The Handbook of International Migration: The American Experience

4. Matters of State: Theorizing Immigration Policy

4 Matters of State: Theorizing Immigration Policy Aristide R. Zolberg "THERE ARE TIMES WHEN, as I look at the regulations of the countries of the world affecting immigrants , I see in my mind's eye the building up of walled-in countries, much like the wall-encircled towns of the medieval period" (Fields 1938, 3). Conjured up two-thirds of a century ago, this metaphoric representation of the outcome of the industrialized world's first immigration crisis is appropriate for our own times as well. Yet considering the pervasiveness of barriers to immigration, mirrored throughout much of the twentieth century by draconian prohibitions on exit that confined hundreds of millions to their countries for many decades, it is remarkable that the role of states in shaping international migration has been largely ignored by immigration theorists. This is acting like the proverbial ostrich, but ignoring the challenge by burying one's head in the sand will not make it go away. A notable case in point is the recent state-ofthe -art assessment published by the Committee on South-North Migration of the International Union for the Scientific Study of Populations (Massey et al. 1993; Massey, Goldring, and Durand 1994).1 Reviewing a wide range of models purporting to explain why international migration begins and persists, the authors maintain an openminded stance, allowing that some of these models might be complementary rather than mutually exclusive . Although Massey and his collaborators adopt as their baseline a neoclassical model of migration as the aggregate result ofincome-maximizing decisions by individuals, they take into account the contributions of heterogeneous approaches, and ultimately suggest (1993, 433) that "it is quite possible ... that individuals act to maximize income while families minimize risk, and that the context within which both decisions are made is shaped by structural forces operating at the national and international levels." Although the committee acknowledges that the "context" includes state regulations, its treatment ofthis factor is itselfentirely atheoretical. For example , barriers to entry into another country are included within the equation representing the calculus of potential migrants as the probability of avoiding deportation from the area of destination (1.0 for legal migrants, < 1.0 for undocumented migrants). Nothing is said, however, about how the parameters determining the category "legal immigrants " are established, and thus how "illegal" ones are defined as well; how large the "legal" category is (absolutely and in relation to demand) in specific instances and for the world as a whole; how these categories vary over time and between countries; and especially, since we are in the realm of theory, how such variation might be explained. The authors handle regulations governing access to the labor market in the same manner, as determinants of the probability of employment (1993, 442), but here again, they make no attempt to describe or explain variation between receiving countries or over time, and in particular the contribution ofimmigration regulation to this variation. Beyond this, the authors' conceptualization of the individual decisionmaking process does not take into consideration calculations by individuals regarding obstacles to "exit," suggesting an implicit assumption that freedom to leave prevails worldwide . Yet this cannot be taken as a given, since until quite recently a very large proportion of the people who wished to move were not able to do so except at considerable risk. This was the case not only for the many hundreds ofmillions living in Communist countries but also, throughout the first half of the twentieth century, for colonial populations that were denied access to the documents required for foreign travel-usually even to the home territory of the colonial power, or to other parts of the empire . Recent changes in these parameters have dramatically enlarged the pool of potential international migrants and concomitantly the regulatory load that states of potential immigration face, but even today many who would rationally be inclined to migrate are not free to do so. In his contribution to this volume, Massey himself acknowledges that, "for theoreticians, a principal challenge is to model the behavior of nation- 72 The Handbook ofInternational Migration states and political actors, filling a void in the general theory of international migration we have constructed here." Indeed, this challenge has stood for over a century, ever since Ernst Georg Ravenstein inaugurated the theoretical quest with his two papers on "The Laws of Migration" presented to the Royal Statistical Society, which adumbrate the rational calculus underlying the theory set forth by Massey and his associates today (1885, 198-99; 1889, 286).2 Four years later he reported further that the propositions generally held true for internal migrations in the other European countries for which data were available, as well as in the United States and Canada; he also verified them for international migration within Europe.3 However, in response to critics who questioned his assumptions, Ravenstein conceded that "laws of population, and economic laws generally , have not the rigidity of physical laws, as they are continually being interfered [with] by human agency," and he noted in particular that "currents of migration which would flow naturally in a certain direction traced out for them in the main by geographical features, may thus be diverted, or stopped altogether, by legislative enactments» (1889, 241, emphasis added). With respect to exit, he observed that fewer people were leaving Russia than was to be expected in the light of that country's economic conditions because of prohibitive government regulations; with regard to entry, however, his only example pertained to internal movement, notably an enactment in the reign of the first Elizabeth prohibiting new settlement in London-a prohibition, he relates, that in fact was not enforced (1889, 241). The absence of reference to international restrictions on immigration may be attributable to the fact that earlier in the century England had eliminated all barriers to entry-imposed during the Napoleonic wars for security reasons-as contrary to Manchesterian principles, and that Ravenstein limited his empirical investigation to immigration into European countries, a movement that at this time was generally unfettered as well. However , there were ominous signs of impending change (Zolberg 1978a, 1997). In the very year when Ravenstein's first paper was published, in his native Germany Bismarck massively expelled "undesirable " Poles, including both Jews and nonJews , and enacted measures to restrict and regulate further immigration