Policy Shocks: On the Legal Auspices of Latin American Migration to the United States

Abstract

In this paper, I compare the transition into legal permanent residence (LPR) of Mexicans, Dominicans, and Nicaraguans. Dominicans had the highest likelihood of obtaining residence, mostly sponsored by parents and spouses. Mexicans had the lowest LPR transition rates and presented sharp gender differentials in modes: women mostly legalized through husbands while men were sponsored through IRCA, parents. Nicaraguans stood in-between, presenting few gender differences in rates and modes of transition and a heavy dependence on asylum and special provisions such as IRCA and NACARA. I argue these patterns stem from the interplay of conditions favoring the emigration of and the specific immigration policy context faced by migrant pioneers; the influence of social networks in reproducing the legal character of flows; and differences in the actual use of kinship ties as sponsors. I discuss the implications of these trends on the observed gendered patterns of migration from Latin America.

Broadly conceived, immigration policy represents a response to forces influencing migration and, in some instances, an attempt to shape them. In many circumstances, the formulation and implementation of U.S. immigration policy has reflected foreign policy considerations rather than humanitarian or economic ones (Coffino 2006; Coutin 2000; Massey and Sana 2003; Wasem 1997). As such, flows from a given nation are not only endogenous to the geopolitical context of their times, they are often deeply affected by it, resulting in the differential treatment of national origin groups by the state, beyond what one might expect from the specific conditions leading to emigration from various countries (Coutin 2000; Grasmuck and Pessar 1991; Mitchell 1989, 1994; Portes and Grosfoguel 1994).

The ability of particular cohorts of immigrants to transition out of undocumented and gray legal statuses and into legal permanent residence and citizenship not only alters their own immigration experience, but that of subsequent generations. Hence, even those immigration policies and practices that pertain to specific generations of immigrants greatly influence the context of reception and modes of incorporation for subsequent arrivals (Portes and Rumbaut 2006). The modeling of patterns of transition into legal status is thus relevant to scholars and policymakers seeking to understand immigrant incorporation and the evolution of immigration flows over time (Bratsberg, Ragan, and Nasir 2002; Chavez, Flores, and Lopezgarza 1992; Donato 1993; Donato, Aguilera, and Wakabayashi 2005; Massey and Bartley 2005; Menjivar 2006; Phillips and Massey 1999; Stodolska 2006; White, Biddlecom, and Guo 1993).

A substantial amount of work has been devoted to the study of naturalization (Bloemraad 2006; Gilbertson and Singer 2003; Jasso and Rosenzweig 1986, 1989; Pantoja and Gershon 2006; Portes and Curtis 1987; Van Hook, Brown, and Bean 2006; Woodrow-Lafield et al. 2004; Yang 1994a, 1994b). Naturalization, however, is conditional on the prior attainment of legal permanent residence, so the playing field among naturalizing migrants has, in a sense, already been leveled. There are many fewer studies of the transition into legal permanent residence (LPR, also referred to as “immigrant admission”).¹ In addition, most of the work devoted to studying LPRs has looked only at the profile and modes of entry of those admitted to permanent resident status (Jasso and Rosenzweig 1986, 1989; Massey and Malone 2003; Newbold 2000; Newbold and Achjar 2002; Polgreen and Simpson 2006; Tyree and Donato 1985).

As these studies exclusively examine permanent residents after admission, they cannot measure the likelihood that an individual obtains LPR in the first place, or determine what factors influence this likelihood (for an exception studying Mexicans, see Malone 2004). Levels and determinants of legalization are likely to be particularly important in populations with large undocumented components, as is the case for many Latin American groups, especially Mexicans (Passel 2005). In this article I examine legal immigration from Latin America from a cross-national perspective, comparing rates and patterns of transition to LPR among Mexicans, Dominicans, and Nicaraguans, situating each transition in the context of evolving U.S. immigration policies with respect to the country in question.

Across these nations I find quite distinct transition patterns reflecting three principal factors: the interplay between U.S. foreign policy concerns and the conditions that originally prompted out-migration; the effect of social networks in reproducing the original character of the outflow; and differences in kinship ties to LPRs that are not entirely explained by prior differences in naturalization. I argue that these three factors have generally worked to exacerbate initial cross-national differences in patterns of migration of men and women, differences heretofore explained mainly in terms of contrasting family systems (Massey, Fischer, and Capoferro 2006). Before presenting and discussing the results, I review the evolution of the migratory flows in the context of U.S. policy.

U.S. Immigration Policy and Migration From Latin America

During its first century and a half of existence, the United States placed no limits on the number of immigrants that could be admitted for permanent residence, except for those from East Asia and later Africa. Prior to the 1920s the United States set only qualitative restrictions on immigration, declaring certain kinds of people ineligible for admission. Quantitative limits to immigration were first established in 1921 and tightened in 1924, but they only applied to immigrants from the Eastern Hemisphere (i.e. Europe and the Middle East). Africans and Asians were banned entirely. The 1952 Immigration and Nationality Act (INA) reaffirmed these national origin restrictions but created a new four-category preference system favoring skills and family reunification to determine priorities for entry within the Eastern Hemisphere (Vialet 1991).

Latin American Immigration in the Pre-Quota Era

For most of U.S. history there were no quantitative limits on immigration from the Western Hemisphere. Despite the absence of numerical barriers, however, few Latin American nations sent significant numbers of migrants to the United States prior to 1970. Until then, the only sizable flows came from Mexico, Puerto Rico, and Cuba, reflecting a shared history and geography with the United States. The first sizable admission of Mexicans into the U.S. came in 1848 as a consequence of shifting national boundaries after the Mexican-American War and not from population movement per se. Texas’s gaining independence from Mexico and its later annexation by the United States—along with portions of California, Arizona, New Mexico, Utah, Colorado, and Nevada—allowed some 50,000 Mexican nationals living in these territories to become U.S. citizens under terms set by the Treaty of Guadalupe Hidalgo (Vázquez and Meyer 1989).²

It was not until the late nineteenth century that significant migration from Mexico began, when U.S. agents arrived to actively recruit people living in the Central and Western states to work on the railroads and farms and in the mines of the southwest (Gamio 1930). During the Great Depression migration from all over the world came to a halt and Mexicans were deported en masse, including many legal residents and U.S. citizens (Hoffman 1974). The flow from south of the border only returned with the advent of the Bracero Program in 1942 (Calavita 1992), which reactivated old regional networks of Mexican farm workers and paved the way for the entry of Mexicans into industrial cities of the Midwest, notably Chicago.

