or asylees. Among those immigrants who were ultimately granted residence as refugees or asylum seekers, 83% declared their intention to seek this status upon entry: 3% entered initially as undocumented border crossers and 11% came as visa abusers.

3. Who Uses Which Pathway?

Table 2 examines the national origins of each pathway by showing the top five countries of nationality contributing to each category. Among immigrants in general, Mexico is by far the largest single source country, comprising 15% of the entire 1996 cohort, followed in turn by immigrants from the former USSR (8%), China (6%), the Philippines (also 6%), and Vietnam (5%). Given its relative importance as the leading source for U.S. immigrants, it is not surprising that Mexico figures prominently in four of the seven pathways we have identified. In one category, however, Mexico far exceeds its overall contribution to the immigrant cohort: illegal border crossers, 41% of whom were Mexican. Given the fact that Mexico is the only developing country to share a land border with the United States, and knowing that 90% of those apprehended by the Border Patrol for unauthorized entry are from this country (U.S. Immigration and Naturalization Service 1997), this finding is hardly surprising.

Table 2.

Country of origin by type of prior US experience

Country of origin	Type of prior US experience
	All respondents		New arrival		Illegal border crosser		Visa abuser		Non-resident visitor		Non-resident worker		Student/exchange		Refugee/asylee
Largest	Mex.	15.0	Phil.	11.9	Mex.	41.2	Mex.	13.0	Mex.	15.8	Can.	14.9	China	23.0	USSR	41.9
Second largest	USSR	8.1	India	8.9	El Sal.	5.9	Viet.	7.6	Jam.	5.7	Phil.	13.8	Japan	5.4	Viet.	29.6
Third largest	China	6.0	Mex.	6.9	Peru	4.6	China	7.5	USSR	5.3	USSR	10.7	Korea	5.1	Cuba	8.8
Fourth largest	Phil.	6.0	China	6.7	Guat.	4.4	Phil.	5.7	China	4.6	Jam.	6.1	India	4.8	Yugo.	7.7
Fifth largest	Viet.	5.2	USSR	5.1	Ecuad.	4.4	Colom.	5.4	Italy	3.9	China	4.6	Mex./Tai.	4.2	Iraq/Sudan	2.2
Top 5 total	40.3		39.5		60.5		39.2		35.3		50.1		46.7		92.4

In some ways, what is most surprising is that Mexico is not more dominant among illegal border crossers. Some 39% of illegal border crossers come from somewhere else, mostly in Latin America: 6% are from El Salvador, 5% from Peru, 4% from Guatemala, and 4% from Ecuador. Thus, Mexico appears to serve as an important platform for illegal entry for immigrants from throughout the hemisphere. Given its unique geographic location, it serves as Latin America’s principal conduit for surreptitious border-crossing. In total, 81% of illegal border crossers are from Latin America or the Caribbean (tabulation not shown).

Mexico also leads other countries in the visa abuser category, although at 13% its share is quite close to its overall share in the immigrant cohort. The next most important contributor to visa abuse is Vietnam, which at 8% is 50% higher than one would expect from its weight in the total cohort, China (8%), the Philippines (6%), and Colombia (5%). In general, however, no country dominates the category of visa abusers the way that Mexico dominates illegal border crossers. Whereas Mexico by itself accounts for 41% of all border crossers, the six countries listed in Table 2 combined only make up 39% of all visa abusers.

Non-resident visitors are also relatively unconcentrated with respect to national origins: the countries listed in Table 2 again make up only around 40% of all immigrants in this category, with Mexico leading the way in rough proportion (16%) to its overall representation in the cohort, followed by Jamaica (6%), the countries of the former USSR (5%), China (5%), and Italy (4%). With the exception of Jamaica, which is more than double its percentage among immigrants as a whole, these percentages are roughly proportional to their relative weights in the total cohort. The over-representation of Jamaicans among non-resident visitors suggests that a large number of these erstwhile tourists may, in fact, be visa abusers seeking to enter the United States and work illegally.

Among non-resident workers, the leading sending countries are Canada (15%) and the Philippines (14%), two nations with exceptionally close economic, political, and cultural ties to the United States. The Philippines, in particular, a former American colony, is a place in which English is widely spoken, with an educational system modelled closely on that of the United States. In fact, the Philippines has numerous training programs explicitly designed to channel prospective immigrants into U.S. employment either under the various H-visa temporary worker programs or as an employment-based candidate for legal permanent residence. The relative numbers of Filipinos and Canadians taking this pathway to legalization far exceed their relative shares in the immigrant cohort. The percentages of immigrants from the former USSR (11%) and China (5%) are roughly comparable to their overall shares in the entering cohort; but once again Jamaica exceeds its expected by a factor of more than two.

Among immigrants with prior experience as students or exchange visitors, by far the largest contributor is China, which at 23% is four times its overall percentage in the total cohort. The fact that the Chinese comprise nearly a quarter of all immigrants entering the U.S. first as students or exchange visitors may indicate a significant amount of undetected visa abuse, or perhaps simply a strong attraction to U.S. educational credentials among the Chinese. Whatever the case, the next closest countries are Japan (5%), Korea (5%), and India (5%). Taiwan and Mexico are tied for fifth place, each contributing a little more than 4% to the total. Of all countries contributing immigrants to the United States through the student/exchange pathway, five of the top six are Asian, representing 44% of the total. In all, Asians make up 63% of those who entered initially as students or exchange visitors.