from eastern Europe. Even more startlingly, in the brief interval between the first and second papers, rising concern within England itself over the arrival of poor Jews from eastern Europe prompted the founding of the Society for the Suppression of the Immigration of Destitute Aliens and the establishment by the House of Commons of the Select Committee on Immigration to consider whether new legislation was required to deal with the problem (Gainer 1972, 8, 80). The discussion that followed Ravenstein 's second presentation affords us a suggestive glimpse of the changing climate: one of the fellows expressed concern over the influx of foreigners into England "as one of the causes of the emigration of Englishmen" (Ravenstein 1889, 304). In the wake of this, in 1905, England enacted what was then the industrial world's most restrictive immigration law, prohibiting the landing of poor aliens of any kind, at the will of portof -entry officials. Had Ravenstein addressed himself to transoceanic movements, he would undoubtedly have noted that the United States was also moving away from laissez-faire by way of a spate of national measures, including draconian restrictions on the immigration of Chinese women (1875) and Chinese workers (1882); the first comprehensive federal law regulating general immigration (1882); and a prohibition of all immigration involving labor contracts (1885). Also in 1885 one of the new progressive economists, reckoning that massive immigration harmed American workers, proposed subjecting prospective immigrants to a literacy test so as to reduce the flow of unskilled labor; this was passed by both houses of Congress a decade later and became law in 1917. In retrospect, it can be seen that these initiatives constituted the first stones of a global wall erected by the rich industrial states to protect themselves from "invasion" by the world's poor. Pushed out of rural areas by demographic growth combined with the spreading dynamics of the market and rendered more mobile by a revolution in transportation , the world's poor constituted a much larger and more diverse pool of potential international migrants than ever before. Although construction of the wall was well under way before World War I, the work was then speeded up by security concerns and the onset of the Bolshevik revolution. It is especially noteworthy that most of the wall's national segments were completed before the Great Depression set in, so that the overall phenomenon cannot be attributed simply to deteriorating economic conditions in the receiving countries. In the United States, which then as now consti- tuted one of the largest components of the international migration system, the wall's major elements were put in place in the period from 1896 to 1924 and included the imposition of heavy fees on entering immigrants; a proliferation of prohibited categories based on health, moral, and political criteria; a literacy requirement designed to reduce the flow of unskilled immigrants, especially from eastern and southern Europe; the nearly absolute prohibition of Asian immigration; the reduction of immigration from Europe to about one-fourth its ongoing level; and the allocation of entries on the basis of national origins so as to severely limit arrivals from eastern and southern Europe , with the explicit objective of preserving the character of the United States as a predominantly "Anglo-Saxon," Protestant community (Higham 1994; Hutchinson 1981). Concurrently, the United States elaborated an innovative apparatus of "remote control" that enabled it to deter immigration by regulating embarkation at or near the point of origin (as discussed later in the chapter). In the decades after World War II, the gates were widened to accommodate political refugees in keeping with cold war imperatives; to facilitate the procurement of temporary labor in the course of the "Bretton Woods boom"; and to allow for family reunion in response to constituency pressures (in the United States), newly institutionalized "family rights" (in most of Europe), and related obligations arising from the recently established international human rights regime. As a number of sociologists have pointed out, over time the flows took on a life of their own by way of the development of networks that lowered the risks and costs of outward movement (Massey et al. 1993,448-50). The result was a revival ofimmigration as a significant social phenomenon in most of the traditional "immigration" countries, notably the United States, as well as its appearance de novo in a number of newly affluent countries that had been prominent sources of emigration, such as Italy and Norway. Reflecting these developments, international migration revived as a social science subject, particularly in sociology and economics. However, focusing on the incoming streams, social scientists paid little or no attention to the fact that the streams were flowing through gates, and that these openings were surrounded by high walls. Since a great deal of international movement was in fact taking place, there was little incentive to surmise about why more was not taking place. Matters of State 73 Hence, as the new literature came into being, the "legislative obstacles" to which Ravenstein had alluded did not stimulate the curiosity of social scientists engaged in theory construction. Of course, the ubiquity of state actions purporting to regulate exit and entry does not necessarily mean that such measures play a significant role in shaping international migrations; they might prove largely ineffective, and in that case, theorists might be justified in ignoring them as a "residual error" that does not warrant explanatory efforts. For example , the IUSSP report's senior author, Douglas Massey, has repeatedly expressed considerable skepticism regarding the effectiveness of U.S. regulations in determining the level and composition of U.S. immigration in the twentieth century, both legal and unauthorized (Donato, Durand, and Massey 1992; Massey and Singer 1995; Massey 1995). In the same vein, the authors of a recent comparative study of immigration policy and policy outcomes in nine industrialized democracies suggest that "the gap between the goals of national immigration policy . . . and the actual results of policies in this area ... is wide and growing wider in all major industrialized democracies" (Cornelius, Martin, and Hollifield 1994b, 3). In the face of such widespread skepticism, I am compelled by my contention that the neglect of the role of state policies, broadly speaking, constitutes a major flaw in an important theoretical undertaking , to begin by falsifying what amounts to a "null hypothesis," namely, the contention that state regulations regarding exit and entry do not significantly shape international migration. If that hypothesis can be laid to rest-and I believe that a persuasive case