After the termination of the Bracero Program in 1964, Mexican immigration continued mostly outside the legal system. Many migrants were needed, and most wished only to come only on a seasonal or temporary basis (Massey et al. 1987). Given the absence of a legal program to accommodate these realities, most migrants entered and left without authorization (see Massey, Durand, and Malone 2002). As settlement patterns changed with the solidification of social networks and the formation of ethnic communities (Cornelius 1992), undocumented migration from Mexico increased steadily between 1965 and 1985. For the most part, men migrated first in undocumented status, followed by wives and children, some of whom entered legally if their male relatives had been able to adjust their status to LPR(Cerrutti and Massey 2001; Donato 1993).

Compared with Mexico, the history of immigration from other countries in Latin America is much more recent. In the Dominican Republic (DR), in-migration and especially out-migration was severely restricted until the 1960s by long-ruling dictator Rafael Leónidas Trujillo. After his assassination in 1961 the restrictions disappeared and political turmoil related to the succession ensued. In this context, for the first time a significant number of Dominicans were both willing and able to leave the country, perhaps including a nontrivial subset of people who would have migrated long before the fall of Trujillo if they could have, yielding a critical mass of people seeking to leave for the United States.

The Cuban Revolution and its shift to communism not only stimulated the creation of a large and influential population of Cuban émigrés in south Florida, but also made a deep imprint on the geopolitical climate of the region for decades to come. With high priority attached to the Alliance for Progress in 1961 and strong memories of the Cuban Missile Crisis in 1962, the Kennedy and Johnson administrations were eager to facilitate the departure of Dominicans from the island in order to defuse political pressure there. John B. Martin, the U.S. Ambassador to the DR at the time, reports that he accommodated the increased demand for visas by obtaining State Department authorization to build a new consulate and hire more consular officers to solve the “visa mess”—the large number of applicants waiting in line for residence visas (Grasmuck and Pessar 1991; Martin 1966, Chapter 3).

Official immigration statistics are consistent with these accounts, with a sharp upturn in both immigrant and non-immigrant admissions from the DR after 1961 (Grasmuck and Pessar 1991: Tables 1 and 2).³ These flows were comprised mostly of working and middle-class people from Santo Domingo and Santiago, the two main urban centers (Grasmuck and Pessar 1991; Portes and Grosfoguel 1994). As time elapsed, Dominican migration continued to grow, mostly in a documented fashion (see Grasmuck and Pessar 1991, 23 and references therein).

Table 1.

Means and standard errors on legal migration and legalization

	Mexico		DR		Nicaragua
	A. Men
Born in the U.S.	0.012	(0.001)	0.027	(0.003)	0.008	(0.001)
Non-U.S.-born
Ever a U.S. migrant	0.208	(0.002)	0.129	(0.006)	0.062	(0.003)
Migrants with 2+ trips	0.325	(0.006)	0.136	(0.018)	0.133	(0.018)
Ever a permanent resident	0.033	(0.001)	0.105	(0.006)	0.026	(0.002)
LPR | ever a U.S. migrant	0.162	(0.005)	0.837	(0.021)	0.491	(0.030)
Made first U.S. trip before becoming LPR	0.832	(0.012)	0.195	(0.024)	0.919	(0.024)
Pct. undocumented	0.936	(0.009)	0.956	(0.031)	0.916	(0.027)
Made last U.S. trip before becoming LPR	0.460	(0.016)	0.154	(0.022)	0.867	(0.029)
Pct. undocumented	0.936	(0.009)	0.956	(0.031)	0.916	(0.027)
No. of individuals in sample	30,495		2,664		5,286
	B. Women
Born in the U.S.	0.012	(0.001)	0.022	(0.003)	0.009	(0.001)
Non-U.S.-born
Ever a U.S. migrant	0.071	(0.001)	0.118	(0.006)	0.048	(0.003)
Migrants with 2+ trips	0.159	(0.008)	0.093	(0.015)	0.113	(0.019)
Ever a permanent resident	0.018	(0.001)	0.104	(0.006)	0.022	(0.002)
LPR | ever a U.S. migrant	0.254	(0.010)	0.896	(0.017)	0.539	(0.033)
Made first U.S. trip before becoming LPR	0.643	(0.020)	0.130	(0.020)	0.871	(0.030)
Pct. undocumented	0.949	(0.013)	0.923	(0.052)	0.928	(0.026)
Made last U.S. trip before becoming LPR	0.486	(0.021)	0.113	(0.018)	0.855	(0.032)
Pct. undocumented	0.949	(0.013)	0.923	(0.052)	0.928	(0.026)
No. of individuals in sample	31,462		2,877		5,771
	C. Both Sexes
Born in the U.S.	0.012	(0.000)	0.023	(0.002)	0.008	(0.001)
Non-U.S.-born
Ever a U.S. migrant	0.138	(0.001)	0.123	(0.004)	0.055	(0.002)
Migrants with 2+ trips	0.282	(0.005)	0.115	(0.012)	0.124	(0.013)
Ever a permanent resident	0.026	(0.001)	0.104	(0.004)	0.024	(0.001)
LPR | ever a U.S. migrant	0.186	(0.004)	0.866	(0.014)	0.513	(0.022)
Made first U.S. trip before becoming LPR	0.763	(0.011)	0.161	(0.015)	0.896	(0.019)
Pct. undocumented	0.940	(0.008)	0.944	(0.027)	0.922	(0.019)
Made last U.S. trip before becoming LPR	0.469	(0.013)	0.133	(0.014)	0.861	(0.021)
Pct. undocumented	0.940	(0.008)	0.944	(0.027)	0.922	(0.019)
No. of individuals in sample	61,957		5,553		11,057

Table 2.

