Unintended Consequences of US Immigration Policy: Explaining the Post-1965 Surge from Latin America

Abstract

The year 1965 is often cited as a turning point in the history of US immigration, but what happened in the ensuing years is not well understood. Amendments to the Immigration and Nationality Act passed in that year repealed the national origins quotas, which had been enacted during the 1920s in a deliberate attempt to limit the entry of Southern and Eastern European immigrants—or more specifically Jews from the Russian Pale and Catholics from Poland and Italy, groups at the time deemed “unassimilable.” The quotas supplemented prohibitions already in place that effectively banned the entry of Asians and Africans. The 1965 amendments were intended to purge immigration law of its racist legacy by replacing the old quotas with a new system that allocated residence visas according to a neutral preference system based on family reunification and labor force needs. The new system is widely credited with having sparked a shift in the composition of immigration away from Europe toward Asia and Latin America, along with a substantial increase in the number of immigrants.

Indeed, after 1965 the number of immigrants entering the country did increase, and the flows did come to be dominated by Asians and Latin Americans. Although the amendments may have opened the door to greater immigration from Asia, however, the surge in immigration from Latin America occurred in spite of rather than because of the new system. Countries in the Western Hemisphere had never been included in the national origins quotas, nor was the entry of their residents prohibited as that of Africans and Asians had been. Indeed, before 1965 there were no numerical limits at all on immigration from Latin America or the Caribbean, only qualitative restrictions. The 1965 amendments changed all that, imposing an annual cap of 120,000 on entries from the Western Hemisphere. Subsequent amendments further limited immigration from the region by limiting the number of residence visas for any single country to just 20,000 per year (in 1976), folding the separate hemispheric caps into a worldwide ceiling of 290,000 visas (in 1978), and then reducing the ceiling to 270,000 visas (in 1980). These restrictions did not apply to spouses, parents, and children of US citizens, however.

Thus the 1965 legislation in no way can be invoked to account for the rise in immigration from Latin America. Nonetheless, Latin American migration did grow. Legal immigration from the region grew from a total of around 459,000 during the decade of the 1950s to peak at 4.2 million during the 1990s, by which time it made up 44 percent of the entire flow, compared with 29 percent for Asia, 14 percent for Europe, 6 percent for Africa, and 7 percent for the rest of the world (US Department of Homeland Security 2012). The population of unauthorized immigrants from Latin America also rose from near zero in 1965 to peak at around 9.6 million in 2008, accounting for around 80 percent of the total present without authorization (Hoefer, Rytina, and Baker 2011; Wasem 2011). How this happened is a complicated tale of unintended consequences, political opportunism, bureaucratic entrepreneurship, media guile, and most likely a healthy dose of racial and ethnic prejudice. In this article, we lay out the sequence of events that culminated in record levels of immigration from Latin America during the 1990s. We focus particularly on the case of Mexico, which accounted for two-thirds of legal immigration during the decade and for three-quarters of all illegal migration from the region.

The unintended legacy of immigration reform

Paradoxical as it may seem, US immigration policy often has very little to do with trends and patterns of immigration. Even when policies respond explicitly to shifts in immigration, rarely are they grounded in any real understanding of the forces that govern international migration. Instead, over time the relative openness or restrictiveness of US policies is more strongly shaped by prevailing economic circumstances and political ideologies (Timmer and Williamson 1998; Massey 1999; Meyers 2004). In the United States, especially, immigrants carry significant symbolic weight in the narrative of American peoplehood (Smith 1997, 2003), and how they are depicted in the media, portrayed by politicians, and treated by legislators probably reveals more about America’s aspirations and hopes—and its fears and insecurities—than anything to do with immigration itself (Tichenor 2002; Ngai 2003).

Americans’ fears and apprehensions prevailed in the 1920s and led to the passage of the discriminatory quotas. In response to rising economic inequality along with new currents of scientific racism, xenophobia, and conservative ideology, the Quota Acts of 1921 and 1924 sought to reduce the number of immigrants entering the United States and shift their origins away from Southern and Eastern Europe and toward Northern and Western Europe, while maintaining outright prohibitions on all immigration from Asia and Africa (Zolberg 2006). In contrast, hopes and aspirations were the dominant force of the 1960s, and legislators sought to enact liberalizing reforms and introduce greater openness into the immigration system. As the civil rights movement gathered force, discriminatory quotas against certain Europeans and prohibitions on African and Asian immigration came to be seen as intolerably racist and were duly repealed by Congress in 1965.

At that time, the United States was nearing the end of a “long hiatus” with respect to immigration (Massey 1995). Indeed, with so few immigrants settling in the United States, the foreign-born percentage in 1970 dipped below 5 percent for the first and only time in US history (compared with 12.4 percent in 2010). As a result, immigration was not a very salient issue in 1965, except to a few conservative senators who actually favored the restrictive quotas. To the extent that legislators were concerned at all, their attention focused mainly on the consequences of opening the door to Asian and African immigration, not immigration from Latin America. Newly elected Senator Ted Kennedy, who served as floor leader for the bill, assured his colleagues that “our cities will not be flooded with a million immigrants annually…[and] the ethnic mix of this country will not be upset…. Contrary to the charges in some quarters, [the bill] will not inundate America with immigrants from any one country or area, or the most populated and economically deprived nations of Africa and Asia” (Congressional Digest, May 1965, p. 152).

The second order of business for immigration reformers with a civil rights agenda was ending a scheme admitting short-term foreign workers known as the Bracero Program. Originally established in 1942 as a temporary wartime measure, the program was extended by Congress and massively expanded in the latter half of the 1950s. By the 1960s, however, the Bracero Program had come to be seen as an exploitive labor regime on a par with Southern sharecropping, and in 1964, over vociferous objections from Mexico, Congress voted to terminate it (Calavita 1992; Massey, Durand, and Malone 2002). The program was phased out between 1965 and 1967 and the flow went to zero in 1968, the same year the new cap on immigration from the Western Hemisphere took effect.

Immigration was a back burner issue for most Americans in the 1960s. They concerned themselves instead with civil rights, the war in Vietnam, the sexual revolution, and urban riots. Nonetheless, out of view of both citizens and politicians, immigration from Latin America had been quietly but steadily growing, especially from Mexico. Owing to the Bracero Program, however, the lion’s share of the migration was temporary and circular and, hence, invisible to citizens. During the period 1955–59, around half a million Mexicans were entering the country each year, the number fluctuating around 450,000 temporary Bracero migrants and 50,000 permanent residents (Massey 2011). Given this large annual inflow, the sudden elimination of the Bracero Program clearly would have dramatic consequences on migration between Mexico and the United States; and with the imposition of a hemispheric cap, and eventually country quotas, the displaced temporary migrants were not going to be accommodated within the system for legal immigration.

In short, as a result of shifts in US immigration policy between the late 1950s and the late 1970s, Mexico went from annual access to around 450,000 guestworker visas and a theoretically unlimited number of resident visas in the United States (in practice averaging around 50,000 per year) to a new situation in which there were no guestworker visas and just 20,000 resident visas annually. The effect of these new limits on the system of Mexican migration that had evolved during the Bracero era was predictable and is illustrated in Figure 1, which shows Mexican entries into the United States in three categories by legal status for the period 1955–1995: temporary migrants (Braceros before 1965 and H-visa holders thereafter), legal immigrants (those entering with permanent resident visas), and illegal immigrants (proxied here by a series based on the annual number of apprehensions per thousand Border Patrol agents—see Table A1). Former illegal migrants adjusting under the 1986 Immigration Reform and Control Act (IRCA) have been removed from the series on legal immigrants as they are already accounted for in the series on illegal immigrants.