The national origin composition of refugees closely parallels what one would expect given U.S. foreign policy over the past several decades, with all of the most important contributors coming from the communist or former communist world. Nations of the former USSR alone comprise 42% of all immigrants with prior experience as refugees, followed by Vietnam at 30%, Cuba at 9%, and the former Yugoslavia at 8%. Together these four countries comprise 89% of all persons becoming immigrants through the refugee/asylee pathway.

Finally, among the relative minority of immigrants who enter the United States with absolutely no prior U.S. experience, the largest national contributors tend to be countries characterized by strong family ties to the United States and significant rates of marriage migration: the Philippines, India, Mexico, China, and the former USSR. All of these countries either support a strong network of kinship ties to U.S. citizens and resident aliens or a well-developed mail-order bride industry, or both. The business in mail-order brides has grown astronomically with the advent of the Internet. American men (typically older) can now search the Web to select foreign women (typically younger) to marry and sponsor as legal immigrants to the United States. Poor foreign women offer their youth, good looks, and pleasing personalities to American men in return for a green card and life in the United States.

Inter-country differences in the mix of pathways are suggested by Table 3, which presents distributions within selected countries and regions of origin (owing to space constraints, this table sums to 100% within rows rather than columns). Compared with other nationalities, immigrants from Mexico are very unlikely to be new arrivals – only 16% had never been to the United States before in their lives. Given its status as a bordering nation, it is not surprising that a clear majority (57%) of Mexican immigrants have prior experience as illegal border crossers, and another 9% are visa abusers. Thus two-thirds of all “newly arriving” Mexican immigrants have prior experience in the United States as undocumented migrants, which is comparable to the figure of 62% reported by Portes (1979) in his 1972–1973 survey. About two-thirds of Salvadorans likewise have prior undocumented experience, but the mix is more tilted toward illegal border crossing (59%) compared with the abuse of visas (7%); the share of new entrants is likewise relatively higher at 28%.

Table 3.

Selected countries of origin by type of prior US experience

Place of origin	Type of prior US experience
		Total (%)	New arrival	Illegal border crosser	Visa abuser	Non-resident visitor	Non-resident worker	Student/exchange	Refugee/asylee
Europe	12,765	100	27.6	10.0	7.7	18.1	5.2	4.6	26.9
Poland	2,076	100	52.0	16.3	14.8	11.2	5.6	–	–
Asia	20,879	100	43.7	8.0	12.5	11.1	3.7	10.4	10.7
China	3,877	100	37.4	13.9	13.1	11.2	3.3	21.0	–
India	2,616	100	74.5	–	11.3	5.3	2.4	6.5	–
Korea	742	100	12.9	5.7	12.9	41.4	2.9	24.3	–
Philippines	3,856	100	67.6	3.8	10.2	8.5	9.9	–	–
North America	1,261	100	23.5	3.4	18.5	16.0	32.8	–	5.9
Latin America	25,910	100	26.4	38.6	10.8	16.9	2.2	2.2	2.9
Dominican Republic	1,780	100	50.0	25.0	8.3	12.5	4.2	–	–
El Salvador	1,324	100	28.0	59.2	7.2	–	–	5.6	–
Mexico	9,703	100	15.5	56.7	9.2	15.5	0.9	1.5	0.8
Africa	3,697	100	55.3	10.6	5.2	7.2	8.0	5.7	8.0
Oceania	85	100	–	–	–	50.0	50.0	–	–

Immigrants from the Dominican Republic are considerably more likely to be new arrivals than are Mexicans or Central Americans (50%); but like Mexicans, those with prior experience are relatively likely to report being undocumented migrants, with 25% entering as illegal border crossers and 8% as visa abusers (with another 4% coming as non-resident visitors). Although the Dominican Republic is an island, illegal border-crossing is accomplished without difficulty by undertaking a clandestine boat ride across Mona Passage to the neighboring island of Puerto Rico (about 100 miles away). .

It is rather hard to discern an overall Asian pattern from the figures in Table 3. Whereas India displays the highest percentage of newly arriving immigrants (75%), Korea displays the lowest (13%) and China and the Philippines lie in-between at 37% and 68%, respectively. The most frequent means by which Koreans achieve legal U.S. residence is through a non-resident visa (41%), followed by students or exchange visitors (24%) and visa abusers (13%). As one would expect given Korea’s location, illegal border crossing was uncommon, but at 6% it was not completely absent either. Among Chinese immigrants, the incidence of illegal border crossing was even higher at 14%, and the relative frequency of visa abuse was also quite high at 13%. Aside from new arrivals, the student/exchange category represents the most common pathway to legalization at 21%. Owing to China’s relatively low level of economic development, Chinese are less able than other groups (e.g., Koreans) to qualify for a non-resident visa, and relatively few make use of this pathway to gain U.S. residence (11%).

Other countries that report a relatively high level of visa abuse are Poland (15%) and North America (19%, nearly all from Canada). Among Polish immigrants, however, over half (52%) were new arrivals, 16% were illegal border-crossers, and 6% were non-resident workers. In general, Europeans display a relatively high frequency of immigrants in the refugee/asylee category (27%, mainly from Eastern Europe), as well as relatively large numbers among non-resident visitors (18%), illegal border crossers (10%), and to a lesser extent visa abusers (8%), not to mention a significant representation of new arrivals (28%).