Probability that a person with U.S. migration experience becomes a permanent resident through a specific sponsor

	Mexico		DR		Nicaragua
	A. Men
Family	0.052	(0.003)	0.799	(0.023)	0.092	(0.019)
Work	0.021	(0.002)	0.006	(0.005)	0.017	(0.008)
IRCA - General	0.063	(0.003)	0.010	(0.006)	0.050	(0.014)
IRCA - SAW	0.018	(0.002)	0.003	(0.003)	0.000	(0.000)
Other	0.004	(0.001)	0.016	(0.007)	0.255	(0.028)
No. with U.S. experience	6,353		343		328
No. with permanent residence	999		276		135
	B. Women
Family	0.160	(0.008)	0.862	(0.020)	0.224	(0.029)
Work	0.016	(0.003)	0.013	(0.006)	0.015	(0.009)
IRCA - General	0.061	(0.005)	0.010	(0.006)	0.025	(0.011)
IRCA - SAW	0.008	(0.002)	0.000	(0.000)	0.000	(0.000)
Other	0.003	(0.001)	0.010	(0.006)	0.214	(0.029)
No. with U.S. experience	2,222		339		277
No. with permanent residence	569		298		124
	C. Both sexes
Family	0.079	(0.003)	0.831	(0.015)	0.152	(0.017)
Work	0.020	(0.002)	0.010	(0.004)	0.016	(0.006)
IRCA - General	0.062	(0.003)	0.010	(0.004)	0.039	(0.009)
IRCA - SAW	0.015	(0.001)	0.002	(0.002)	0.000	(0.000)
Other	0.004	(0.001)	0.013	(0.005)	0.236	(0.020)
No. with U.S. experience	8,575		682		605
No. with permanent residence	1,568		574		259

Quotas, Central American Violence, and IRCA

Amendments to the INA in 1965 placed an annual ceiling of 170,000 on immigrants from the Eastern Hemisphere and a 20,000 person limit on immigration from any one country. Visas were allocated according to a seven-category preference system that sought to facilitate family reunification, attract skilled immigrants, and offer a safe haven to certain kinds of refugees. The 1965 amendments also established the first-ever numerical limits on immigration from the Western Hemisphere. Effective July 1, 1968, immigration from the Americas was capped at 120,000 per year, though neither the per-country limits nor the preference system initially applied there. It was not until 1976 that the INA was further amended to apply the country quotas and preference system to nations in the Western Hemisphere, and in 1978 the separate hemispheric quotas were abolished to create a single worldwide ceiling of 290,000 that was subsequently reduced to 270,000 by the Refugee Act of 1980, which took refugees out of the preference system and allowed foreign policy and humanitarian considerations to determine eligibility for refugee status (Coffino 2006) and asylum (Wasem and Ester 2006, 2). (The guidelines for withholding deportation would change after the passage of the Illegal Immigration Reform and Immigrant Responsibility Act–IIRIRA in 1996. I discuss some of these below in the context of the passage of the 1997 Nicaraguan Adjustment and Central American Relief Act–NACARA.)

Migration from Central America accelerated in the 1980s as political violence spread and insurgencies strengthened, leading thousands of Salvadorans, Guatemalans, and Nicaraguans to flee northward.⁴ After Nicaragua’s Samoza dictatorship was deposed by the communist-inspired Sandinista front in 1979, the Carter administration initially provided economic aid, but the Reagan administration suspended aid in 1981 and soon began to support paramilitary groups operating from Honduras, known as the Contras. Although Congress banned military aid to the Contras after the Sandinistas held elections in 1984, the Reagan administration nonetheless organized a trade embargo with Nicaragua and secretly diverted funds to paramilitary forces (Wasem 1997). Along with this diversion of funds and damaging effects of Hurricane Joan in 1988, the embargo paralyzed the Nicaraguan economy until a peace agreement was finally signed in 1989.⁵

The violence and economic deprivation resulting from this conflict motivated many Nicaraguans to leave during the late 1970s and 1980s (see Alvarado and Massey, this volume). Many of these people, especially those of slightly higher socioeconomic standing, went to the United States (Funkhouser 1992; Lundquist and Massey 2005; Massey and Sana 2003), typically entering the country without documents or overstaying tourist visas and later applying for some kind of relief.⁶

Compared with other Central Americans, Nicaraguans who applied for asylum enjoyed relatively higher success rates through the late 1980s. They also benefitted disproportionately from two major regularization programs: the Immigration Reform and Control Act (IRCA) of 1986 and the 1997 NACARA 1997 (both reviewed in greater detail below). Although Mexicans were the main beneficiaries of IRCA, Central Americans who left during the early stages of the civil conflict were able to take advantage of its main amnesty provision, general amnesty, which authorized regularization for those present continuously in the country since 1982.⁷ In addition to granting permanent residence to over 2 million Mexicans, some 136,000 Salvadorans, 50,000 Guatemalans, and 15,000 Nicaraguans also obtained LPR status through IRCA (Wasem 1997).

Given IRCA’s residency requirements, most Central Americans who fled the region after 1982 were not eligible for amnesty. While Nicaraguans had relatively high asylum approval rates during most of the 1980s, as high as 80 percent in fiscal year 1987, these fell by the end of the decade, reaching just 20 percent in 1990 (Wasem 1997, 14). Those denied asylum received unique treatment by way of the Nicaraguan Review Program (NRP), a special office established in 1987 under Attorney General Edwin Meese as a reaction to the Cardoza-Fonseca Supreme Court ruling that shifted the standard for a “well-founded” fear of persecution from demonstrating a clear probability of persecution to only showing a “reasonable” one (see Wasem 1997, note 16). The de facto temporary admission of Nicaraguans through establishment of the NRP would prove to be instrumental in their search for permanent residence.