FIGURE 1. Mexican immigration to the United States in three categories, 1955–95.

SOURCE: US Department of Homeland Security (2012). See text and Table A1.

Table A1.

Legal and illegal US migration from Mexico and political and policy indicators, 1955–2009

Year	Legal migrants (1)	Temporary migrants (2)	Estimated illegal migrants (3)	Apprehensions (4)	Deportations (5)	border Patrol agents (6)	border Patrol budget ($1,000) (7)	Line- watch hours (8)	Percent conservative (9)
1955	50772	398650	164035	242608	17695	1479	NA	919000	NA
1956	65047	445197	45475	72442	9006	1593	NA	990000	NA
1957	49154	436049	29813	44451	5989	1491	NA	927000	NA
1958	26712	432857	24437	37242	7875	1524	NA	947000	NA
1959	23061	437643	20131	30196	8468	1500	NA	932000	NA
1960	32084	315846	19847	29651	7240	1494	NA	929000	NA
1961	41632	291420	18093	29817	8181	1648	NA	1024000	NA
1962	55291	194978	18756	30272	8025	1614	NA	1003000	NA
1963	55253	186865	26205	39124	7763	1493	NA	928000	NA
1964	32967	177736	29366	43844	9167	1493	NA	928000	NA
1965	37969	20286	37116	55340	10572	1491	41.7	920000	18.2
1966	45163	8647	60195	89751	9680	1491	49.2	930000	19.3
1967	42371	7703	72508	108327	9728	1494	52.3	937566	20.5
1968	43563	0	101543	151705	9590	1494	47.8	941860	21.6
1969	44623	0	134964	201636	11030	1494	46.3	909623	22.7
1970	44469	0	177125	277377	17469	1566	47.0	1148854	23.8
1971	50103	0	224053	348178	18294	1554	55.9	1118710	24.9
1972	64040	0	263288	430213	16883	1634	60.5	1259430	26.0
1973	70141	0	338511	576823	17346	1704	64.7	1279198	27.8
1974	71586	0	408257	709959	19413	1739	62.6	1253191	29.5
1975	62205	0	377367	680392	24432	1803	63.1	1465423	30.1
1976	74449	0	394883	781474	38471	1979	67.3	1775890	31.7
1977	44646	2011	464160	954778	31263	2057	67.9	1740446	31.8
1978	92681	2271	446170	976667	29277	2189	78.1	1762616	33.8
1979	52479	1725	427033	998830	26825	2339	65.5	1935926	33.9
1980	56680	3323	329098	817479	18013	2484	82.6	1815797	34.0
1981	101268	4719	357788	874433	17379	2444	85.6	1929448	33.4
1982	56106	4966	356705	887481	15216	2488	98.7	1871173	32.7
1983	59079	5014	473850	1172306	19211	2474	110.1	1976162	34.5
1984	57820	5336	473229	1170769	18696	2474	114.1	1843179	36.5
1985	61290	9622	392017	1266999	23105	3232	141.9	1912895	36.5
1986	66753	12029	452602	1671458	24592	3693	150.9	2401575	35.8
1987	72511	13393	307752	1139606	24336	3703	194.6	2546397	32.5
1988	95170	16802	255783	949722	25829	3713	205.3	2069498	36.7
1989	66933	27168	232418	865292	34427	3723	246.4	2570311	32.3
1990	57667	16891	292606	1092298	30039	3733	262.6	2781317	36.9
1991	54622	19148	309918	1131510	33189	3651	298.7	2638720	32.2
1992	91658	19813	294298	1199560	43671	4076	399.3	2642227	34.4
1993	109108	23169	318661	1263490	42542	3965	361.7	2713937	36.3
1994	107012	24896	244124	1031668	45674	4226	400.0	3073758	37.0
1995	87073	26512	271297	1324202	50924	4881	451.5	3397049	37.1
1996	160138	35949	263674	1549876	69680	5878	568.0	4073542	37.2
1997	144641	35949	205371	1412953	114432	6880	717.4	4807669	36.1
1998	129970	66197	194911	1555776	174813	7982	877.1	6660692	35.0
1999	146432	86424	189080	1579010	183114	8351	916.8	8740258	34.5
2000	173161	104155	181984	1676438	188467	9212	1055.4	8999552	34.0
2001	205560	116157	131200	1266214	189026	9651	1146.5	9802081	34.5
2002	218822	118835	96476	955310	165168	9902	1416.3	9183667	35.1
2003	115585	114673	88375	931557	211098	10541	1420.3	9457060	36.8
2004	175411	117999	108175	1160395	240665	10727	1212.9	9830697	38.4
2005	161445	169786	107069	1189108	246431	11106	1525.3	10474078	36.7
2006	173749	225680	86419	1089136	280974	12603	1561.8	11562715	34.7
2007	148640	300346	62179	876787	319382	14101	2277.5	14055363	35.2
2008	189989	360903	41365	723840	358886	17499	3002.2	17852594	35.7
2009	164920	301558	27637	556032	392862	20119	3501.3	20657122	40.0

SOURCES: Columns 1–8: US Dept. of Homeland Security; Column 9: General Social Survey.

The number of apprehensions recorded in any year is a joint function of the number of migrants attempting unauthorized entry and the effort made to apprehend them. Raw apprehension counts in themselves are seriously flawed as indicators of the volume of illegal migration. However, once the enforcement effort is accounted for, here by dividing the number of migrants arrested by the number of Border Patrol agents looking for them, the adjusted apprehension counts offer serviceable indicators of trends in illegal migration. We do not claim that the estimates yielded by this procedure capture the true number of undocumented entries, only that fluctuations in them over time reflect trends in the volume of undocumented migration.

As already noted, by the late 1950s a massive circular flow of Mexican migrants had become deeply embedded in employer practices and migrant expectations and had come to be sustained by well-developed and widely accessible migrant networks (Massey, Durand, and Malone 2002). As a result, when avenues for legal entry were suddenly curtailed after 1965, the migratory flows did not disappear but simply continued without authorization or documents. As shown in the figure, the end of the Bracero Program corresponded exactly in time with the rise of illegal migration. From a figure of around 40,000 in 1965, the number of apprehensions per thousand agents rose to peak at around 460,000 in 1977. It then fluctuated between 330,000 and 460,000 from 1978 to 1986 whereupon it fell into the range of 240,000–320,000 per year after passage of IRCA. This act offered legal status to millions of undocumented migrants who before 1965 had moved back and forth across the border and contributed to the annual count of apprehensions.

In sum, illegal migration rose after 1965 not because there was a sudden surge in Mexican migration, but because the temporary labor program had been terminated and the number of permanent resident visas had been capped, leaving no legal way to accommodate the long-established flows. With permanent resident visas capped, moreover, the inflow of legal immigrants could not rise and remained below 50,000 through the early 1970s and thereafter fluctuated between 50,000 and 100,000 per year. The number of legal immigrants was able to exceed the annual statutory cap of 20,000 because parents, spouses, and minor children of US citizens were exempted from numerical limitation (we return to this point later).