4. Migratory Background

The top panel of Table 4 presents information about the legal category under which migrants from each pathway ultimately achieved permanent resident status. Overall, 70% of the cohort attained permanent residence through a family tie to someone already present in the United States, either a citizen or resident alien, reflecting the strong emphasis on kinship built into U.S. immigration law. In contrast, only 12% of the immigrants entered under employment-based classifications, 12% were refugees, and 6% were diversity immigrants. The diversity category thus lends relatively little actual diversity to the immigrant flow.

Table 4.

Demographic characteristics of immigrants taking different pathways to legalization

Demographic characteristic	Total respondents (%)	Type of prior U.S. experience
		New arrival* (%)	Illegal border crosser (%)	Visa abuser (%)	Non-resident visitor (%)	Non-resident worker (%)	Student/exchange (%)	Refugee/asylee (%)
Status at entry
All new admissions	100	100	100	100	100	100	100	100
Non-quota family	40.5	36.7	50.0	46.7	59.2	24.1	50.1	3.3
Quota family	29.4	46.9	39.4	25.4	17.9	–	–	1.1
1st Preference	2.3	5.1	1.7	1.1	0.8	–	–	–
2nd Preference	17.3	22.8	29.4	12.0	15.6	–	–	–
3rd Preference	4.1	9.2	1.7	4.3	1.6	–	–	–
4th Preference	12.1	9.9	6.7	7.6	–	–	–	1.1
Employment	12.1	2.1	7.3	15.2	17.4	70.5	47.8	0.3
Refugee	12.1	–	1.7	12	2.3	2.7	–	95.3
Diversity	6.0	14.3	1.7	1.1	3.1	2.7	2.1	–
Prior U.S. experience
Age at first trip	33.3	35.3	28.3	33.3	33.5	31.8	27.0	40.6
Year of first trip	91.2	96.0	87.7	86.2	88.6	89.8	89.2	93.0
Duration of first trip (years)	2.5	0.1	4.4	3.5	3.5	2.6	4.3	2.6
Total number of trips	1.9	1.1	2.3	2.5	2.6	3.5	2.2	1.1
Total time in U.S. (years)	4.1	0.1	7.4	7.4	6.2	5.7	6.2	2.6
% Adjusting status	43.0	2.3	7.0	55.0	100.0	100.0	100.0	100.0
% With multiple trips	21.8	2.5	40.4	35.3	35.7	34.3	39.8	1.4
% With U.S. job before U.S. residence papers	22.7	0.0	30.0	60.6	0.0	100.0	51.6	28.8

^*Positive prior U.S. experience values are due to post-admission entries and exits and pre-admission waiting periods.

In two pathways, the share of family-based immigrants exceeds the overall percentage of 70% in the total cohort: new arrivals, where the figure was 85%, and illegal border crossers, where it was 84%. Family immigrants who are new arrivals mostly consist of the spouses and children of citizens and resident aliens. Those admitted without numerical limitation (37%) principally represent the spouses and children of U.S. citizens, whereas the second preference category under numerically limited visas (23%) includes the spouses and children of legal resident aliens. The tilt toward immediate family members is even stronger among illegal border crossers, with 48% entering as the spouse or child of a citizen, and 27% entering as the spouse or a child of a resident alien (second preference).

The pattern of visa abusers resembles that of illegal border crossers, except that relatively fewer are admitted in the second preference category. Although the percentage of family members entering without limitation (immediate dependents of citizens) is nearly identical at 47%, at 12% the share coming in through the second preference category is less than half the share observed among illegal border crossers. Thus, patterns observed for illegal border crossers and visa abusers again suggest that immediate family join their citizen or resident alien relatives abroad while they await legalization themselves, living in the United States as undocumented migrants until the time when they can return to their country of origin to pick up their documents and re-enter the country as “new” immigrants.

Among non-resident visitors, spouses and children of citizens figure the most prominently of all at 59% of the total; but not surprisingly, the large majority of non-resident workers qualified for permanent residence under employment-based criteria (71%), although a significant share (24%) are immediate relatives (mostly spouses) of U.S. citizens. Those becoming residents through the student/exchange pathway are about evenly split between spouses of citizens and employment-based classifications (50% versus 48%), with the remaining 2% coming in as diversity immigrants. Finally, in almost tautological fashion, 95% of immigrants following the refugee/asylee pathway qualify for permanent residence as refugees.

The bottom panel of Table 4 presents information about the amount and nature of prior U.S. experience for immigrants within each pathway category. Illegal border crossers generally took their first trip to the United States around 1987 at the relatively young age of 28 and spent an average of 4.4 years in the country before returning home (the distribution, however, has a very long tail and the median is much lower). Over their lifetimes, illegal border crossers took an average of 2.3 trips and accumulated a total of 7.4 years in the United States. Some 40% of illegal border crossers reported taking multiple trips to the United States, and 30% indicated that they began their current U.S. job before becoming a legal resident alien, which obviously understates the degree of U.S. labor market experience, but nonetheless puts a lower bound on the incidence of pre-green card work histories. Given the fact that these migrants at some earlier point entered the United States surreptitiously, it is not surprising that relatively few (7%) were adjusting from some recognized legal status.