The Post-IRCA Period

The Immigration Act of 1990 reorganized the preference system into three tiers that continued to favor family reunification but also increased annual employment-based immigration and created a new category of diversity visas available to nationals of countries that were underrepresented in recent flows. It also established a flexible worldwide cap that was significantly higher than the previous limits. The new system allowed unused visas in employment-based categories to be available for family preference immigrants the following year (and vice versa). The new cap was set at 700,000 for 1992–1994, and went to 675,000 in 1995. Currently, 71.1 percent (480,000) of the worldwide cap is devoted to family reunification, 20.7 percent (140,000) to employment-based admissions, and the remaining 8.2 percent (55,000) to diversity visas.

Family reunification provisions allow U.S. citizens many more opportunities to sponsor the entry of relatives than permanent residents. For example, immediate relatives of U.S. citizens (defined by the INA as spouses, unmarried minor children, and parents of adult U.S. citizens) are not subject to numerical limitation (see Wasem 2006).⁸ Citizens can also sponsor the entry of other relatives subject to numerical limits—unmarried adult children (the first family preference, set at 23,400 visas plus those unused in the fourth preference); married adult children (third family preference, set at 23,400 visas plus those not used in the first and second family preferences); and siblings age 21 or older (fourth family preference, set at 65,000 visas plus those not required by the other three preferences). In contrast, permanent residents can only sponsor spouses and unmarried children, both subject to numerical limitation (second family preference, set at 114,200 visas plus those unused by the first preference (Wasem 2006: Table 1), although 75 percent of these visas are exempt from the calculation of per-country numerical limits. Employment-based visa priorities follow an analogous procedure based on the skill level of the applicant and domestic needs for specific abilities (Wasem 2006).

All these exceptions explain why immigration consistently surpasses worldwide levels and per-country limits in any given year, as it has for most years since 1980 (United States. Office of Immigration Statistics 2007). The INA also specifies that per-country limits be held below 7 percent of the worldwide level of immigrant admissions (that is, of the effective number and not the worldwide limit), with a few exceptions. First, ever since the mid-1990s, 75 percent of the visas allocated to spouses and children of LPRs (i.e. the first tier of the second family preference) are not subject to the per-country ceiling. Second, since 2000 per-country ceilings for employment-based immigrants can be surpassed as long as visas are available within the worldwide limit for employment preferences (Wasem 2006, 5). Third, per-country ceilings also account for the number of unused visas that roll over from previous years, sometimes due to processing backlogs in a specific category. When the number of immigrants eligible for admission exceeds the per-country limit, immigrant visas are prorated according to the preference system. That is, within the family-based category, priority is given to unmarried adult children of U.S. citizens, then to spouses and unmarried children of LPRs, then to married children of citizens, and finally to adult siblings of citizens.⁹

Despite the fact that numerical limits to immigration have not been reduced, immigration policy nonetheless tightened during the latter part of the 1990s in diverse ways. Besides increased enforcement directed at unauthorized border crossers, especially in high transit corridors,¹⁰ certain aspects of the admission process became more rigid. In 1996, welfare reform legislation restricted the eligibility of legal immigrants for means-tested public programs and the IIRIRA toughened the requirements for immigrant sponsorship.¹¹ IIRIRA also significantly reduced the prospects for obtaining temporary and long-term relief from deportation. The old procedure for suspension of deportation was replaced by a new procedure (cancellation of removal) that applied stricter standards for obtaining relief and capped the number who could receive it at just 4,000 per year. Under the old suspension of deportation rules, relief could be obtained under circumstances of “extreme hardship to the alien, the alien’s citizen or permanent resident alien spouse, children, or parent” (Wasem 1997, 3). The new procedure changed these standards to “exceptional and extremely unusual hardship.” In addition, it increased the amount of time the person seeking relief had to be physically present in the United States (except for short absences) from 7 to 10 years. More important, it stated that the calculation of this time would end at the moment “the alien receives a notice to appear (the document that initiates removal proceedings) or when the alien commits a serious crime” (Wasem 1997: p. 4).

Although most of the rule changes enacted by IIRIRA applied only to cases started on or after April 1, 1997, that Act did contain language implying the retroactive application of the “time-stop” rule (Eig 1998), potentially affecting around 300,000 Central Americans. As a result, some 40,000 Nicaraguans who had received other forms of relief, were in the process of being deported, or still in asylum proceedings and who entered after the phase-out of the Nicaraguan Review Program in mid-1995 became vulnerable (Wasem 1997, 8)

NACARA addressed the retroactive nature of the cancellation of removal rules and gave preferential treatment to Nicaraguans relative to Salvadorans and Guatemalans. The former (along with their spouses and minor children) were allowed to adjust directly to permanent residence simply by presenting proof of their continuous presence in the United States before December 1, 1995 and with no numerical limit considerations (Eig 1998). Immigrant admissions from Nicaragua accordingly grew from the 4,000–6,000 range typical in the post-IRCA 1990s to 13,000 in 1999, 21,000 in 2000, and 11,000 in 2002 before returning to circa 4,000 per annum in 2003–2007 (United States Office of Immigration Statistics 2007).¹²

In sum, the three cases studied here differ markedly in terms of the immigration policy context in which they originated and evolved. Dominicans were granted relatively easy and swift access to permanent residence visas in response to the geopolitical situation in the Caribbean at the time. Nicaraguans, on the other hand, began migrating in significant numbers during a later time when both refugee flows and undocumented migration had significantly increased to create a less welcoming environment. Although Cold War politics did benefit them to a certain extent (Wasem 1997), their struggle for residence lasted for several years, was stricter in terms of the resources needed to achieve permanent residence, and relied on special regularization programs to a larger extent than Dominicans. Although Mexicans benefited from regularization under IRCA, given the large number of undocumented migrants who entered the country before 1982, this one-time event has gradually waned in importance and legal entry has grown progressively more difficult.

Family reunification provisions built into U.S. immigration law are an important avenue for legal migration and legalization in all groups. After all, it is the main immigration preference stipulated under the INA. Nonetheless, this avenue is more accessible to nationals of countries that early on built up a critical mass of LPRs relative to the total size of the flow. We therefore expect Dominicans to enjoy the greatest access to legal visas, followed by Nicaraguans, and then Mexicans. The next section looks at trends in LPR transition rates in order to verify whether they are consistent with these expectations and to assess the relative importance of family reunification, work-related criteria, and one-time regularization programs in conditioning legal migration from each country.