Rise of the Latino threat narrative

In the absence of access to any avenue of legal entry, the post-1965 increase in illegal migration was attributable almost entirely to the termination of the Bracero Program. Once the status quo ante of circular migration had been reestablished under undocumented auspices in the late 1970s, growth in illegal migration ceased and ultimately declined in the wake of IRCA’s legalization. The increase in illegal migration from 1965 through the late 1970s is critically important to understanding the dynamics of policy responses in the years that followed, however, for it was this development that enabled political activists and bureaucratic entrepreneurs to frame Latino immigration as a grave threat to the nation.

Chavez (2001, 2008) has documented the rise of what he calls the “Latino threat” narrative in American media after the 1960s. When he coded national magazine covers on immigration as positive, negative, or neutral he found a steady rise of negative portrayals through the 1970s, 1980s, and 1990s. The rise in the threat narrative occurred during a time of increasing income inequality, and as social psychologist Susan Fiske (2011: 89) has shown, “feeling individually deprived… may alert a person to feeling collectively deprived…[and] this collective feeling leads to blaming out-groups (immigrants, rich elites, the party in power).”

The most common negative framing depicted immigration as a “crisis” for the nation. Initially marine metaphors were used to dramatize the crisis, with Latino immigration being labeled a “rising tide” or a “tidal wave” that was poised to “inundate” the United States and “drown” its culture while “flooding” American society with unwanted foreigners (Santa Ana 2002). Over time, marine metaphors increasingly gave way to martial imagery, with illegal immigration being depicted as an “invasion” in which “outgunned” Border Patrol agents sought to “hold the line” in a vain attempt to “defend” the border against “attacks” from “alien invaders” who launched “banzai charges” to overwhelm American defenses (Nevins 2001; Chavez 2008).

To document the rising use of such metaphors in the American media, we used the Proquest Historical Newspaper Files to search for instances in which the words “undocumented,” “illegal,” or “unauthorized” were paired with “Mexico” or “Mexican immigrants” and the words “crisis,” “flood,” or “invasion.” We focused our analysis on the country’s four leading papers: the New York Times, Washington Post, Wall Street Journal, and Los Angeles Times. To control for random noise and isolate the underlying trends, we computed three-year moving averages. The results are plotted in Figure 2. As can be seen, the use of the negative metaphors to describe Mexican immigration was virtually nonexistent in 1965, at least in major newspapers, but thereafter rose steadily, slowly at first and then rapidly during the 1970s to reach a peak in the late 1970s, roughly at the same time illegal migration itself peaked. From 1965 through 1977 the correlation between the illegal migration series shown in Figure 1 and the negative metaphor series shown in Figure 2 is 0.911.

FIGURE 2. Frequency of pairing of the terms “flood,” “crisis,” or “invasion” with “Mexico“ or “Mexican immigrants,” in four leading US newspapers (three-year moving average), 1965–1995.

SOURCE: Proquest Historical Newspaper Files.

The framing of immigration as a “crisis” and the increasing use of martial imagery were actively promulgated by immigration officials. In 1976, for example, the Commissioner of the Immigration and Naturalization Service (INS) published an article in Reader’s Digest entitled “Illegal Aliens: Time to Call a Halt!” in which he told readers that “when I became commissioner [of the INS] in 1973, we were out-manned, under-budgeted, and confronted by a growing, silent invasion of illegal aliens. Despite our best efforts, the problem—critical then—now threatens to become a national disaster” (Chapman 1976: 188). Similarly, in 1992 the Chief of the San Diego Sector of the Border Patrol, Gustavo de la Viña, filmed and released a video entitled “Border Under Siege,” which in one vivid scene showed migrants scrambling over cars and dodging traffic on Interstate 5 to enter the United States without inspection in and around the San Ysidro Port of Entry, then the country’s busiest border crossing (Rotella 1998).

Politicians quickly discovered the political advantages to be gained by demonizing Latino immigrants and illegal migration. Ronald Reagan, for example, asserted that illegal immigration was a question of “national security,” and in a 1986 speech he told Americans that “terrorists and subversives are just two days’ driving time from [the border crossing at] Harlingen, Texas” (Kamen 1990). In his 1992 reelection campaign, California Governor Pete Wilson called on Congress to “stop the invasion” and borrowed footage from “Border Under Siege” for a series of attack ads. As images of migrants dashing through traffic rolled, a narrator intoned, “they keep coming. Two million illegal immigrants in California. The federal government won’t stop them at the border yet requires us to pay billions to take care of them” (Massey, Durand, and Malone 2002: 89).

The media discovered that the trope of a border under siege made for dramatic copy and good visuals and happily played handmaiden to aspiring politicians and bureaucrats. Later, a host of pundits joined the anti-immigrant chorus to attract attention and sell books. Lou Dobbs (2006) framed the “invasion of illegal aliens” as part of a broader “war on the middle class.” Patrick Buchanan (2006) charged it was part of an “Aztlan Plot” hatched by Mexicans to recapture lands lost in 1848, stating that “if we do not get control of our borders and stop this greatest invasion in history, I see the dissolution of the U.S. and the loss of the American southwest” (Time, 28 August, p. 6). From his lofty Harvard position, Samuel Huntington (2004) warned Americans that “the persistent inflow of Hispanic immigrants threatens to divide the United States into two peoples, two cultures, and two languages… The United States ignores this challenge at its peril.” All these views received extensive coverage in print and broadcast media throughout the country.

Effects on public opinion

The shift in the legal auspices of Mexican migration thus transformed what had been a largely invisible circulation of innocuous workers into a yearly and highly visible violation of American sovereignty by hostile aliens who were increasingly framed as invaders and criminals. The relentless propagandizing that accompanied the shift had a pervasive effect on public opinion, turning it decidedly more conservative on issues of immigration even as it was turning more conservative with respect to social issues more generally. Indeed, the rise of illegal migration remains inadequately acknowledged as a factor in the rightward shift of American public opinion.

To support this argument, we extracted data from the General Social Survey (GSS) from its inception in 1972 to the present and estimated the effect that the annual number of border apprehensions had on the likelihood that a respondent self-identified as conservative (Massey and Pren 2012). We controlled for individual demographic, social, and economic characteristics as well as the country’s overall economic climate. Holding these factors constant, we found that border apprehensions were strongly associated with the likelihood of self-identifying as conservative. An increase in the number of apprehensions from its minimum to its maximum doubled the likelihood that a respondent self-identified as conservative.

Admittedly, self-identification as conservative does not necessarily predict anti-immigrant sentiment; but annual data on attitudes toward immigration are not available. To establish this link, we drew on GSS data for 1996 and 2004, two years in which the survey questionnaires had asked about attitudes toward immigration, enabling us to create a reliable scale of support for exclusionist policies (ibid.). We regressed this measure on respondents’ self-identification as slightly conservative or as extremely conservative (compared to not conservative at all), controlling for age, education, income, occupation, race, region, city size, and national economic climate. We found that conservative self-identification strongly predicted support for exclusionist policies. Both effects were highly significant. Other things equal, conservative self-identification and exclusionist sentiments are strongly interrelated.

Policy feedbacks

Not surprisingly, the rise of the Latino threat narrative and the concomitant increase in conservatism were associated over time with the passage of increasingly restrictionist immigration legislation and the implementation of ever more stringent enforcement policies. Table 1 presents a cumulative list of 15 restrictive immigration bills passed from 1965 to 2010. As can be seen, over time restrictionist bills were passed at an increasingly rapid pace. In the 30 years from 1965 to 1995, for example, only six major immigration bills were enacted, whereas in the decade from 1996 to 2006, eight pieces of legislation were signed into law. Table 2 lists 16 named enforcement operations launched between 1993 and 2010. They typically were announced with great fanfare, including official releases, press conferences, and saturated media coverage. The pace at which such operations were launched also increased over time. Moreover, they became more sweeping in scope, covering locations within the United States as well as along the Mexico–US border.