Visa abusers entered slightly later in time and at older ages compared with border-crossers. The average date of entry was 1986 and the average age at that time was 33 years. Compared with illegal border crossers, the mean duration of the first trip was shorter at 3.5 years, and the average total number of trips was 2.5, yielding 7.4 years of total time spent in the United States (the same as illegal border crossers). A majority of visa abusers (61%) reported beginning their current job before receiving their green card, but 55% were adjusting status from some recognized non-immigrant legal status. These people thus had some document that entitled them to be in the United States prior to their legal immigration, but unbeknownst to the INS they had at some point violated the terms of that documentation.

Compared with the two categories of undocumented migrants, nonresident visitors tended to enter slightly more recently (1989), although at roughly the same age as visa abusers (33). The average length of stay on these visitors’ first trip was 3.5 years and the average number of trips was 2.6, yielding 6.2 years of total U.S. experience. As the typical length of a tourist visa is only 6 months, an average trip length of several years suggests there may be a substantial amount of hidden visa abuse among non-resident visitors. In order to be conservative in our estimation of visa abuse, however, we only considered tourists to be “abusers” if their stay exceeded one year, since it is possible under certain circumstances to get an extension of a tourist visa. In obtaining their green cards, all of the non-resident visitors were adjusting status from a legal category recognized by the INS. Nonetheless, 36% reported multiple trips, although none admitted beginning their current U.S. job before becoming a permanent resident alien.

On average, those in the non-resident worker pathway entered the U.S. in 1990 at the age of 32, accumulated 2.6 years on their first trip, and over the course of 3.5 trips accumulated an average of 5.7 total years in the United States. In claiming permanent residence, all non-resident workers were adjusting status and 100% work at the same job they held before making the adjustment. Those in the student/exchange category, meanwhile, entered about half a year earlier than non-resident workers (early 1989) and at a significantly younger age (27) as befitting their student status. The average length of stay on the first trip was 4.3 years, but over the course of 2.2 trips these students and exchange visitors accumulated 6.2 years of total time in the United States. Not surprisingly, all report they are adjusting status in becoming permanent resident aliens, and 40% began their current U.S. job before receiving their papers. Finally, refugees are rather old (average age 41 years) and recently arrived (1993). Most have made only one trip and have been in the country 2.5 years. All are adjusting status from a classification recognized by the INS, and 28% began their current U.S. job before receiving their residence papers.

5. Demographic and Social Background

Social and demographic characteristics pertaining to each of the pathways are presented in Table 5. New arrivals and students/exchange have the largest share of females at 61% and 58%, respectively. Students, however, are more heavily concentrated in the primary labor force ages of 20–34 years (72% compared with 48%) and their average age is considerably lower (31 compared with 36). They are also considerably more likely to be married (90% compared with 65%); but even so, they have fewer children, on average, than new arrivals (0.7 rather than 1.8). Not surprisingly, students and exchange visitors have very high levels of education, with some 69% having completed college, yielding an average education of just over 18 years (compared with figures of 10% and 11.8 years among new arrivals). Likewise, 85% of those becoming immigrants from the student/exchange pathway speak English fairly or very well (compared with 30% among new arrivals).

Table 5.

Social characteristics of immigrants taking different pathways to legalization

Demographic characteristic	Total respondents	Type of prior U.S. experience
		New arrival	Illegal border crosser	Visa abuser	Non-resident visitor	Non-resident worker	Student/exchange	Refugee/asylee
Sex
% Male	44.9	38.8	50.4	58.5	29.6	64.8	42.1	55.1
Current age
% 20–34	49.3	47.8	56.2	36.6	45.2	52.5	72.2	46.0
Average age	36.5	35.5	33.8	40.7	38.6	36.4	31.0	40.9
Marital status
% Currently married	71.4	64.7	75.7	71.3	80.5	78.2	89.9	59.5
Household composition
Mean household size	4.1	4.5	4.4	3.7	3.7	3.2	2.9	3.9
Mean number of children	1.7	1.8	1.9	2.0	1.7	1.2	0.7	1.8
Years of education completed
% Less than 7 years	13.4	14.5	25.0	10.2	10.4	2.7	0.0	5.5
% More than 16 years	17.3	10.3	8.0	16.6	21.8	49.0	68.7	12.2
Mean years of education	12.5	11.8	10.4	12.9	13.1	16.2	18.3	12.6
English language ability
% Speak fairly or very well	38.5	30.0	31.6	44.7	48.9	73.6	84.5	20.0
% Speak not well or not at all	34.7	42.4	37.2	25.7	29.7	6.9	2.4	49.3
% Understand not well or not at all	28.6	37.8	24.2	24.0	26.0	6.1	0.6	39.4

Half of all illegal border crossers are male and just 56% are of prime labor force age. Some three-quarters are currently married, and they have a mean household size of 4.4 persons with 1.9 children. With respect to schooling, a quarter have a primary education or less, and only 8% have graduated from college, yielding a low average educational attainment of 10.4 years, roughly that of someone who dropped out in their sophomore year of high school. Only a third (32%) of illegal border crosses speak English fairly or very well (although 39% say they understand it fairly or very well), and 24% report that they speak it not well or not at all.