Legal Migration by Mexicans, Dominicans, and Nicaraguans

I use comparable data from surveys done by the Mexican Migration Project (MMP) and the Latin American Migration Project (LAMP) (see Donato et al. this volume, also see Massey and Capoferro 2004; Massey and Sana 2003). In addition to 7 Dominican and 9 Nicaraguan communities (surveyed in 1998–1999 and 2002–2003 respectively), I include data from 66 Mexican communities surveyed over roughly the same period (1998–2004), where information is most comparable with that from the LAMP samples. Specifically, I use the U.S. migration and permanent residence modules from the complete household roster (i.e. household members plus children of the head). As the evolution of migration from these countries has also followed clear gendered patterns (Massey et al. 2006), I examine trends for men and women separately.

Legal Migration and Legalization

Table 1 shows means and standard errors for legal migration and legalization dynamics by country in two separate panels according to the respondents’ gender. A rather small proportion of the sample consists of U.S.-born citizens: roughly 1.2 percent of the Mexican sample, 2.7 percent of the Dominican sample, and slightly below 1 percent of the Nicaraguan sample. Given the focus on legal migration and legalization, I eliminated these individuals from the ensuing analyses.

In terms of prevalence, Mexican males are the most likely to have U.S. migration experience, at 21 percent. Dominican men report the second highest migration prevalence, at 13 percent, but they are not significantly different from Dominican women, at 12 percent. Mexican women rank fourth behind Dominicans, at 7 percent, while 6 percent of Nicaraguan men and 5 percent of Nicaraguan women reported U.S. migration experience. Whereas almost a third of Mexican men with U.S. migration experience reported more than one U.S. trip (reflecting shorter distances, seasonal occupations, and lower rates of settlement—see Lindstrom 1996), only 13 percent of Dominican and Nicaraguan men had a second trip. Female gradients are relatively similar but more nuanced. Whereas 16 percent of Mexican women reported more than one U.S. trip, only 11 percent and 9 percent of Nicaraguan women did so.

A clear inter-country (and gendered) pattern can be seen in the prevalence of U.S. permanent residence, where Dominicans stand alone vis-à-vis the other groups. Holding a green card is a relatively rare event in most countries, with the prevalence in the 2–3 percent range for Mexico and Nicaragua. In clear contrast to Nicaraguans and Mexicans, 10 percent of Dominicans in the sample (including non-migrants) reported holding U.S. permanent residence. In addition, contrary to what happens in Mexico, Dominican women are just as likely as men to hold a green card.

The higher prevalence of green cards among men in Mexico reflects the higher propensity of males to become migrants in the first place. If we only consider those with U.S. migration experience, Mexican migrant women are in fact more likely than migrant men to have permanent residence (25 percent vs. 16 percent). Both groups, however, are less likely to have permanent residence than migrants of other national origins. Dominican migrant men and women had by far the highest probability of having a green card, with rates of 84 and 90 percent, respectively, which are not significantly different from each other. Nicaraguan women and men come in a distant but still substantial second, with figures of 54 percent and 49 percent, respectively.

As expected, given the evolution of immigration from these countries and the policies in place when the flows began and expanded, the timing of legalization also varies conspicuously across countries. Not surprisingly, the vast majority of Mexicans and Nicaraguans made their first U.S. trip before obtaining legal permanent residence. Some 83 percent of Mexican men and 92 percent of those from Nicaragua who eventually obtained permanent residence entered the U.S. before obtaining their green card, while only 20 percent of Dominican men did so. In all cases women were less likely to report entering the U.S. before obtaining legal status, but only in Mexico was the difference large and statistically significant (83 percent of men vs. 64 percent of women), suggesting that Mexican women are less likely to migrate without documents and that they tend to follow a parent or spouse who had already obtained permanent residence (Cerrutti and Massey 2001; Donato 1993).

Compared with first trips, the percent of all country-gender groups who obtain a green card after initiating the last U.S. trip is lower; but the difference is large only for that group evincing large proportions with more than one trip: Mexicans, especially males. The relative share of Mexican men who entered on their last trip before obtaining permanent residence was 46 percent, compared with 83 percent on the first trip. Among women, the difference was 64 percent on the first trip versus 48 percent on the last trip. In contrast, the differences are not statistically significant among Dominicans and Nicaraguans.

In recent years, most immigrant admissions have been adjustments of status within the United States rather than new arrivals processed in embassies and consulates abroad (see Wasem 2006, Figure 2). Significant proportions of Latin Americans admitted for permanent residence have previous experience in the United States, either as temporary workers, students, business visitors, tourists, or undocumented workers (Massey and Malone 2003). These trends in preadmission migration may simply reflect where people are choosing to wait for their visas to become available. Regardless of country of origin, gender, or previous number of trips, some 92–96 percent of LPRs who entered the United States before achieving legal residence reported doing so without documents or with a tourist visa. The data thus suggest two very different forms of legal immigration: initial entry with legal residence documents versus undocumented entry or visa violation followed by later legalization. The first is characteristic of Dominicans and to a lesser extent Mexican women, whereas the second pertains to Mexican men and Nicaraguans of both sexes.

Types of Sponsorship

Table 2 shows the sponsorship categories by which men and women from each country obtained their green cards.¹³ It is clear that the bulk of the Dominican advantage in legal transition stems from their use of family members as sponsors, which is also consistent with the evidence on preadmission U.S. experience just presented and historical accounts of Dominican migration (see Grasmuck and Pessar 1991, chapter 2 and references therein). The likelihood that Dominican migrant men and women obtained permanent residence through a family member is very high—79 percent and 86 percent, respectively. These figures account not only for the vast majority of Dominican LPRs but for Dominican migrants generally. In contrast, the share obtaining legal residence through a relative is much lower in all other countries and never above 25 percent for either men or women. Thus, Nicaraguan and Mexican women come in a distant third and fourth with respective shares of 22 percent and 16 percent, and the figure is only 5 percent for Mexican men. The share achieving permanent residence through work-related criteria is much smaller, at or below 2 percent, for all groups.