TABLE 1.

Restrictive immigration legislation enacted by Congress affecting Latin Americans, 1965–2010

1965	Hart–Cellar Act Imposed first-ever annual cap of 120,000 visas for immigrants from Western Hemisphere
1976	Amendments to Immigration and Nationality Act Put Western Hemisphere under preference system and country quotas
1978	Amendments to Immigration and Nationality Act Combined separate hemispheric caps into single worldwide ceiling of 290,000
1980	Refugee Act Abolished refugee preference and reduced worldwide ceiling to 270,000
1986	Immigration Reform and Control Act Criminalized undocumented hiring and authorized expansion of Border Patrol
1990	Amendments to the Immigration and Nationality Act Sought to cap visas going to spouse and children of resident aliens
1996	Anti-Terrorism and Effective Death Penalty Act Authorized expedited removal of noncitizens and deportation of aggravated felons
1996	Illegal Immigration Reform and Immigrant Responsibility Act Increased resources for border enforcement, narrowed criteria for asylum, and increased income threshold required to sponsor immigrants
1996	Personal Responsibility and Work Opportunity Act Declared documented and undocumented migrants ineligible for certain entitlements
1997	Nicaraguan and Central American Relief Act Allowed registered asylum seekers from Central America (mostly Nicaraguans) in the US for at least 5 years since December 1, 1995 to obtain legal status; but prohibited legalization and ordered deportation for those who lacked a valid visa or who previously violated US immigration laws (mostly Guatemalans, Hondurans, and Salvadorans)
2001	USA PATRIOT Act Created Department of Homeland Security, increased funding for surveillance and deportation of foreigners, and authorized deportation of noncitizens without due process
2004	National Intelligence Reform and Terrorism Protection Act Funded new equipment, aircraft, Border Patrol agents, immigration investigators, and detention centers for border enforcement
2005	Real ID Act Sharply increased the data requirements, documentation, and verification procedures for state issuance of drivers licenses
2006	Secure Fence Act Authorized construction of additional fencing, vehicle barriers, checkpoints, lighting and funding for new cameras, satellites, and unmanned drones for border enforcement
2010	border Security Act Funded hiring 3,000 more Border Patrol agents and increased BP budget by $244 million

TABLE 2.

Restrictive enforcement operations launched by the Immigration and Naturalization Service or the Department of Homeland Security 1993–2010

1993	Operation blockade Border Patrol’s (BP) militarization of the El Paso Sector
1994	Operation Gatekeeper BP’s militarization of the San Diego Sector
1998	Operation Rio Grande BP program to restrict the movement of migrants across the Texas and New Mexico border with Mexico
1999	Operation Safeguard BP’s militarization of the Tucson Sector
2003	Operation Endgame Plan launched by Immigration and Customs Enforcement (ICE) to detain and deport all removable noncitizens and “suspected terrorists” living in the United States
2004	Operation Frontline Program launched by ICE to address “vulnerabilities in immigration and trade” by focusing on immigration violators who pose an “enhanced public safety or national security threat”
2004	Arizona border Control Initiative Multi-agency effort supporting Homeland Security’s anti-terrorism mission through the detection, arrest, and deterrence of all persons engaged in cross-border illicit activity
2004	Operation Stonegarden Federal grant program administered through the State Homeland Security Grant Program to provide funding to state and local agencies to improve immigration enforcement
2005	Secure borders Initiative Comprehensive multi-year plan launched by IC E to secure America’s borders and reduce illegal migration
2005	Operation Streamline Program mandating criminal charges for illegal migrants, including first-time offenders
2006	Operation Return to Sender Sweep of illegal immigrants by ICE to detain those deemed most dangerous, including convicted felons, gang members, and repeat illegal immigrants
2006	Operation Jump Start Program authorizing the deployment of National Guard troops along the US–Mexico border
2007	Secure Communities Program ICE program to identify and deport criminal noncitizens arrested by state and local authorities
2007	Operation Rapid REPAT Program to Remove Eligible Parolees Accepted for Transfer by allowing selected criminal noncitizens incarcerated in US prisons and jails to accept early release in exchange for voluntary deportation
2008	Operation Scheduled Departure ICE operation to facilitate the voluntary deportation of 457,000 eligible illegal migrants from selected cities
2010	Operation Copper Cactus Deployment of Arizona National Guard troops to assist BP in apprehension of illegal migrants

This sustained, accelerating accumulation of anti-immigrant legislation and enforcement operations produced a massive increase in border apprehensions after the late 1970s, when the underlying flow of migrants had actually leveled off. For any given number of undocumented entry attempts, more restrictive legislation and more stringent enforcement operations generate more apprehensions, which politicians and bureaucrats can then use to in-flame public opinion, which leads to more conservatism and voter demands for even stricter laws and more enforcement operations, which generates more apprehensions, thus bringing the process full circle. In short, the rise of illegal migration, its framing as a threat to the nation, and the resulting conservative reaction set off a self-feeding chain reaction of enforcement that generated more apprehensions even though the flow of undocumented migrants had stabilized in the late 1970s and actually dropped during the late 1980s and early 1990s.

The dimensions of the paradox are illustrated in Figure 3, which contrasts estimated illegal entries (from Figure 1) with total apprehensions (which reflect both the enforcement effort and underlying traffic). Controlling for the enforcement effort, we see that illegal migration rose from 1965 to 1977, as the circulation of the Bracero Era was reestablished, but thereafter leveled off and fluctuated before ultimately falling. In contrast, the total number of apprehensions grew at a faster pace after 1977, peaking at 1.7 million in 1986 before declining to around 900,000 and then rising again to 1.3 million by 1995. As one would expect, temporal variation in the total number of apprehensions is closely related to the use of threatening immigration metaphors. From 1965 through 1995, the temporal correlation between the frequency of newspaper allusions to crises, floods, and invasions and the total number of apprehensions is 0.956.

FIGURE 3. Annual number of apprehensions and estimated illegal entries, 1955–1995.

SOURCE: US Department of Homeland Security (2012). See text and Table A1.

After the late 1970s, in other words, anti-immigrant sentiment increasingly fed off itself to drive the bureaucratic machinery of enforcement forward to new heights, despite the lack of any real increase in illegal migration. In Figure 4 we model these hypothesized causal paths using a two-stage least squares estimation strategy on time-series data for the 30-year period from 1965 to 1995 (see Table A1). Data on apprehensions, the number of Border Patrol agents, the size of the Border Patrol budget, and the number of line-watch hours spent patrolling the border were obtained from the Office of Immigration Statistics of the US Department of Homeland Security. Proportions conservative from 1972 onward come from the General Social Survey. From 1965 through 1971, we used various national surveys conducted by Gallup, Newsweek, and other polling organizations to provide figures for individual years before the launching of the GSS. The cumulative numbers of pieces of restrictive legislation passed and restrictive operations launched were coded from Tables 1 and 2.

FIGURE 4.