The profile of a visa abuser is slightly different. Relatively more are male (59%), a somewhat smaller share are married (71%), and they are substantially older than illegal border-crossers, with only 40% of prime working age and an average age of over 41 years. On average, they come from households of 3.7 persons with around two children, but they are considerably better educated than illegal border crossers. Only 10% have a primary education or less, 17% are college graduates, and the mean of 12.9 years exceeds the average for all immigrants. Some 45% speak English fairly or very well, and a majority (55%) say they understand the language with some facility.

Compared with visa abusers, non-resident visitors are more likely to be female (only 30% are male), younger (average age 39 years), and married (81% are currently married). More than 21% have graduated from college, yielding an average education of 13 years. Nearly half (49%) non-resident visitors also speak English fairly or very well, and a substantial majority (56%) understand it. From these facts, we conclude that older persons with greater English language fluency and better educations – people who are in a better position to secure a visa – enter the United States on a non-resident document such as a tourist visa, whereupon a significant share violate the terms of the document by staying too long or taking a job. Many of these people are wives of male undocumented migrants who entered via other means. In contrast, younger migrants with less education and fewer skills in English become undocumented migrants through illicit border crossing.

Non-resident workers are the most heavily male of all pathway groups (65%), and not surprisingly they are concentrated in the prime working ages (53%) with an average age of 36 years. Nearly 80% are married and come from households that average 3.2 persons and 1.2 children. A relatively large share (49%) have a college education, yielding an average education of 16 years. Nearly three-quarters (74%) speak English fairly or very well, and 82% report understanding it with fluency. In contrast, refugees and asylees are much less likely to speak English at this level (only 20%) and they also have lower levels of education (an average of 13 years with only 12% finishing college). A relatively low share of refugees are currently married (at 60% they display the lowest marriage ratio), although their household size is rather large at 3.9 persons, on average. They are also rather old, with an average age of about 41 years.

6. Economic Situation in the U.S

The various pathways to legal immigration are also associated with distinct economic characteristics, which are explored in Table 6. In general, new arrivals exhibit a relatively low rate of labor force participation at the time of the survey (47%), the lowest rates of self-employment (2.4%), and one of the lowest concentrations in white collar occupations (just 10% in executive/managerial occupations and 20% in sales and clerical groups). Among those with jobs, however, newly arriving immigrants are the most likely to have received their job through a family member (40% reported being referred by a relative). Thus, new immigrants are relatively unlikely to be employed, or to be well-employed, and appear to be highly dependent on family members for the few jobs they have been able to obtain.

Table 6.

Economic characteristics of immigrants taking different pathways to legalization

Demographic characteristic	Total respondents (%)	Type of prior U.S. experience
		New arrival (%)	Illegal border crosser (%)	Visa abuser (%)	Non-resident visitor (%)	Non-resident worker (%)	Student/exchange (%)	Refugee/asylee (%)
Currently employed	57.1	46.6	58.8	73.1	44.8	99.2	79.7	59.5
Occupation
Exec/Prof/Mang/Tech	21.9	10.0	12.3	23.6	20.8	63.7	66.3	9.7
Sales/clerical	17.3	20.0	15.4	21.0	18.5	4.6	15.4	18.4
Household service	7.2	6.6	9.4	4.9	9.4	5.4	–	11.1
Other service	18.4	18.3	19.7	19.1	26.8	13.9	13.5	12.9
Agriculture	3.0	1.5	11.3	1.5	–	–	–	–
Operative	15.1	20.4	15.9	13.4	13.4	1.5	2.6	22.1
Transportation/moving	2.0	1.5	3.1	3.4	2.3	–	–	1.8
Laborer	15.1	21.7	12.8	13.2	8.9	10.8	2.2	23.9
Self-employed	7.2	2.4	9.6	12.1	7.5	13.6	4.2	5.5
Same pre-immigration employer	2.9	0.0	3.0	2.4	1.1	24.6	1.6	–
Referred by relative	24.6	40.3	26.8	14.7	8.9	3.5	26.8
Related to employer	3.4	4.0	1.3	8.2	4.3	2.7	–	1.9
Average hours per (usual) week	41.4 h	39.8 h	43.6 h	40.7 h	39.9 h	48.6 h	42.0 h	38.4 h

Both categories of undocumented migrants display high employment rates. About 59% of immigrants with a background of illegal border-crossing were employed at the time of the survey, compared with 73% of those who were visa abusers. Their respective occupational profiles were quite different, however. Visa abusers were substantially more likely to be in executive/professional occupations, with 24% in this category and another 21% in sales/clerical occupations, compared with respective figures of 12% and 15% among illegal border crossers. The latter are also considerably more likely to work in agriculture than visa abusers. Among visa abusers, only 1% worked in agriculture, compared with 11% of illegal border crossers. Visa- abusers also report slightly higher levels of self-employment compared with border-crossers (12% compared with 10%), and the latter were also considerably more likely to be dependent on relatives for employment. Whereas 27% of illegal border-crossers reported getting their job through a relative, only 15% of visa abusers did so. In addition, illegal border crossers tended to work more hours, reporting an average of 44 hours per week compared with 41 hours among visa abusers. Aside from these differences, however, members of the two groups were about equally likely to hold occupations in the services, operative, and laborer categories.