can be made to that effect-then surely international migration theory must encompass a theoretically grounded account of state agency. In the second part of this chapter, I review various attempts to come to grips with the subject and incorporate them into a tentative comprehensive framework. How EFFECTIVE ARE IMMIGRATION CONTROLS? Contrary to many assertions, evidence from historical cases as well as recent developments in immigration policy substantiate Ravenstein's (1889) admonition that "currents of migration which would flow naturally in a certain direction ... may . .. be diverted, or stopped altogether, by legisla- 74 The Handbook ofInternational Migration tive enactments." To begin with, the null hypothesis is easily rejected with regard to emigration, since it is evident that prohibitions on exit can be highly effective. As already noted by Ravenstein a century ago, "Great Russian" emigration was minuscule in relation to the base population and prevailing conditions in the source country: Since there can be no question that today most states that choose to block exit can harness the capacity to do so (Dowty 1987), my discussion will center on the regulation of entry.s In the present climate, assessments of the capacity of states to control immigration also have important public policy implications , because widespread perceptions of a catastrophic "loss of control" may precipitate and legitimize the enactment of draconian measures that conflict with other societal objectives and desirable values. On the immigration side, the null hypothesis is demonstrably false with regard to movement into the United States from noncontiguous countries -movement that today, as in the past, accounts for the largest share of recorded immigration -as well as with regard to immigration to the other English-speaking overseas countries and contemporary immigration to Europe. To what extent regulation is effective with regard to immigration into the United States from Mexico is much less clear, not only because the evidence regarding the effects of the relevant "legislative enactments" is mixed, but especially because American policy objectives are themselves highly ambiguous. Given the focus of this volume, I shall limit the discussion to the United States, focusing on three instances referred to by Massey. u.s. Q1wta Period, 1921 tv 1965 Boldly questioning common wisdom that the draconian restrictions of the post-World War I period were effective, Massey asserts that "although the national origins quotas, combined with earlier bans on Asian immigration enacted in 1882 and 1917, did playa role in reducing the number of immigrants, I believe their influence has been overstated" (Massey 1995, 635). In support of his contention, he points out that the quotas did not apply to immigrants from the Western Hemisphere , and that after World War I, which induced a sharp drop in immigration, "European immigration began to revive, despite the restrictive quotas" (636). I shall first demonstrate that the second point is not supported by the relevant statistics , and I shall deal with Massey's observation regarding the Western Hemisphere in the next section. As evidence that U.S. quotas did not reduce immigration significantly, Massey cites statistical evidence to the effect that during the period from 1921 to 1930 some 412,000 immigrants arrived from Germany, 455,000 from Italy, 227,000 from Poland, and 102,000 from Czechoslovakia, supplementing "large numbers" from countries that were not limited by quotas. However, he fails to examine the distribution of these decennial numbers in the years before and after enactment of the restrictive legislation. When this is done, the evidence clearly contradicts his assertion. The demonstration requires a detailed consideration of the regulations and their consequences. It is certainly true that World War I induced a sharp drop in immigration-mostly because European belligerents did not allow their populations to leave-and that European immigration rapidly revived afterward, when European men were demobilized and civilian shipping resumed. The most important legislative barrier imposed in the intervening period was a literacy requirement (in the immigrant's own language), enacted in 1917, three decades after it was first advocated as a deterrent . Its ineffectiveness at the time under consideration is hardly surprising, since by World War I the pool of literate males was very large, even among the lower classes of the less developed regions of Europe. Recorded immigration rose from a mere 24,627 in 1919 to 246,295 the following year, and 652,364 in 1921, the last year when it was unrestricted by numerical limits (U.S. Bureau of the Census 1971, 56-59). The escalation suggested that arrivals were well on their way toward the one-million level of the immediate prewar years; given European political and economic upheavals , there was reason to expect that this would soon be exceeded-possibly in keeping with the previous experience of a doubling of incoming numbers in each of the four prewar decades. Interpreted in the light of the pervasive nativism, itself exacerbated by wartime nationalism and reactions to the Russian revolution, these developments and expectations prompted decisionmakers to impose an unprecedented quantitative limit on immigration from Europe (Higham 1994). They further allocated the newly rarefied entries by way of nationality quotas designed to reduce immigration from southern and eastern Europe, an approach first proposed in a congressional report shortly before the war. Applicable to the independent countries of the "Eastern Hemisphere" (excluding China, subject to near-absolute exclusion), the 1921 law limited the yearly number of immigrants of each nationality to 3 percent of the number of foreign-born of that nationality enumerated in the United States in 1910, making for a maximum of 356,000 (Hutchinson 1981, 483-87). As a result, European immigration immediately dipped to 216,385, but it then rose again to 307,920 in 1923, and to 364,339 in 1924 (as recent immigrants brought in their spouses and children). The Congress then devised a more draconian law, signed in 1924, which provided for a maximum of 150,000 a year, to be allocated in proportion to the "national origins " of the American population as of 1920. The measure was designed to enlarge the British quota while concomitantly reducing the German, Irish, and Scandinavian allotments, and even more drastically reducing the quotas granted to more recent arrivals, who had not yet produced a large progeny . Since the notion of "national origins" constituted a putative ascendancy, for which there were no official records of any sort, the system required elaborate calculations and was not scheduled to go into effect until 1929; meanwhile, admissions were cut to 2 percent ofeach of the foreign-born groups as of 