Feedback loop between apprehensions and border enforcement, 1965–1995

The exogenous variable, number of undocumented entries, was independently estimated using life-history data obtained from household heads interviewed by the Mexican Migration Project (MMP), a binational study that each year surveys Mexicans on both sides of the border and draws on the information they provide to create a cumulative database on patterns and processes of documented and undocumented migration («http://mmp.opr.princeton.edu/»). At this writing, the database contains life histories of 21,475 household heads. Each history includes a complete history of migration that can be used to compute annual probabilities of illegal migration (see Massey and Singer 1995; Massey, Durand, and Pren 2009). We applied these estimated probabilities to annual population counts derived from Mexico’s National Institute for Statistics and Geographic Information (the Mexican census bureau) to estimate the gross annual inflow of illegal migrants.

Because this number is exogenous to the paths that define the policy feedback loop, it offers a means of estimating that feedback. The hypothesized causal chain begins with rising undocumented entries, which generate more apprehensions, which are then transformed into a conservative anti-immigrant reaction through the activities of entrepreneurial politicians, ambitious bureaucrats, and elements of the media, and this reaction, in turn, generates more restrictive immigration laws and border operations, which increase the number of Border Patrol agents and the size of the Border Patrol budget, which ultimately produce more linewatch hours, which generates more apprehensions. The effect of the feedback—what might be termed the “enforcement loop”—is to increase the conservative reaction independently of the actual number of illegal entries.

Figure 4 also shows estimated path coefficients for each relationship in the model, allowing judgments of the relative strength of the various pathways. The total effect of any pathway is computed as the product of coefficients along the pathway. Thus the effect of illegal entries on the percent conservative is 0.905 * 0.937, or 0.848. To put it in words, as illegal entries rise so do apprehensions, thus transforming clandestine crossings into threatening events, which prompts more Americans to identify as conservative. This effect is quite strong, as one would expect; but an increasing percent conservative also sets in motion a number of feedback effects that are quite powerful. The total indirect effect through enforcement feedbacks is 0.692 (0.820 * 0.873 * 0.935 * 0.936 + 0.820 * 0.377 * 0.029 * 0.936 + 0.336 * 0.402 * 0.029 * 0.936 + 0.336 * 0.181 * 0.935 * 0.936). Of the two principal pathways in the enforcement loop—the upper one through restrictive legislation (0.820 * 0.873 * 0.935 * 0.936 = 0.626) and the lower one through restrictive border operations (0.336 * 0.402 * 0.029 * 0.936 = 0.004)—the former is obviously by far the stronger, accounting for 90 percent of the total feedback effect (0.626/0.691=0.906). Although not proving causality, this model nonetheless suggests that between 1965 and 1995 rising apprehensions produced a conservative reaction that led to strengthened enforcement and hence more apprehensions, further exacerbating the conservative reaction.

The end result of this feedback loop is depicted in Figure 5, which shows three indicators of the intensity of border enforcement relative to the number of standardized apprehensions, our indicator of the underlying traffic in illegal migrants. Each series is divided by its value in 1977, the year in which undocumented migration reached its peak. The relative number of apprehensions per thousand agents rose from the late 1950s to 1977 and thereafter fluctuated around 1.0 before dropping in the late 1980s. In contrast, from 1977 to 1995 the number of linewatch hours doubled, the number of Border Patrol agents increased 2.5 times, and the Border Patrol budget rose by a factor of 6.5. During and after the 1970s, in other words, the border build-up was increasingly disconnected from the actual traffic in illegal migrants.

FIGURE 5. Intensity of border enforcement, 1955–1995.

SOURCE: US Department of Homeland Security (2012). See text and Table A1.

Enter the war on terrorism

The feedback loop connecting apprehensions, public attitudes, legislation, and enforcement was fully established by the 1990s when a series of exogenous terrorist events drove the enforcement cycle to new heights. Major attacks unleashed in the United States and against US installations abroad in the 1990s and early 2000s helped transform the first decade of the twenty-first century into another era of fear and insecurity. Although anti-immigrant sentiments based on material and ideological considerations were in ample evidence before the 1990s, a string of visible and politically charged terrorist attacks that began in 1993 brought such concerns to new levels. In response to the 1993 attack on the World Trade Center and the 1995 bombing of the Murrah Federal Building in Oklahoma City, Congress in 1996 passed the Anti-Terrorism and Effective Death Penalty Act as well as the Illegal Immigration Reform and Immigrant Responsibility Act. Then upon the heels of the 1998 bombing of the USS Cole in Yemen, the 2000 bombings of the US Embassies in Kenya and Tanzania, and the catastrophic attacks on the World Trade Center and the Pentagon in 2001, Congress enacted the USA PATRIOT Act.

These measures not only further strengthened border enforcement, which had been rising for some time, but more dramatically increased the number of arrests, detentions, and deportations within the United States. Before 1996, internal enforcement activities had not played a very significant role in immigration enforcement; afterward these activities rose to levels not seen since the deportation campaigns of the Great Depression (Hoffman 1974; Balderrama 1995). The conflation of the war on terrorism with the deportation of immigrants is suggested by Figure 6, which plots deportations by year from 1955 through 2009 (US Department of Homeland Security 2012). Prior to the mid-1990s the annual number of deportations had not exceeded 50,000 for decades, but with the passage of the 1996 legislation this threshold was breached and by the turn of the century deportations were running at just under 200,000 annually. With the passage of the Patriot Act in late 2001, the number of deportations increased again and reached nearly 400,000 in 2009. None of the terrorist attacks involved Mexicans, and none of the terrorists entered through Mexico. Indeed, all came to the United States on legal visas. Yet, as Figure 6 clearly indicates, Mexicans nonetheless bore the brunt of the deportation campaign launched in the name of the war on terrorism, comprising 72 percent of those removed in 2009.

FIGURE 6. Annual deportations from the United States, 1955–2009.

SOURCE: US Department of Homeland Security (2012). See text and Table A1.

The likely dynamic by which this outcome unfolded is summarized by the path diagram shown in Figure 7, which uses time-series data from 1965 through 2009 to estimate the effects of the two principal Islamic terrorist attacks on the number of deportations and through them on the conservative reaction, in addition to the ongoing effect through border apprehensions. (See Table A1.) The 1993 terrorist attack indicator was given a value of 1 from 1993 onward and zero otherwise, whereas the 2001 terrorist attack indicator was coded 1 during 2001–09 and zero at earlier dates.

FIGURE 7.

Feedback loop between deportations and internal enforcement, 1965–2009

In this model, the effect of apprehensions continues, but its influence on the conservative reaction is now much more direct (0.806) than indirect through the feedback loop, which is essentially zero ([0.806 * 0.304 * −0.187 * 0.886 * 0.269] + [0.806 * 0.006 * 1.143 * 0.886 * 0.269] = −0.010). In addition, the rising number of apprehensions is no longer significantly connected to the rise in deportations (0.806 * 0.304 * −0.187 * 0.886 + 0.806 * 0.006 * 1.143 * 0.886 = −0.036). The upsurge in deportations arises mainly from the two principal terrorist attacks and the resulting increase in anti-immigrant legislation and enforcement operations. In general, the 1993 attack was more powerful in generating deportations (0.513 * −0.187 * 0.886 + 0.317 * 1.143 * 0.886 = 0.236) than the 2001 attack (0.323 * −0.187 * 0.866 + 0.674 * 1.143 * 0.866 = 0.615), with most of the effect occurring through the rise in enforcement operations in the wake of the Anti-Terrorism and Effective Death Penalty Act. The feedback effects of the terrorist attacks in provoking a conservative reaction through deportations are also relatively modest for both the 1993 attack (0.513 * −0.187 * 0.886 * 0.269 + 0.317 * 1.143 * 0.886 * 0.269 = 0.063) and the 2001 attack ([0.323 * −0.187 * 0.866 * 0.269] + [0.674 * 1.143 * 0.866 * 0.26] = 0.159). Though fairly large by usual social science standards, these effect sizes pale in comparison to the feedback effect of 0.692 observed in the earlier model. Most of the increase in deportations occurred as a response to the two rounds of terrorist attacks.