The employment rate of non-resident visitors (45%) is the lowest of any group. Among those who were employed, however, 40% held a white collar job (21% in the executive-professional category and 19% as sales or clerical workers). None of the immigrants following this pathway worked in agriculture, but 36% worked in services (9% in household services and 27% in other services). The remainder of the immigrants in this category held either operative (13%) or laborer jobs (9%). Their jobs were mostly obtained without the help of a family member and the average number of hours worked per week was nearly 40.

Not surprisingly, virtually all persons who become resident aliens after serving as non-resident workers were employed at the time of the survey; but the rate of employment for students and exchange visitors is also extremely high at 80%. Among the employed, these two pathways display similar pro-files, except that non-resident workers logged considerably more hours on the job (49 hours per week) than the latter (42 hours per week) and were also much more likely to be self-employed. Indeed, at 14% non-resident workers had the highest rate of self-employment of any category, more than three times the paltry 4% rate among students. Nonetheless, both groups had high concentrations in executive/professional occupations (66% versus 64%), with smaller concentrations in services.

The employment situation of refugees and asylees resembles that of former undocumented migrants, with a rate of employment of 59% and relatively strong dependence on relatives for job referrals (27% got their job through a family member). Employed immigrants in the refugee pathway are heavily concentrated in operative (22%), laborer (24%), sales (18%), and service occupations (24%), and like illegal border-crossers, few held professional or managerial occupations (10%). The principal difference is that refugees have a lower rate of self- employment (6%) and none works in agriculture. Indeed, no group comes close to displaying the concentration in agriculture exhibited by illegal border crossers, although at 10%, 90% of this group do not work in the agrarian sector, contradicting a common stereotype.

7. The Potential for Bias

Our descriptive results to this point have established conclusively that the vast majority of “new” immigrants to the United States are not, in fact, new to these shores. Indeed, roughly two-thirds of all entering immigrants report prior U.S. experience under a variety of different statuses that differ substantially from one another with respect to national origins, migratory history, social status, and economic characteristics. Standard residual methods of measuring time spent in the United States can thus be expected to yield estimates that are systematically biased, leaving considerable unmeasured variation in the amount and nature of pre-green card experiences in the United States.

In Table 7 we assess the degree to which statistical models may be biased by such unmeasured heterogeneity by estimating two earnings regressions. The first set of columns shows the estimates that would result using variables typically available from published sources. The second set of columns adds in variables pertaining to U.S. experience acquired before the respondent achieved permanent residence status. In each equation we regress the log of the respondent’s reported weekly earnings on selected demographic and employment characteristics, as well as measures of human and social capital, controlling for regions of origin and destination. The second equation includes dummy variables for each pathway was well as indicators of the degree and nature of the respondent’s prior U.S. experience: total number of prior entries, total time accumulated in the United States, whether or not the respondent adjusted status, and whether or not the respondent got the current job before receiving his or her green card.

Table 7.

Ordinary least squares statistics estimating logged weekly wages

Independent variables	Baseline model			Expanded model
	Coef. est.	s.e.	p-value	Coef. est.	s.e.	p-value
Intercept	3.008**	0.367	0.000	3.498**	0.348	0.000
Demographic characteristics
Male	0.246***	0.068	0.000	0.182**	0.063	0.004
Age	0.051**	0.020	0.012	0.033*	0.019	0.089
Age squared	−0.001**	0.000	0.030	0.000	0.000	0.151
Married	0.070	0.083	0.396	0.041	0.080	0.612
Number of children	−0.036	0.030	0.231	−0.030	0.028	0.300
Human capital
Years of education	0.037**	0.008	0.000	0.020**	0.009	0.023
Speaks English well or fairly well	0.371**	0.073	0.000	0.280**	0.070	0.000
Number of months since admission	−0.014	0.011	0.185	−0.003	0.010	0.734
Social capital
Related to employer	−0.230	0.212	0.277	−0.225	0.196	0.252
Referred to job by relative	−0.069	0.091	0.448	0.009	0.086	0.920
Employment characteristics
Self-employed	−0.212	0.138	0.123	−0.337**	0.132	0.011
Number of hours worked per week	0.030**	0.003	0.000	0.026*	0.003	0.000
U.S. region of residence
Northeast	–	–	–	–	–	–
South	−0.009	0.079	0.913	0.006	0.076	0.939
North central	−0.057	0.116	0.626	−0.095	0.109	0.383
West	−0.076	0.087	0.382	−0.073	0.082	0.373
U.S. territories	0.076	0.776	0.923	0.293	0.715	0.682
Global region of origin
Europe	–	–	–	–	–	–
Asia	−0.168*	0.090	0.062	−0.170**	0.086	0.049
North America	0.367*	0.195	0.060	0.234	0.191	0.220
Latin America	−0.140	0.098	0.152	−0.281**	0.098	0.004
Africa	−0.232	0.154	0.132	−0.186	0.144	0.199
Oceania	−0.353	0.780	0.651	−0.125	0.720	0.862
Pathway to legalization
New arrival	–	–	–	–	–	–
Illegal border crosser	–	–	–	0.246**	0.116	0.035
Visa abuser	–	–	–	0.107	0.168	0.525
Non-resident visitor	–	–	–	0.058	0.186	0.757
Non-resident worker	–	–	–	0.462**	0.192	0.017
Student/exchange	–	–	–	0.377**	0.193	0.051
Refugee/asylee	-	–	–	−0.174	0.187	0.351
Prior U.S. experience
Adjusted immigrant status	–	–	–	0.019	0.148	0.896
Number of entries	–	–	–	0.029**	0.010	0.003
Total time spent in the U.S.	–	–	–	0.006	0.007	0.370
U.S. employment prior to legalization	–	–	–	0.195**	0.090	0.031
Adjusted R-square		0.3910			0.4570

– = Omited variable.