1890 (rather than 1910, so as to roll back the southern and eastern tide of the intervening years). The smaller percentage and the change of base year immediately reduced the annual total by more than half, to 165,000. Accordingly, entries from Europe as a whole fell to 148,366 in 1925 and remained below 170,000 a year for the remainder of the decade. Until 1929 the number of annual entries available to German nationals was quite large; hence, the fact cited by Massey that 412,000 German immigrants arrived in the 1920s is irrelevant. However , when the national-origin system came into effect, the German quota was lowered to 25,957, and this was reflected in an immediate reduction of admissions from 46,751 in 1928 to 26,569 in 1929. As for the other countries he mentions, there can be no doubt as to the acute and immediate effects of restriction. In support of his contention , he points out that the Italian decennial total reached 455,000; however, fully half (222,260) of these came in the single year 1921, before the nationality quotas went into effect: A first numerical limitation was then imposed, followed by a more severe one in 1924, as a result of which the average annual level of Italian arrivals for the remainder of the decade (1925 to 1930) fell to 14,969about 7 percent of 1921 entries.? The effect of Matters of State 75 restriction was even more acute for Poles: of the decennial total, 179,068-79 percent-came in 1921 alone, and the average for the postquota years was 8,111, slightly less than 5 percent of the 1921 level. Similar demonstrations of the effectiveness of restrictive regulation can be provided for other periods and other groups as well. During the Great Depression decade (1931 to 1940), U.S. immigration from all sources amounted to only 464,275, the lowest level since the 1840s, and immigration from Europe alone numbered 348,289-well below the authorized quota level, which would have amounted to some 1.5 million for the decade. Although this might be interpreted as prima facie validation of the "individual choice" approach as prospective immigrants became aware that the probability of employment was well below 1.0, in reality matters were not that simple. In the face of rising unemployment, Congress attempted to suspend immigration altogether; however, at the request of President Hoover, the State Department determined that it might be reduced well below the ongoing level without new legislation merely by strictly enforcing the clause prohibiting the issuance ofvisas to persons "likely to become a public charge" (LPC), following the successful use of this approach to reduce Mexican immigration (discussed later). Accordingly, State Department officials instructed European consuls that they were under obligation to consider any applicant dependent on wages as falling automatically into the LPC category (Divine 1957; Hodgdon 1931). This resulted in the immediate turndown of about 80 percent of applicants and the creation of extremely long waiting lists, whose existence indicates quite clearly that the dramatic reduction in European immigration in the early 1930s cannot be attributed in the main to the immigrants' calculations .8 The effectiveness of state power is attributable to the elaboration of the unprecedented system of "remote control" referred to earlier. Launched in 1924, this system required all foreign nationals coming from overseas to produce an entry visa prior to boarding a U.S.-bound vessel-a procedure now commonplace in air travel throughout the world. The possibility of developing such a system arose from the fact that, in relation to Europe , the United States was effectively an island. However, for this potential to be actualized, it was necessary to create practically de novo a body of professional consular officials with the authority to exercise gatekeeping functions (Hodgdon 1931) 76 The Handbook ofInternational Migration as well as additional regulatory power over U.S. and even foreign shipping companies. Once the system was put in place, prospective immigrants were screened before departure from their home country in accordance with the multiple criteria indicated , including a medical examination administered by U.S. Public Health physicians or local doctors certified by the U.S. government, to the point of making Ellis Island altogether obsolete. Those familiar with the contemporary scene might surmise that applicants turned down as immigrants might enter as visitors and then remain in the United States as "overstayers." However, the probability of success was low because American consuls were instructed to deny visitor visas to persons judged to be of the "immigrant class."9 It was still possible to enter surreptitiously by way of Mexico or Canada, but the latter country shortly adopted remote control as well, and Mexico was difficult to reach by direct travel from Europe.lO It is notorious that these regulatory mechanisms made it almost impossible for most refugees to gain admission to the United States from 1933 to 1945, a period of mounting political persecution and violence, despite a huge increase in demand (Feingold 1970; Wyman 1968, 1984). The immense visa waiting lists of the period from 1945 to 1965, extending up to fifty years into the future for the least favored nationalities, further suggest that despite various ad hoc laws that raised immigration from certain European countries in the postwar period above the quota level, absent the apparatus of restriction, incoming numbers would have been much higher. Asian exclusion provides an equally powerful demonstration of the effectiveness of regulation. As analyzed by Bill Ong Hing (1993a), measures of growing severity from 1875 to 1943 reduced decennial Chinese immigration from a high of 123,201 in the 1870s to 14,799 in the 1890s; the level rose slowly to 29,907 in the 1920s and was then reduced to 4,928 in the following decade. American policy not only prevented the expansion of established communities but, thanks to concerted efforts to keep out Chinese women, effected an absolute decline of the population of Chinese ancestry from 118,746 in 1900 to a low of85,202 in 1920, after which it grew slightly to 106,334 in 1940-still below the 1900 level. This successful containment of an abhorred population group constitutes a unique event in American history. Overall there can be no better demonstration of state effectiveness than the fact that in the 1920s the United States asserted to the world that it no longer wished to be "a nation of immigrants" and managed to counteract the dynamics of the international population system, including the effect of the networks that had developed in the wake of the immense immigration wave of previous decades . As a result of these efforts, by 1965 the proportion of foreign-born in the population had reached approximately 5 percent, its lowest level since Tocqueville