Explaining the Latino immigration boom

It is evident that the massive increase in border enforcement (Figure 5) and the exponential increase in deportations (Figure 6) were not successful in preventing the entry of millions of unauthorized Mexicans after 1965 (Redburn, Reuter, and Majmundar 2011: 34). From 1980, when Warren and Passel (1987) first estimated the number of undocumented Mexicans living in the United States to be 1.13 million, the population grew to 2.04 million in 1990, reached 4.68 million in 2000, and then peaked at 7.03 million in 2008 (Wasem 2011). Most of the remaining growth in the undocumented population came from Central America. When the United States intervened in the region by launching the Contra War and funding paramilitary groups, the level of violence increased; and as the region’s economy deteriorated, the resulting wave of emigrants met with the same restrictions on legal entry that had earlier blocked the entry of former Braceros from Mexico. The sole exception was Nicaragua, whose emigrants were fleeing a leftist regime and were thus allowed to overstay tourist visas and ultimately adjust their status to become legal permanent residents (Lundquist and Massey 2005). In contrast, the flows of undocumented Salvadorans, Guatemalans, and Hondurans remained predominantly illegal, accumulating total undocumented populations of 570,000, 430,000, and 300,000, respectively, by 2008 (Hoefer, Rytina, and Baker 2009). Central Americans and Mexicans together accounted for nearly three-quarters of the increase in the undocumented population.

Not only did the massive enforcement effort fail to prevent the entry of unauthorized Latin Americans, in an unanticipated way it actually accelerated the net inflow. Although the inflow of undocumented Mexicans was largely unaffected by the build-up of enforcement resources along the border, the outflow was drastically curtailed (Redburn, Reuter, and Majmundar 2011). As the costs and risks of unauthorized border crossing mounted, migrants minimized them by shifting from a circular to a settled pattern of migration, essentially hunkering down and staying once they had successfully run the gauntlet at the border (Massey, Durand, and Malone 2002). It was thus a sharp decline in the outflow of undocumented migrants, not an increase in the inflow of undocumented migrants, that was responsible for the acceleration of undocumented population growth during the 1990s and early 2000s, and this decline in return migration was to a great extent a product of US enforcement efforts (Redburn, Reuter, and Majmundar 2011).

It was not just undocumented migration that grew after 1965. As we have seen, legal immigration from Latin America also grew despite the imposition of caps and quotas. With a country quota of 20,000 visas per year beginning in 1976, the expected number of entries for any country over a decade is 200,000, yet the decadal number of legal Mexican immigrants, which amounted to 442,000 in the 1960s, rose to 621,000 in the 1970s, reached 1 million in the 1980s, and peaked at 2.8 million in the 1990s before dropping back to 1.7 million in the first decade of the new century. Over the same period, legal immigration from the rest of Latin America rose from 544,000 in the 1960s to 734,000 in the 1970s, 935,000 in the 1980s, 1.5 million in the 1990s, and 1.6 million in the 2000s (US Department of Homeland Security 2012).

The key to understanding the dynamic of this growth again lies with decisions taken by Congress that had unintended consequences: the decision in 1965 to exempt close relatives of US citizens from the country quotas and a series of decisions from 1986 onward that systematically privileged US citizens, limited the rights and liberties of legally resident noncitizens (“green card” holders), and increased the vulnerability of noncitizens to deportation. The quota exemptions by themselves probably would not have led to a sharp increase in legal immigration from Latin America, but when combined with the rising burden placed on noncitizens by US immigration legislation, they produced a dynamic interplay between naturalization and family reunification that drove legal immigration to new heights.

When the citizenship exemption was enacted in 1965, Congress was not really thinking about naturalized citizens. In that year the typical naturalized citizen was an elderly European who had arrived before 1929 and whose children had been born in the United States. Instead, the prototypical citizen seeking to sponsor an immigrant spouse and children in 1965 was a US serviceman returning from duty in Southeast Asia, or coming home from a deployment in Germany, South Korea, the Philippines, Panama, or other countries containing large US military bases. In the context of the Vietnam War, Congress certainly did not want to block entry by wives and children of American soldiers. In addition, the Catholic Church, immigrant organizations, and humanitarian groups rallied around the principle of “family reunification” as the moral cornerstone of US policy and succeeded in including the parents of US citizens under the quota exemption.

For many years, the quota exemption for citizen relatives had little effect on levels and patterns of immigration. For Latin Americans, there was no special advantage to assuming US citizenship once legal permanent residence had been achieved. The entry of spouses and minor children was numerically limited, but they were given a high position in the “preference system” of visa allocation; and with the exception of eligibility for federal employment and the right to vote, there was no particular advantage to naturalization. This calculus began to change in 1986, and in subsequent years the benefits of citizenship increased as the costs of noncitizenship proliferated.

When Congress passed the Immigration Reform and Control Act in 1986, it authorized two legalization programs, one for farm workers and another for long-time residents, and the terms of these programs for the first time made prerequisites to citizenship a requirement for legalization. Prior to adjusting status to a legal one, immigrants were required to provide evidence of facility or instruction in English and a knowledge or evidence of instruction in American history. In this way, 2.7 million former undocumented migrants were induced to satisfy the requirements previously asked only of people wishing to become US citizens (Massey, Durand, and Malone 2002). Five years after regularization, millions of people would suddenly become eligible for naturalization. Having satisfied the language and civics requirements, all they would have to do is pay an application fee and wait.

The first cohort of persons legalized under IRCA became eligible for citizenship in 1992, with the bulk following in 1993 and 1994. In 1990, Congress undertook the first of a series of actions that began to ratchet up the costs of noncitizenship. In an effort to slow immigration from Latin America and other major immigrant-sending countries, amendments passed in that year sought to limit visas issued for purposes of family reunification by setting an annual cap, but one that could be “pierced” by taking family reunification visas that formerly went to relatives of legal residents and giving them to relatives of citizens. The net effect was to permanently reduce access to legal visas by relatives of legal resident aliens, which greatly increased their waiting time (Zolberg 2006).

Congress followed up in 1996 with three pieces of legislation, all of which bore down heavily on legal immigrants. The Illegal Immigration Reform and Immigrant Responsibility Act authorized the deportation of noncitizens from ports of entry without judicial hearing and, in an effort to restrict family migration still further, required sponsors of legal immigrants to provide affidavits of support that demonstrated a household income at least 125 percent of the federal poverty line. Not only would resident noncitizens have to wait in a long line to sponsor the entry of a spouse or child, when their turn finally came, they would need more money as well. Meanwhile, the Personal Responsibility and Work Opportunity Reconciliation Act placed new restrictions on legal permanent residents’ access to public services, barring them from receiving food stamps, Supplemental Security Income, and other means-tested benefits for five years after admission (Newton 2008).