The first set of columns yields results typical for immigrant earnings regressions. With respect to demographic influences, wages are higher for men; they rise during the early labor force years, peak, and then decline; and they are generally greater for those who are married (although this effect is not significant statistically). The acquisition of human capital significantly raises the rate of remuneration among immigrants: each year of education raises weekly earnings by 3.7%, and the ability to speak English fairly well or very well carries a sizeable 37% earnings premium. Time since arrival is usually an important explanatory factor in models of immigrant assimilation, but the index “months since receipt of green card” is not significant here, owing largely to the lack of effective variance. Since all immigrants in the sample arrived at essentially the same time, most of the observed variance in time since arrival stems from variation in the date of the interview, which we would not expect to be related to rates of pay.

None of the social capital variables we included in the baseline model seem to influence wages significantly: being related to one’s employer or receiving a job reference from a relative had no significant effect on the amount earned per week, and among employment characteristics, only hours worked had the expected strong and positive effect (other things equal, the more one works per week, the higher the weekly earnings). Among the regional controls, the coefficients for Asia and North America approach significance, but in opposite directions: whereas Asians earn 17% less than Europeans (the reference category), North Americans (i.e., Canadians) earn 37% more.

The expanded model shown in the second set of columns reveals that the inclusion of information about pre-green card experiences significantly increases the model’s explanatory power, producing an 20% increase in the proportion of variance explained as the R² goes from 0.39 to 0.47. However, most of the coefficients included in the standard model did not change significantly with the inclusion of the expanded variable set. Although overall patterns of statistical significance changed slightly (age and North American origin decline in importance, while self-employment and Latin American origin now attain statistical significance in the negative direction), only one of the shifts was meaningful in a statistical sense: the apparent effect of education dropped by roughly half. Whereas each year of schooling yielded 3.7% higher wages in the baseline model, the gain was only 2% in the expanded model.

Thus, although models typically estimated from INS and Census data may yield a specification error by omitting powerful variables crucial to understanding the process of economic assimilation in the United States, they do not appear to yield substantially biased coefficients for other variables. This outcome replicates earlier findings by Lindstrom and Massey (1994) using an entirely independent, and very different, data set. The principal errors of interpretation are thus likely to stem from our inability to measure the effects of factors crucial to the process of economic achievement in the United States: the pathway by which immigrants reached legal status and the amount and nature of their prior U.S. experience.

Among the pathway groups, non-resident workers generally report the highest average earnings subsequent to their legal immigration (46% higher than new immigrants, the reference category), followed by students and exchange visitors (38% higher) and illegal border-crossers (25% higher). When they work, immigrants following the other pathways to legalization – visa abuse, non-immigrant residence, or refugee status – generally report earnings that are the same as new immigrants, other things equal.

Of course, other things are not equal, and the various pathways are associated with other advantages that we measure and include in the model. As the number of prior entries increases, for example, weekly earnings tend to rise. Each additional entry yields a 2.9% increase in earnings. Likewise, immigrants who obtained their current U.S. job prior to receiving their green card experience earnings that were 20% higher than those who got their job after legally entering the United States as permanent residents (most likely reflecting the effects of job-specific human capital and the advantages of incumbency).

8. Conclusion

Contrary to widespread beliefs, most legal immigrants to the United States are not arriving for the first time when they receive their visa for permanent residence. The large majority have been here before in one guise or another, following one of six basic pathways to legal immigration: illegal border crossers (21% of all immigrants), visa abusers (11%), non-resident visitors (15%), non-resident workers (4%), students and exchange visitors (6%), and refugees/asylees (11%), although the precise mix varies from country to country and region to region. Together immigrants in these categories comprise about two-thirds of all new immigrants to the United States, meaning that only one-third are really new arrivals at the point they receive their green card.

A detailed analysis of the various pathways revealed distinct differences in the mix of nationalities: Latin Americans in general and Mexicans in particular dominate among illegal border-crossers, whereas Asians and especially Chinese dominate among students and exchange visitors. Refugees and asylees are heavily European, with smaller concentrations of Vietnamese, whereas visa abusers are composed of a mix of Asians, Latin Americans, and Caribbeans. Non- resident visitors are spread between Latin America, Europe, and Asia, in contrast to non-resident workers, who come from the Caribbean, Europe, and Asia.