came to America in 1830 and saw it as a nation of settled "Anglo-Americans." Immigration from Mexico Against this background of generally effective state action in regulating immigration from overseas, still the major portion of the U.S. intake, the persistence of unregulated movement from Mexico stands out as a deviant case and therefore warrants special theoretical attention.ll However, it is by no means clear what sort of deviant case it is. Considered in the light of perennial pronouncements by the U.S. government regarding its determination to eliminate illegal immigration, it constitutes an egregious instance of policy failure. However, the odd bundle of legal measures and administrative practices that have accumulated over the decades can also be thought of as amounting to a purposefully weak regulatory system that allows a relatively stable informal "guest worker" program to stay in place, in keeping with the clearly expressed preferences of powerful regional economic entrepreneurs , whose interests have come to be assumed as an obligation by the American state. Assessed in that perspective, on balance the policy must be reckoned a "success." Mexicans began migrating toward the Texas cotton fields in the later decades of the nineteenth century, when the unprecedented influx of American capital into mining, farming, and transportation transformed Mexico's isolated northern states into "the Border" (Katz 1981, 7). As steadily more draconian measures were deployed to prevent further Asian immigration, Californian fruit and vegetable growers also turned south to find workers to man the rising "factories in the fields" (Coolidge 1909). After 1910 many Mexicans fled northward to escape the upheavals of the Mexican revolution and the ensuing "secret war" waged by the United States against the revolutionaries. Although this was a period of steadily more forceful regulation of overseas immigration, by and large laissez-faire prevailed along the southern border. Following the entry of the United States into World War I, Congress explicitly exempted some seventy thousand Mexicans recruited as agricultural and railroad workers from the array of "qualitative" requirements designed to deter immigration from Europe, notably the LPC clause, the hefty head tax, the prohibition of advance contracts , and the recently enacted literacy test (Rico 1992, 243). After the war the exemption was terminated , but spokesmen for agricultural interests testified before Congress that they objected to any interference with free cross-border recruitment and pointed out that there was no way the requirements could be enforced. They also insisted that Mexicans came to the United States to work but not to live and were endowed with a "homing pigeon instinct" that drove them to return. Thus, Mexican laborers-commonly referred to as "wetbacks " (los mojados)-continued to flow freely throughout the period (Calavita 1992; Craig 1971; Garcia 1980, 142; Garcia y Griego 1988). Concurrently, formal immigration from Mexico expanded as well, despite the restrictive "qualitative regulations."l2 After the level passed the 50,000 mark in 1920 and reached nearly 90,000 in 1924, the congressional anti-immigration camp urged the adoption of limitative quotas to keep out Mexican "half-breeds," who, as nonwhites, were subject to Jim Crow laws in Texas. However, the State Department argued that this would be interpreted by the Mexican government as an unfriendly act and would interfere with ongoing negotiations to obtain compensation for U.S. properties nationalized by its more revolutionary predecessor. As an alternative, they pledged stricter enforcement of the literacy test and the LPC clause; given widespread poverty and illiteracy, this strategy would prevent ordinary Mexicans from qualifYing as legal immigrants (Divine 1957). In keeping with this objective, documented Mexican immigration was reduced to some 40,000 in 1929, 12,703 in 1930, and 3,333 the next year. The average was only 2,232 a year for the 1930s as a whole. The hybrid regulatory system that emerged along the southern border in the 1920s-an effectively enforced barrier to formal immigrationcombined with laissez-faire regarding the movement of labor, makes sense if the objective was to deter, not the physical entry of Mexicans into the United States, but rather their social and political entry into American society. American policy was highly effective in meeting this goal: Mexicans tended to return home when no longer needed, and in any case, their illegal status made it easy for American authorities, whenever they wished, to Matters of State 77 expel them without complicated legal procedures. Those who stayed on were confined by way of residential segregation and, in sharp contrast with the treatment of immigrants from Europe, were not subject to Americanization efforts. "Success" with Mexican immigrants was indicated by an extremely low rate of naturalization. How much permanent settlement resulted is difficult to estimate because the Bureau of the Census tended to view all persons of Mexican origin-including many born in the United States-as only temporary residents (Lorey 1990). Following the outbreak of World War II, the Mexican government's bargaining power increased because the United States wanted to secure its strategic support. Accordingly, the two states agreed to a labor program based on governmentsupervised contracts, generally known as the bracero program. However, informal recruitment continued alongside the bracero program because American employers sought to reduce costs by avoiding the regulations and workers sought to avoid exploitation by the Mexican officials who controlled the awarding of the contracts. At the end of the war the United States unilaterally terminated the program under conjoined pressure from employers and organized labor, which had abhorred "contract labor" of any kind ever since the heyday of anti-Chinese agitation. Informal recruitment then resumed, often with the cooperation of U.S. authorities, who legalized "wetbacks" whenever needed by rounding them up on the American side, pushing them over the border, and then readmitting them as temporary workers. Although a commission appointed by President Truman recommended the enactment of sanctions against employers of illegal foreign workers, in accordance with organized labor's wishes this remained a dead letter. The bracero program was reinstated in 1952 when the outbreak of the Korean War occasioned an acute labor shortage in the United States and concomitantly once again improved Mexico's bargaining position. This time Mexico demanded that the program also provide for employer sanctions , without which there would be no incentive for them to use legal braceros rather than the cheaper and more docile "wetbacks," and it also insisted that Texas be excluded because of its Jim Crow laws. In 1952 