The legislative troika was completed with passage of the Anti-Terrorism and Effective Death Penalty Act, which declared that any noncitizen who had ever committed a crime, no matter how long ago, was subject to immediate deportation. It also gave the federal government broad new powers for the “expedited exclusion” of any noncitizen who had ever crossed the border without documents, no matter what his or her current legal status. In addition, the bill granted the executive branch new authority to designate any organization as “terrorist,” thereby making all members immediately excludable. At the same time, the bill severely circumscribed the possibilities for judicial review of deportation orders (Legomsky 2000).

Finally, in response to the terrorist attacks of September 11, on October 26, 2001 Congress passed, without significant debate, the USA PATRIOT Act, which granted executive authorities even greater powers to deport, without hearings or any presentation of evidence, all noncitizens—legal or illegal, temporary or permanent—who the Attorney General had “reason to believe” might commit, further, or facilitate acts of terrorism. For the first time since the Alien and Sedition Act of 1798, Congress authorized the arrest, imprisonment, and deportation of noncitizens upon the orders of the Attorney General without judicial review (Zolberg 2006). These repressive federal laws have been accompanied in recent years by a surge in anti-immigrant measures enacted at the state and local level (Hopkins 2010).

The net effect of these legislative acts was to dramatically increase the pressure on noncitizens, regardless of legal status. In response to the rising pressure, more immigrants adopted a strategy of “defensive naturalization” in order to protect themselves and their families from detention and deportation and to guarantee their continued access to public benefits. The trend is most clearly seen among Mexicans, who historically exhibited one of the lowest rates of naturalization of any major immigrant group. From 1965 through 1985, the number of Mexicans who naturalized averaged just 8,200 per year. During the immediate post-IRCA period (1986–95), it grew to 29,400 and in the period just after the 1996 Acts (1996–2001) it mushroomed to 168,500, before finally falling back to 99,000 per year in the years after the 2001 terrorist attack (2002–10). If we take the average number of naturalizations per year from 1965 to 1985 as the pre-IRCA norm, then an additional 2 million Mexicans naturalized after 1986 compared to what would have been the case under the previous circumstances.

As the number of newly minted citizens grew, so did the number of Mexicans admitted to the United States outside the quotas as relatives of US citizens, rising from an annual average of 24,000 before 1986 to 34,000 during 1986–95 and 80,000 during 1996–01 before peaking at 98,000 during 2002–10 (US Department of Homeland Security 2012). Whereas a legal permanent resident may petition for the entry of a spouse and minor children, such dependents must wait in an increasingly long line for an increasingly scarce allocation of visas. If the same person naturalizes, however, his or her spouse, children, and parents can enter immediately without limitation. Each new citizen thus generates more potential immigrants over time. By pushing immigrants toward defensive naturalization, Congress in effect increased future legal immigration.

As a result of this dynamic, the proportion of Mexicans admitted to the United States as relatives of US citizens has steadily increased since 1991, when the costs of noncitizenship began seriously to accumulate. As shown in Figure 8, the percentage admitted outside the country quotas rose from just 5 percent in 1990 to peak at 69 percent in 2002 before falling to 45 percent in 2005 and then rebounding to 64 percent in 2010. The more young naturalized citizens there are in the population, the larger the share of entries by citizen relatives. Indeed, from 1990 to 2010 the correlation between the cumulative total of naturalizations and the percentage entering as citizen relatives is 0.854. In other words, the larger the number of naturalized citizens in the population, the higher the percentage of immigrants able to enter outside the national quota.

FIGURE 8. Percentage of Mexicans admitted outside the country quota as relatives of US citizens, 1990–2010.

SOURCE: US Department of Homeland Security (2012).

The new American demography

We have argued that the post-1965 surge in Mexican, Central American, and to a lesser extent South American immigration was not a direct result of immigration reforms that occurred in the mid-1960s, but instead arose indirectly through an accumulation of unintended consequences that unfolded afterward. Before 1965 a circular flow of migrants between Mexico and the United States had been established under the aegis of the Bracero Program and institutionalized informally through the spread and elaboration of migrant networks, which connected sending communities in Mexico to work sites in the United States. By the late 1950s a circulatory cross-border migration system was entrenched and the annual inflow averaged roughly 500,000 persons per year, with around 90 percent entering on temporary work visas.

A host of immigration reforms enacted beginning in the mid-1960s eliminated the temporary work program while simultaneously imposing new quantitative limits on immigration from the Western Hemisphere. With opportunities for legal entry constrained, the well-established migratory flow simply continued informally without authorization. Illegal migration rose steadily after 1965 to peak in the late 1970s. Thereafter it fluctuated in tandem with economic trends north and south of the border. In essence, in 1965 the United States shifted from a de jure guestworker program based on the circulation of bracero migrants to a de facto program based on the circulation of undocumented migrants.

The substitution of a de facto for a de jure system of labor migration might have been the end of it, if not for the fact that the associated rise in apprehensions offered aspiring politicians an opportunity to mobilize voters and entrepreneurial bureaucrats a chance to obtain status and resources by framing illegal migration as a grave threat to the United States. These actors portrayed the undocumented inflow as a crisis, framing it as a “tidal wave” threatening to “flood” the country and “drown” its society, or as an “alien invasion” that threatened national security, defined with reference first to the Cold War and then to the war on terrorism. The media disseminated the “Latino threat” narrative and depicted a “border under siege” and increasingly deployed threatening marine and martial imagery in newspapers, magazines, and on the air.

The rising number of border apprehensions and the intensifying threat narrative, in turn, had profound political consequences, galvanizing a shift toward conservatism among voters and increasing support for more stringent immigration and enforcement policies, setting off a chain reaction in the public sphere. Rising apprehensions led to greater conservatism, which produced more restrictionist legislation and enforcement policies, which translated into a larger Border Patrol with bigger budgets, which produced more linewatch hours, which boosted apprehensions, which then fed back on conservatism. After 1979 rising apprehensions were driven increasingly by this feedback effect, bringing about a growing divergence between apprehensions and the actual volume of undocumented migration.

The massive increase in border enforcement that arose from this feedback loop had the unintended and unexpected result not of deterring departures from Mexico but of reducing returns (Redburn, Reuter, and Majmundar 2011). The volume of US out-migration plummeted as the volume of in-migration continued, and the net rate of undocumented population growth rose sharply to bring about a large increase in the size of this population during the 1990s and early 2000s. On top of this growth in undocumented migration from Mexico, US intervention in Central America during the Cold War contributed to a further destabilization of the region, leading to large-scale migration northward. Although Nicaraguan emigrés were for the most part welcomed as refugees from communist tyranny, those from Salvador, Guatemala, and Honduras encountered the same restricted opportunities for legal entry as Mexicans. As a result, most entered without authorization or grudgingly received temporary protected status only to have it revoked after the Cold War ended. As a result, nearly three-quarters of the roughly 11 million undocumented migrants in the United States today are from Mexico or Central America.

In the 1990s the Cold War was replaced by the threat of terrorism. The Anti-Terrorism and Effective Death Penalty Act of 1996 and the 2001 Patriot Act intensified border enforcement and, more importantly, brought about a sharp rise in deportations from the United States. Deportations replaced border apprehensions as the visible manifestation of the Latino threat. Although the resulting feedback loop was not as powerful as the apprehensions-based loop that prevailed from 1965 to 1995, it was potent nonetheless and deportations expanded even as apprehensions fell in the decade after 2000.