Each pathway to legal immigration is associated with a different set of migratory experiences. A large share of immigrants who began as illegal border crossers, visa abusers, and non-resident visitors appear to be the spouses or dependents of U.S. citizens or resident aliens, as evidenced by large concentrations in the non-quota and second preference categories. These people enter the United States surreptitiously or with tourist visas and then stay on to be with family members until their papers are approved, whereupon they return to their country of origin to pick up the documents and re-enter as “new” immigrants. In these categories, the typical immigrant made 2–4 prior trips and accumulated 5–8 years of experience in the United States before legalizing.

Among undocumented migrants, those who began as illegal border-crossers generally display the lowest levels of human capital, with an average of just 10 years of schooling and 25% reporting a primary education or less; only a third speak English with any fluency. In contrast, only 10% of visa abusers have less than a primary education and 17% are college graduates, yielding an average attainment of nearly 13 years; and 45% speak English fluently. Non-resident visitors have even more advantaged traits, suggesting that immigrants seeking to spend time with family members while awaiting legalization are channeled in two separate directions depending on their human capital: those with education and language ability get a visa that they can overstay or periodically renew, whereas those without these resources resort to clandestine border crossing. In general, non-resident workers, students, and exchange visitors display the highest levels of human capital attainment. Most are college graduates, and at least three-quarters speak English fluently, and upwards of 80% understand it well. Immigrants who began as refugees or asylees generally have high school educations, and relatively few speak English fluently.

These contrasts in human capital are reflected in the occupational distribution associated with each pathway. Illegal border crossers principally work in services, agriculture, manufacturing, or as unskilled laborers, whereas both visa abusers and non-resident visitors display concentrations in executive/professional occupations as well in sales and clerical jobs. Those in blue collar occupations are generally employed as service workers, operatives, or unskilled laborers; virtually none are farmworkers. Those entering initially as students or non-resident workers are heavily concentrated in the upper reaches of the occupational distribution, with nearly two-thirds of both groups reporting executive/professional occupations. In contrast, refugees and asylees resemble illegal border crossers, except that none work in agriculture; they work mainly as laborers, operatives, and service employees.

Differences in the socioeconomic backgrounds of immigrants following different pathways to immigration carry the potential to bias standard models of immigrant adaptation and assimilation estimated from census or INS data. To assess how the omission of variables pertaining to prior U.S. experience might affect such models, we estimated two wage regressions – one containing variables typically available from these sources and another incorporating various measures of pre-green card experience in the United States. Although the omission of the latter variables yields a specification error by overlooking factors that strongly influence earnings garnered after legal immigration to the United States, their exclusion generally does not appear to yield biased coefficients for the variables that are measured, with the possible exception of education, whose effect was overstated. It thus seems that conclusions drawn from estimates based on census or INS data will be more or less valid as far as they go; they will simply fail to show the whole picture, and yield a rather narrow and incomplete conceptualization of the process of immigrant adaptation and assimilation.

Our analysis suggests that traditional methods for measuring experience in the United States need to be updated and adapted to the realities of circular migration and the extensive movement back and forth that occur before legalization. Asking immigrants to state when they came to the United States to stay is not the best way of conceptualizing international migration. While it may not be practical to replace this item from the Decennial Census and replace it with multiple questions, on the Current Population Survey it would be preferable to supplant it with a set of questions similar to those employed in the NISP, asking respondents to report the date and duration of their first U.S. trip regardless of legal status, whether or not they ever left the country after that entry, and when relevant, the date and duration of the last trip (as well as the total number of entries). This information will enable social scientists to construct more accurate indicators of experience in the United States and to derive more valid and realistic models of immigrant adaptation and assimilation.

Appendix

Table A.

Algorithm used to assign sample of legal immigrants in 1996 cohort to pathways classifications on the basis of prior U.S. experience

1.	If R reports no U.S. visits prior to admission→	New arrival
2.	If R reports U.S. visits prior to admission	Illegal border crosser
	and R reports having no documents on first or last visit→
3.	If R reports U.S. visits prior to admission
	and R is not previously assigned to a pathway
	and [(R reports first entry with a tourist visa and
	duration of first tip was 12 months or longer) or
	or (R reports last entry with a tourist visa and
	duration of last tip was 12 months or longer)]
	or
	If R reports U.S. visits prior to admission
	and R is not previously assigned to a pathway
	and R reports beginning of current trip prior to admission
	and INS classifies R as a new arrival
	or
	If R reports U.S. visits prior to admission
	and R is not previously assigned to a pathway
	and INS classifies R as adjusting from undocumented status
	or legalizing under IRCA
	or
	IF R reports U.S. visits prior to admission
	and R is not previously assigned to a pathway
	and R reports pre-admission documentation prohibiting employment
	and R reports that current job began before admission→	Visa abuser
4.	IF R reports U.S. visits prior to admission
	and R is not previously assigned to a pathway
	and R reports pre-admission documentation	Student/Exchange
	as student or exchange visitor→
5.	IF R reports U.S. visits prior to admission
	and R is not previously assigned to a pathway
	and R reports pre-admission documentation	Refugee/asylee
	as a refugee, asylee, or parolee→
6.	IF R reports U.S. visits prior to admission
	and R is not previously assigned to a pathway
	and R reports pre-admission documentation that forbids employment	Non-resident visitor
	and there is no evidence of pre-admission employment→
5.	IF R reports U.S. visits prior to admission
	and R is not previously assigned to a pathway
	and R reports pre-admission documentation	Non-resident Worker
	that permits employment in U.S.→