Legalization programs were as or more instrumental in providing Mexicans and Nicaraguans with access to permanent residence. In addition to the amnesty for long-term undocumented residents of the United States, IRCA also offered a special legalization to farm workers who had worked in agriculture during the year prior to the law’s passage. Together these two provisions accounted for the legalization of 7.7 percent of all Mexican migrants, roughly the same as the 7.8 percent who achieved legal status through a family member (see Table 2, panel C). Indeed, despite the one-time nature of IRCA’s regularization programs, these provisions represented a more important avenue for legalization among Mexican men than family provisions (8 percent vs. 5 percent, a significant difference at p<.05).

Only 4 percent of Nicaraguan migrants benefited from IRCA. Many more of them benefited from NACARA. Although this specific program was not included as a separate option in the questionnaire and the actual open-ended responses are not available in the Nicaraguan database, reports from the fieldwork coordinator¹⁴ and analyses of the timing of migration and legalization (not shown here) suggest that at least half of the 24 percent of Nicaraguan migrants who reported being sponsored by “other” provisions used NACARA criteria, with the other half possibly involving asylum cases. A conservative assessment thus suggests that IRCA and NACARA were just as relevant as family provisions in providing Nicaraguans with access to legal residence.

Family Provisions in More Detail

Given the relevance of family provisions in the legalization process and the rather large differences in rates of LPR acquisition across countries, I consider which specific family members were responsible for sponsorship. Table 3 shows the share of migrants who achieved permanent residence through specific family ties by country and gender. As both residents and citizens can sponsor the entry of spouses and minor children, it is unsurprising that spouses and parents are the most likely conduits to legal permanent residence, though there are interesting differences across countries in which tie dominates. Spouses are the most common sponsors of women from Mexico and Nicaragua, accounting for 8.9 percent of sponsorships among the former and 11.4 percent among the latter. Women are more likely than men in both countries to rely on spouses for permanent residence. Mexican and Nicaraguan men tend to rely more on parents than on spouses, though they still have lower probabilities than women of becoming a legal resident through a parent. Whereas 2.6 percent of Mexican men and 3.8 percent of Nicaraguan men with U.S. experience obtained residence through a parent, only 1.6 percent of the former and 1.3 percent of the latter were sponsored by a wife.

Table 3.

Probability that a person with U.S. migration experience becomes a permanent resident sponsored by a specific relative

	Mexico		DR		Nicaragua
	A. Men
Spouse	0.016	(0.002)	0.262	(0.025)	0.013	(0.007)
Parent	0.026	(0.002)	0.485	(0.028)	0.038	(0.012)
Child	0.004	(0.001)	0.019	(0.008)	0.029	(0.011)
Sibling	0.005	(0.001)	0.032	(0.010)	0.013	(0.007)
	B. Women
Spouse	0.089	(0.006)	0.301	(0.026)	0.114	(0.022)
Parent	0.050	(0.005)	0.497	(0.028)	0.065	(0.017)
Child	0.015	(0.003)	0.035	(0.010)	0.040	(0.014)
Sibling	0.006	(0.002)	0.029	(0.009)	0.005	(0.005)
	C. Both Sexes
Spouse	0.035	(0.002)	0.282	(0.018)	0.059	(0.011)
Parent	0.032	(0.002)	0.491	(0.020)	0.050	(0.010)
Child	0.007	(0.001)	0.027	(0.007)	0.034	(0.009)
Sibling	0.006	(0.001)	0.031	(0.007)	0.009	(0.005)

The situation of Dominicans is, again, quite different. As in the Mexican and Nicaraguan cases, Dominican men are considerably more likely to be sponsored by a parent than a spouse (49 percent vs. 26 percent of Dominican men with U.S. experience). However, in contrast to Mexican and Nicaraguan men, who tend to rely on spouses at a much lower rate than their female counterparts, Dominican men are almost as likely to obtain a green card through a spouse, with 26 percent reporting legalization through a wife versus 30 percent of Dominican women reporting legalization through a husband (difference not significant at p<.05). Likewise in contrast to Mexico and Nicaragua, where women are more likely to be sponsored by a husband than by parents, Dominican women were much more likely to be sponsored by parents than by husbands (i.e. 50 percent vs. 30 percent among Dominican women with U.S. experience).

The foregoing differences in patterns between Mexico, Nicaragua, and the DR are eclipsed by sheer differences in the magnitude of rates across countries. Dominican men and women are far more likely than Mexicans and Nicaraguans to be sponsored by either a spouse or a parent. Dominican women (at 30 percent) are 3.4 and 2.6 times more likely to be sponsored by husbands than are Mexican (9 percent) or Nicaraguan (11 percent) women. they are 9.8 and 7.6 times more likely to be sponsored by a parent (50 percent for Dominicans vs. 6.5 percent and 5 percent for Nicaraguans and Mexicans, respectively). These differences are even more striking for men. Dominican men (at 26 percent) are 16 and 20 times more likely to be sponsored by wives than Mexican (1.6 percent) or Nicaraguan (1.3 percent) men. Likewise, they are 18 and 12 times more likely to be sponsored by a parent (49 percent among Dominicans with U.S. experience compared with only 3.8 percent and 2.6 percent of Nicaraguan and Mexican men).