the United States duly enacted a law that prohibited aiding, harboring, and concealing illegal immigrants; however, agricultural interests secured the insertion of a clause (popularly known as the "Texas proviso") specify- 78 The Handbook ofInternational Migration ing that the usual and normal practices incident to employment, even the provision of shelter, "shall not be deemed to constitute harboring." The measure thus in effect institutionalized the procurement of illegal immigrants as an informal component of U.S. immigration policy, which continued alongside the bracero system. In the 1950s the state of California emerged as an actor with distinctive interests in this sphere: the preference of its growers for informal recruitment had occasioned mounting welfare obligations , for which the state had failed to secure additional revenue. Accordingly, the state's Republican governor, Earl Warren, urged his friends in the Eisenhower administration to implement more effective border controls. In response, the administration , from 1954 to 1955, mounted "Operation Wetback," which rounded up and summarily expelled several hundred thousand Mexican workers; the main objective, however, was to persuade U.S. employers to employ braceros rather than "wetbacks ." The operation was proclaimed a success when the annual number of contracts soared and apprehensions declined. In subsequent years the growing mechanization of cotton harvesting, stimulated in part by the higher cost of legal braceros, reduced the demand for temporary labor in Texas, leaving the expanding California fruit and vegetable industries as the main users. However, in 1960, after the CBS documentary Harvest ofShame established in the public mind a link between the bracero program and the plight of domestic migrants, the Eisenhower Administration decided to discontinue the program altogether. Its Democratic successors concurred , owing to pressure from labor; following a number of extensions urged by the growers, the bracero program finally expired in 1964. However , in the absence ofmore effective enforcement, this termination amounted once again to a reinstatement of unregulated procurement, whichaccording to apprehension reports, widely publicized by the Immigration and Naturalization Service (INS) to increase its budget-began to soar. Legal immigration again expanded in the postwar period, and by the early 1960s Mexico had emerged as the single largest source country. Arising at a time when the Kennedy administration's immigration reform proposal (discussed in the next section) was wending its way through Congress , this development prompted southern congressional leaders and other defenders of white America to impose an unprecedented ceiling on immigration from the Western Hemisphere as a whole; a decade later, when Mexican immigration stood at around 70,000 a year, the worldwide annual maximum of 20,000 per country (not including nonquota relatives) was imposed as well. This constituted in effect a reinstatement of the hybrid system elaborated in the 1920s. Amid the stagflation occasioned by the oil crisis in the 1970s, employer sanctions twice passed the House only to be defeated in the more conservative Senate. In the face of increasing outcry over the economic and social consequences of illegal immigration and growing agitation in the more and more politically powerful Sunbelt about expanding Chicano urban communities that were said to resist assimilation, President Carter appointed the Select Commission on Immigration and Refugee Policy (SCIRP) to get around the policy stalemate. In 1981 the commissioners recommended a package deal: the imposition ofsanctions on employers of unauthorized foreign workers , and the grant of permanent resident status to illegal aliens who had become de facto permanent residents by entering the United States prior to some specified date (Zolberg 1989). Completed in 1986 after much travail, the Immigration Reform and Control Act (IRCA) offered what was popularly termed an "amnesty" to illegal residents who had entered the United States before 1982. The law also sought to appease the growers byextending a special legalization program to some 1.2 million agricultural workers who had entered later. Although the law made it illegal for any employer to "knowingly" recruit, refer, or hire undocumented workers, it provided only for civil fines and absolved employers from any responsibility for verifying the documents presented to them. Moreover , mechanisms for enforcement remained rudimentary . Altogether, the legalization programs created a pool of some three million permanent residents and prospective citizens, who would eventually stimulate more immigration by way of family reunion. IRCA also provided considerable support for research on the effects of employer sanctions; it ovenvhelmingly suggests that after a slight initial deterrent effect, the impact of sanctions on unauthorized entry and employment remains minimal (Kossoudji 1992; Massey and Espinosa 1997; Massey and Singer 1995; Papademetriou, Lowell, and Cobb-Clark 1991). Although some skeptics contend that sanctions cannot possibly work because of the nature of unauthorized immigration, others attribute their ineffectiveness to inadequate enforcement and insist that sanctions would work if properly implemented. However, this would require enlisting employers in the regulatory process by imposing heavier responsibilities on them, as well as enhancing the federal government's police apparatus, notably by instituting some sort of verifiable ID card. Both possibilities have provoked considerable political opposition, from diverse locations on the political spectrum.13 Employer obligations were raised somewhat in 1996, but it is too early to tell whether this will render the system more effective. Post-1965 "New" Immigration Massey (1995, 638) also questions the relevance of regulation with regard to American immigration in the post-1965 period, asserting that the 1965 law, which eliminated the national-origin quota system, "was neither the sole nor the most important cause of the increase in numbers or the shift in origins," and he again suggests that "scholars have generally overstated the role of the 1965 amendments in bringing about the new immigration." It is quite true that the 1965 law was not the sole cause ofthe change in the quantity and characteristics ofimmigration in the intervening third of a century. This does not demonstrate, however, that immigration policy as a whole was irrelevant: the changes in quantity and sources that cannot be attributed to the 1965 law were occasioned mostly by American refugee policy and by the 1986 legalization program discussed previously, both ofwhich were very deliberate policy options. With regard to composition, the architects of the 1965 law mainly set out to eliminate the national -origin quotas that