Led by Mexico, legal immigration from Latin America also rose steadily over the period, especially after 1986 when it came to be powered by a dynamic interplay between rising naturalization rates and the increasing use of quota exemptions by recently naturalized immigrants. Spouses, minor children, and parents of US citizens had long been exempted from the country quotas, but, in response to the continuing increase in legal immigration, rising concerns about terrorism, and growing xenophobia, Congress began to strip civil, social, and economic rights from legal as well as illegal immigrants, prompting more permanent residents to petition for citizenship as part of a strategy of “defensive naturalization.” This dynamic unfolded just as millions of former undocumented migrants became eligible to naturalize after receiving permanent residence under IRCA, and each new citizen simply created more legal migration over time. As the number of naturalized migrants grew into the millions, a rising share of immigrants from Mexico and other countries were admitted as relatives of citizens, thus evading the country quotas.

The end result of these processes has been a massive transformation of the demography of the United States in the past 40 years. In 1970 the Hispanic population of the United States stood at around 9.6 million and comprised just 4.7 percent of the US population. More than 70 percent were native born, 60 percent were Mexican, and just 6 percent were Central or South American, compared with nearly a quarter from the Caribbean. In 2010, after four decades of mass migration, the number of Hispanics had risen to 50.5 million and constituted more than 16 percent of the US population. The percentage US-born among Hispanics had fallen to a little over 60 percent and the distribution by national origin had shifted, with Caribbeans dropping to around 15 percent while Mexicans rose to 63 percent and Central and South Americans to 13 percent, with the remaining 9 percent falling into the “other” category (Ennis, Ríos-Vargas, and Albert 2011).

Over the past four decades, the legal status distribution was likewise markedly transformed. According to estimates from the Census Bureau (Acosta and de la Cruz 2011) and the Department of Homeland Security (Hoefer, Rytina, and Baker 2011), a majority of immigrants from Mexico and Central America are present without authorization, including 58 percent of Mexicans, 57 percent of Salvadorans, 71 percent of Guatemalans, and 77 percent of Hondurans. Even when one considers all Hispanic generations together, the undocumented members constitute large fractions of these origin groups: 21 percent of all persons of Mexican origin, 38 percent of Salvadoran origin, 50 percent of Guatemalan origin, and 52 percent of Honduran origin. In other words, undocumented migrants are no longer a small share of Latinos in the United States. Never before have so many people been outside the law and never before have the undocumented been so concentrated within such a small number of national origins.

A counterfactual scenario

To say that US immigration policies have failed is an understatement. From 1970 to 2010 the population born in Latin America increased more than 11 times. Owing to mass immigration, the total Hispanic population grew by a factor of five, and the percentage of the population born in Latin America residing in the US more than tripled. All these trends unfolded in spite of—and, as we have shown, paradoxically also because of—the progressive limitation of opportunities for legal entry, the massive build-up of enforcement resources at the border, the large rise in deportations, and the systematic restriction of the civil liberties and social rights of noncitizens. If the goal of such actions was to limit immigration from Latin America and prevent the demographic transformation of the United States, they achieved the opposite.

Might events have worked out differently? Possibly. The crux of the problem is that Congress routinely makes consequential policy decisions with scant consideration of the underlying dynamics of the social processes involved. That was certainly the case here, for in orchestrating immigration reforms during the 1960s and 1970s Congress took little notice of the long history of recruitment in the hemisphere; the high degree of circularity that historically had prevailed; the strong connection of flows to the dynamics of labor supply and demand; the key role of networks in sustaining and expanding migration over time; the motivations of migrants and how they change in the course of a migratory career; the structural transformations that occur in sending and receiving areas as a result of mass migration; the likelihood of a migratory response to economic, political, and military intervention; the large size and well-established nature of flows into the United States on the eve of restriction; and most importantly the strong momentum that accrues to migratory flows once underway.

As we have demonstrated, intervening forcefully in complex social and economic systems without understanding their dynamics can lead to unintended consequences and unanticipated policy feedbacks. To be fair, immigration was not a salient issue in the 1960s and 1970s when many of the fateful decisions were taken, and fundamental features of migration processes—network effects, migratory momentum, migrant motivations—are much better understood now than they were four decades ago. Nonetheless, it is possible to imagine a different scenario occurring, particularly if Congress had thought about immigration reform as potentially having wide-ranging effects on a well-established binational migratory system rather than simply as a domestic political issue.

Suppose, for example, that in choosing to reform the Bracero Program, Congress had enacted safeguards to improve the wages, working conditions, and treatment of workers instead of shutting down the program entirely. Suppose that in implementing the new system of ethnically neutral country quotas, Congress had granted special, more generous visa allocations to Canada and Mexico as America’s closest neighbors. Finally, suppose that instead of funding the Contras and other paramilitary operations, generating a cycle of violence and economic disruption, the Reagan Administration had let events in Nicaragua take their course without interference.

Under those circumstances mass migration from Latin America might indeed have been avoided. With the continuation of a reformed temporary labor migration program, the flow from Mexico would have remained predominantly circular. With a generous country quota for permanent resident visas, Mexican workers who established legitimate ties and wished to settle in the United States would have had a legal pathway. Hispanic population growth would thus have been slow and illegal migration would likely not have risen to the high levels that have made it into a major political issue. The Latino threat narrative would not have gained traction, fears of an alien flood or invasion would not have pushed Americans toward greater conservatism, and there would have been little support for restrictionist policies.

At the same time, absent a US intervention in Central America, the collapse of Soviet support would likely have brought down the Sandinista government in Nicaragua and ended insurgencies elsewhere without displacing millions of Central Americans northward. Central Americans would not have contributed so greatly to Hispanic population growth. The illegal population of the United States would consequently have remained rather small, Latinos would still be the country’s second minority after African Americans, and, when the terrorists attacked, the United States would have been a less divided and fractious nation. A softer line on restriction, less punitive enforcement, and greater military restraint could have yielded fewer permanent immigrants, less undocumented migration, and slower population growth.

The era of mass Mexican migration appears to be over, at least for the moment. According to the best estimates, the undocumented population peaked at around 12 million in 2008, fell to 11 million in 2009, and has held steady since then (Wasem 2011). The annual number of legal entries has likewise fallen, dropping from 219,000 in 2002 to 139,000 in 2010. The number of entries by temporary workers, in contrast, has increased to record levels, reaching 517,000 in 2010, the largest number in history (US Department of Homeland Security 2012). With the number of temporary worker entries exceeding those recorded during the Bracero Program, illegal migration at zero, and demand for permanent resident visas apparently falling, the key remaining item on the immigration reform agenda is how to deal with the legacy of past failed policies—the large population of undocumented residents. Despite record deportations and rising anti-immigrant sentiment, the rate of return migration among this population is also near zero.

Of the 11 million persons currently present without authorization, some 3 million entered as children. Absent a criminal record, there is a compelling case that they should be granted amnesty as proposed in pending legislation, such as the Dream Act. After all, it was not their decision to be undocumented. For those who entered the United States illegally as adults, arguably the only sensible and humane solution is an earned legalization program. For example, migrants could accumulate points for learning English, taking civics courses, paying taxes, and having US citizen children. When a specified threshold number of points is reached, they would pay a fine and adjust their status to legal permanent resident. Abundant research has documented that harsher enforcement and rising deportation have not increased the rate of “self-deportation,” but instead have lowered the rate of return migration among the undocumented to record lows. A permanent undocumented population of 11 million can only bring a host of social and economic problems, ones that will worsen the longer an appropriate policy response is deferred.