In sum, Dominicans’ much higher likelihood of becoming LPRs seems to stem from their greater ability to take advantage of family reunification provisions of U.S. immigration law, drawing more frequently on ties to parents and spouses (in that order) rather than regularization programs after an undocumented spell. Given their much higher likelihood of obtaining a green card through parents or spouses, we would not expect Dominicans to have lower rates of doing so than Mexicans or Nicaraguans at any age, but especially during childhood. Table 4 confirms this conjecture by presenting age-specific transition probabilities by country for two major periods: pre-IRCA (1965–1985) and post-IRCA (1986–survey year).¹⁵ Even in the pre-IRCA period, when significant Dominican migration had just started and Mexican migration was already well-established, LPR transition rates were much higher in the former than the latter in all age groups. For instance, whereas 1.9/10,000 of all Mexicans age 0–14 (including non-migrants) obtained a green card, the equivalent figure is 8.1 times higher for Dominicans and 31 times higher for Nicaraguans, suggesting that the pattern of family migration was initiated much earlier from the DR than from Mexico, most likely thanks to Dominicans’ privileged access to legal documents.

Table 4.

Age-specific transition probabilities to permanent resident per 10,000 population in the country of origin

	Mexico		DR		Nicaragua
	A. 1965 – 1985
0 – 14	1.9	(0.23)	15.5	(2.17)	0.5	(0.28)
15 – 34	9.9	(0.63)	63.1	(5.18)	2.9	(0.83)
35+	4.6	(0.73)	45.1	(7.03)	6.9	(2.18)
All ages	5.1	(0.27)	36.6	(2.36)	2.1	(0.43)
	B. 1986 - survey year
0 – 14	4.3	(0.46)	25.7	(4.01)	3.8	(0.95)
15 – 34	13.8	(0.67)	75.3	(5.79)	20.9	(1.97)
35+	9.1	(0.69)	48.1	(5.95)	27.8	(2.91)
All ages	9.8	(0.37)	52.9	(3.19)	17.1	(1.15)
	C. All periods
0 – 14	2.9	(0.22)	19.1	(1.92)	2.0	(0.42)
15 – 24	17.0	(0.53)	69.1	(3.74)	12.4	(1.10)
35+	10.5	(0.60)	48.3	(4.45)	20.8	(2.03)
All ages	10.1	(0.26)	44.4	(1.88)	9.6	(0.60)

Notes: Standard errors of estimates are indicated between brackets

Mexican and Nicaraguan legalization rates increased considerably after the passage of IRCA, which also benefited Dominicans though in relatively lower terms given the large size of the Dominican LPR versus undocumented population (more on this below). The probability that a Mexican obtained legal permanent residence almost doubled, from 5.1 to 9.8 per 10,000 persons between the two periods. This increase pales in comparison to that of Nicaraguans, whose transition probabilities increased from 2.1 to 17.1 per 10,000 persons and became higher than those of Mexicans after age 15 (see Table 4). Both the magnitude and older age pattern of the increase in the Nicaraguan transition probabilities are consistent with the notion that Nicaraguans benefited not only from IRCA after 1986, but also from NACARA after 1997 (see United States Office of Immigration Statistics 2007, Table 2).

Potential Misreporting of LPR Status

As transition rates to permanent residence are much larger among migrants from the DR than among those from Nicaragua or Mexico, one might suspect that Dominicans are possibly over-reporting the possession of legal residence documents. While this possibility cannot be entirely ruled out, there is considerable evidence that this is not the case. As mentioned above, studies using other data sources also suggest Dominicans are not as likely to be undocumented as other groups. Evidence based on administrative data and indirect estimation of undocumented migrants seems to provide a similar picture with respect to Dominican migration. Permanent admissions from the DR reported in immigration statistics are quite sizable relative to the size of the country. Even by the beginning of the 1980s, the DR had the third highest rate of legal admissions per 10,000 population in the home country, at 32.3 (see Grasmuck and Pessar 1991,Table 3). This figure is remarkably similar to and within the 95 percent confidence interval of the annual average rate estimated with LAMP data for the pre-IRCA period (36.6, see Table 4, Panel A).

If there were a large undocumented component not registered in LAMP data, these two figures would most likely differ substantially. Likewise, around 8.3 percent of Dominicans living in New York enumerated in the 1980 census are estimated to be undocumented (i.e. 14,000 out of 169,000, see Grasmuck and Pessar 1991, 22, 163).¹⁶ While most Dominicans in the sample reported going directly to the U.S. mainland (and specifically to New York), it should be noted that the Dominican undocumented component in recent years has been around six times higher in Puerto Rico, which has traditionally been the main entry hub for undocumented Dominicans into the United States (Duany 2005).

Finally, even in the presence of misreporting, the bias in Dominican self-reports relative to those of Mexicans and Nicaraguans would need to be rather large to change the conclusions of this study. One would expect groups with a lower likelihood of return to over-report legal status. As Dominicans are indeed less likely to return relative to both Mexicans and Nicaraguans (Riosmena 2005, chapter 4), this is indeed a possibility.¹⁷ But the amount of additional Dominican misreporting would have to be extremely large to change the results presented here. For instance, assuming that no Nicaraguans falsely reported having a green card, differences between Dominicans and Nicaraguans would be reduced substantially (say, such that their 95 percent confidence intervals would overlap) only if around 25 percent of Dominican migrants were falsely reporting being a LPR. This figure would need to be much larger for Mexicans, such that around 60 percent of Dominicans would need to misreport for results to change substantially. These differences seem rather large, especially as they represent lower bounds: they would need to be even larger in the event that misreporting bias for Mexicans and Nicaraguans was not negligible.¹⁸

Conclusions

This article has presented evidence to suggest that immigration policy leaves an imprint not only on the particular cohorts of people who experience it, but also on subsequent immigrant generations. It does so by creating or restricting a critical mass of people with legal permanent residence who are able to sponsor the legal migration of relatives. If a critical mass of people exists but lacks legal residence, the effect of social networks is to reproduce undocumented or gray-status migration, a cycle that can only be broken by one-time regularization programs.

Of the three groups studied here, Dominicans displayed by far the highest likelihood of obtaining LPR at all ages, mostly through family reunification provisions such as sponsorship by a parent or spouse, leading to a high rate of legal immigration as opposed to legalization after one or more undocumented entries and after considerable durations of time spent in undocumented status. Nicaraguans evinced the second highest transition rates (though a distant second), also with relatively few gender differences. However, their transition to legal status was mostly achieved through post-undocumented migration regularization enabled by special programs authorized by IRCA and NACARA. These legalization programs supplemented a significant number of legal immigrants sponsored by family members and admitted under relatively generous asylum provisions. Mexican migrants displayed the lowest likelihood of achieving legal status and presented sharper gender differentials. Women were more likely to legalize (mostly sponsored by husbands), whereas men benefited from IRCA as much as from family ties, and the family ties they drew upon were mainly parents.