discriminated against southern and eastern Europeans; in their place, they established an annual maximum of 20,000 immigrants per country worldwide, to be allocated by way of ranked "preferences." The law initially facilitated additional immigration from Italy, Greece, and some other European countries, as intended . Although conservative opponents ofthe reform warned that the new universalistic approach would also foster an unprecedented growth of immigration from Asia and Mrica, its liberal supporters dismissed this prospect as a red herring. In retrospect, it is impossible to establish whether they were being disingenuous or merely lacked an understanding of global migration dynamics. In any case, the change ofpolicy must surely be ranked as a "cause" of the changing composition of immigration , even ifsome ofthe consequences it generated were unanticipated. Matters ofState 79 The quantitative expansion of legal immigration is attributable to the law as well. While it established a worldwide annual quota limit of 290,000 (170,000 for the Eastern Hemisphere, 120,000 for the Western)-raised to 366,000 in 1990-it also provided for new additional nonquota categories , notably the parents of citizens in addition to their spouses and children; these nonquota immigrants account for a very large part of the soaring legal immigration in the post-1965 period." In 1992, for example, they numbered 235,484, an increment of 64 percent above the 366,000 "baseline " annual admissions authorized under the several preferences (Fix and Passel 1994). Although SCIRP considered altering the preference system so as to reduce the "chaining" that accounts for this phenomenon, it was sharply divided on the subject, and perennial congressional efforts in this direction (notably in the course of what emerged as the 1990 law) have so far been unsuccessful, largely because of effective pressures from affected ethnic groups and constituencies. As things stand, there is no doubt that American immigration largely reflects the design of the country's immigration policy, which is highly attuned to family networks by virtue of congressional and presidential responsiveness to organized ethnic constituencies . Leaving aside the IRCA legalization program discussed earlier, the balance of the post-1965 growth in legal immigration is attributable to u.s. refugee policy. In 1992, for example, refugees and asylees numbered 134,290, an increment of 37 percent above the baseline level of 366,000. Although the 1965 law itself provided for only 10,200 refugees a year (6 percent of the 170,000 admissions allotted to the Eastern Hemisphere), after its enactment the U.S. government repeatedly invoked foreign and security policy imperatives to initiate much larger refugee programs, as it had done ever since the end of World War II (Zolberg 1995).15 Altogether, between 1945 and 1965, refugee programs increased admissions by about one-third above the level provided for by ordinary immigration regulations. Similar developments took place in the wake of the 1965 reform. Major episodes include the Cuban "freedom flights," initiated the very year the law was enacted;16 the resettlement of South Vietnamese associated with the United States at the fall of Saigon in 1975; the intake of Indochinese "boat people" from 1979 on; and the welcoming of Soviet Jews, whose freedom of exit was secured by holding up the threat of trade sanctions. In 80 The Handbook ofInternational Migration 1980 Congress enacted a new refugee law, which took refugees out of the quota system altogether and provided instead for the admission of individual asylum seekers in keeping with U.S. obligations under international law, as well as of others whose admission might be "in the national interest ," as determined by the president in consultation with Congress. It was estimated that refugees would not exceed 50,000 annually (including some 5,000 qualified asylum seekers); however, by the end of the decade refugees and other humanitarian admissions added between 100,000 and 150,000 a year to the immigration total. It might be noted that after they settle in, asylees and refugees also generate ordinary immigration by way of family reunion. However one might assess the desirability of these additional flows, there can be no doubt that they are not occasioned by a mere increase in the pool of refugees in the world at large but are entirely the result of deliberate American policy choices. The situation with regard to El Salvador and Haiti is more ambiguous; although American authorities generally turned down individual asylum seekers from those countries and, in the case of Haitians, even resorted to proactive interdiction on the high seas, many of them nevertheless secured an extended de facto haven in the United States thanks to legal actions undertaken by political supporters and a residually permissive societal environment. The 1990 immigration law, which provided temporary protected status for several hundred thousand illegal Salvadorans, in effect converted them into ordinary immigrants. Evidence from the World at lArge Additional evidence on behalf of the important role of state regulations in shaping immigration can easily be found outside the United States. One of the most dramatic demonstrations is provided by the United Kingdom. Notwithstanding its standing in the first third of the twentieth century as a country with one of the highest per capita incomes in the world and attractive political conditions , as of 1930 it had the lowest proportion of foreign-born of any industrial country on either side of the Atlantic (Kirk 1946), an outcome for which the immigration restrictions it imposed as early as 1905 and reinforced in the 1920s provide the most likely explanation. The growth and composition of immigration to the United Kingdom in the immediate post-World War II years are also attributable to specific policies, notably the decision to initiate and then discontinue the procurement of European "displaced persons" as supplemental workers; the maintenance ofan "open-door" policy with regard to Ireland, even after it became fully independent; the largely politically motivated 1948 decision to authorize entry and settlement by "New Commonwealth" citizens; and the rectuitment of West Indians into the public sector (notably London Transport) as a way of reducing political tensions in the colonies (Rees 1993). From 1962 on, British immigration policy became increasingly restrictionist, and by the early 1980s legal immigration was reduced to some 50,000 a year-not including the Irish-with a reportedly low level of unauthorized settlement. Although there is no doubt that this outcome is attributable to British policy, there is some question as to whether it represents a policy success or failure: on the one hand, the annual level of immigration is well below the putative "demand" for entry, and