Cross-national differences in the degree and mode of transition into LPR status are the result of three main factors: the interplay between U.S. foreign policy concerns and conditions prompting the original out-migration; the effect of social networks in reproducing the original character of the outflow; and differences in kinship ties to LPRs that are not entirely explained by prior differences in naturalization patterns. All of these have made an imprint on the subsequent character of immigration in addition to the effects of origin country conditions, yielding cross-country differences in gendered and family patterns of migration.

Interplay Between Legal Context of Reception and Conditions at Origin

Dominicans encountered a relatively favorable legal context of reception during the initiation of international migration, originally spurred by the sudden relaxation of emigration restrictions and the unstable political situation in the DR in the early 1960s. This legal context of reception was not more favorable than that encountered by Puerto Ricans (U.S. citizens by birth, though members of a geographically, linguistically, and culturally-distinct nation—see Duany 2002) or Cubans (eligible for quasi-immediate admission through a variety of mechanisms without numerical limits—see Eig 1998; Portes and Rumbaut 2006; Wasem 1997). But it was indeed more lax and much more favorable than the legal context faced by Nicaraguans and Mexicans, especially at the time when these flows originated.

Dominican migration began during a time of more liberal immigration policies, when nations in the Western Hemisphere faced no numerical limitations and were unconstrained by preference systems that favored but nonetheless curbed family migration. Although Mexican immigration also originated in this context, the conditions stimulating migration from these two countries were rather different. Dominican emigration was jump-started by a newfound freedom to emigrate, fueled by political and economic instability, and facilitated by a liberal (albeit short-lived) policy of generously granting nonimmigrant and immigrant visas (Grasmuck and Pessar 1991; Martin 1966).

The Cumulative Causation of Legal Migration

Many studies have demonstrated the relevance of migrant networks in facilitating the migration process (Flores, Hernández-León and Massey 2004; Fussell and Massey 2004; Massey and Espinosa 1997; Munshi 2003; Palloni et al. 2001). As more migrants enter the flow, the costs of migration drop for future immigrants as they have access to networks of people able to provide information and assistance and thereby reduce uncertainty surrounding the move. In the context of a preference system favoring family reunification, specific kinship ties to migrants (and, even more importantly, to naturalized citizens) become critical in determining access to legal migration. When the legal context of reception (initially or by way of posterior legalization programs) and conditions associated with emigration create or fail to create a critical mass of individuals with legal documents, subsequent chain migration reinforces the legal status of the original immigrants (see also Massey et al. 2002).

Issues worthy of future research is to ascertain whether a critical mass of LPRs tends to attract a larger flow of migrants from a given country and whether that flow tends to be legal or undocumented. Of course, one would need to control for the initial conditions associated with the original out-migration from each country. There is some evidence that men with direct ties to LPR family members have a higher likelihood of U.S. migration relative to those having non-LPR migrant relatives (Riosmena 2005, chapter 4); but the analysis did not specify whether the likelihood was larger because individuals migrated as a part of a family unit or because such ties were with legal immigrants, undocumented migrants, or both.

Effectiveness in the Use of Family Reunification Provisions

No matter how broad or narrow the initial critical mass of LPRs, some groups have been more effective in taking advantage of the preference system to sponsor additional immigrants. Dominicans did not benefit from a more liberal immigration policy framework for long, but long enough for them to quickly build up a large legal population, allowing permanent family-based immigration to gain considerable momentum. Transition to permanent residence (i.e. emigration in the Dominican case) was already remarkably high during the first twenty years of Dominican immigration (see Table 3); and even in an era of national quotas (somewhat flexible but clearly below their potential levels, see Jasso and Rosenzweig 1986, 1989), Dominicans continue to exhibit high rates of legal immigration.

Given that that U.S. immigration law explicitly favors family reunification, especially the relatives of citizens and to a lesser extent those of permanent residents, it is not surprising that legal Dominican immigrants were mostly sponsored by parents and spouses. As naturalization data are not available in the Dominican LAMP survey, it is not possible to directly observe whether Dominicans in the sample were more instrumental in sponsoring more relatives (and some of them faster and in an untapped fashion) because of their citizenship status. While this is possible, naturalization is most likely not driving the flows given the relatively low naturalization rates observed for Dominicans during most of the period studied (e.g. Woodrow-Lafield et al. 2004),¹⁹ though naturalization could also be facilitating transnational life for a subset of Dominicans (Gilbertson and Singer 2003). It is, however, a matter of future research to ascertain the relevance of human capital characteristics in explaining differences in sponsorship rates across countries.

The Role of Documentation in Explaining Gendered Patterns of Migration

Scholars have suggested that the gendered pattern of Dominican migration (i.e. few gender differentials and a high likelihood of female migration) stem from the less patriarchal family systems that prevail in the DR (Massey et al. 2006; Massey and Sana 2003). Although it is indeed likely that migration and return patterns for Dominican males and females are similar because the absence of patriarchal constraints allows women to migrate independently and gives them more bargaining power in the household, the decision to migrate is nonetheless greatly facilitated when individuals have access to permanent residence. The sponsor-specific legalization rates and the actual sponsors most often used by Dominicans suggest that families are better able to reunite legally in the United States, and are thus less likely to return to the DR in the short term (see Riosmena 2005, chapter 4). At the very least, differences in access to legal status help to explain how differences in family systems are translated into strikingly different gendered patterns of migration.

Biography

Fernando Riosmena is assistant professor in the Population Program, Institute of Behavioral Science, at the University of Colorado at Boulder. His research interests intersect the fields of formal and social demography with an emphasis on how demographic processes are associated with the social mobility, wellbeing, and development in Latin American societies and immigrant communities from said region in the United States.