much lower than the actual level of immigration to France or Germany. (The contrast is even greater when population size is taken into consideration.) On the other hand, it remains above the United Kingdom's "zero immigration" objective (Freeman 1994; Layton-Henry 1994). Is the glass half full, or is it half empty? The question is applicable to the regulatory capacity of affluent liberal democracies more generally. Overall, as indicated earlier, the pessimist's perspective has gained the upper hand. This judgment is founded on the notion of a gap between the goals of immigration policy and the actual results achieved, but the preceding discussion suggests that the goals themselves are often ambiguous as well as grossly unrealistic. The significance of the checkered history of U.S. policy regarding its southern border can be elucidated by reflecting on parallels with two other cases of regulation, traffic control and gun control. With regard to the first, the United States has emerged without any apparent contention as a "strong state": it deploys an elaborate system of regulations and subjects every aspiring driver to fairly strict testing and the requirement of a relatively foolproof identification document, which can now be checked in a central location by just about any policeman in the country. Although they are administered by a variety of state and local authorities, these regulations are nevertheless fairly uniform nationwide. Overall, this approach has contributed to the achievement of a remarkably low international ranking in the rate of automobile accidents, a situation that has in fact steadily improved. By contrast, with regard to gun control the United States is a "weak state": it has much more permissive regulations than in comparable societies abroad, generally inadequate enforcement of whatever regulations do exist, and consequently a comparatively high rate of homicide and gunshot wounds. Although a number of factors contribute to this outcome, the determination of a well-organized minority to block the elaboration of an effective regulatory system, while effectively enlisting patriotic sentiment in its cause, is generally acknowledged as a major contributing factor. Although the prevalence of traffic control indicates that there are no fundamental structural or cultural impediments in American society to the elaboration of a complex system of regulation involving intrusive law enforcement that affects nearly every individual, the case of gun control suggests that the absence of an effective regulatory system is itself a "policy." American immigration policy with regard to incoming overseas movement may be more like traffic control, but its stance regarding movement across its southern border resembles its stance toward gun control. In that light, the high incidence of unauthorized "back-door" migration can hardly be taken as evidence of the ineffectiveness of policy.17 The example of traffic control brings up the fundamental importance of selecting realistic objectives in evaluating the effectiveness of policy. No regulatory agency expects to achieve zero speeders or zero accidents; rather, the goals usually consist of a combination of improvement in relation to some established baseline (for example, bringing down the annual rate of accidents in relation to the number of cars on the road, or the number of miles traveled) and the achievement of specific targets, such as the elimination of particularly hazardous spots. Although the INS publishes "apprehensions," this is a "rate" only in relation to time (that is, the number per year), not in relation to actual traffic (attempted crossings).18 With regard to immigration regulation, the most relevant denominator would be some estimate of "potential immigrants," with actual immigrantsboth legal and unauthorized-as the numerator. Although it is impossible to establish such a denominator with quantitative precision, it is not very difficult to set forth a credible order of magnitude . It is reasonable to surmise that the pool ofinternational migrants aspiring to relocate to the affluent countries of the North has continued to expand under the combined effects of demographic Matters of State 81 growth in the poorer countries, persistently unequal conditions in the world at large, and widespread information about opportunities elsewhere. Because of the availability of cheap transportation, the nearly total lifting of state-imposed barriers to exit, and the multiplication of networks, potential migrants possess an unprecedented capacity for moving. Yet the combined intake of the affluent countries-including both legal and unauthorized migrants, granting the most generous estimates of the latter-has certainly leveled off since the 1970s, and possibly decreased overall. This is applicable even to the United States, where legal immigration appears to have peaked, and by all reports the rate of unauthorized immigration has reached a plateau as well. In short, international migration theories will remain woefully incomplete so long as they fail to take into account the positive and negative roles of states in shaping international population movements. INTERNATIONAL MIGRATION POLICIES IN A BOUNDED WORLD A Global Perspective The starting point for theorizing about migration policies is an understanding that they enter into play not merely as "error factors" in an equation whose parameters are social and economic, but as constitutive elements of international migration as a distinctive social phenomenon. In short, international migration is an inherently political process, which arises from the organization of the world into a congeries of mutually exclusive sovereign states, commonly referred to as the "Westphalian system" (Zolberg 1978a, 1981a).19 It involves the transfer of a person from the jurisdiction of one state to that of another-in whole or in part-and the eventuality of a change of membership in an inclusive political community. Accordingly, the relevant policies encompass not only the regulation of outward and inward movement across state borders-including the movement ofpersons who are not, or declare that they are not, migrantsbut also the rules governing the acquisition, maintenance , loss, or voluntary relinquishment of "membership" in all its aspects, political, social, economic, and cultural. Migration policies vary enormously, both historically and between states in a given period. States have at times prohibited the exit of just about all of their population and used draconian means to 82 The Handbook ofInternational Migration implement this policy, such as the imposition of galley slavery in seventeenth-century France to prevent Huguenots from departing for Protestant states, or shoot-to-kill policing in the German Democratic Republic as late as a decade ago. At the same time, they have also acted ruthlessly to push out religious, ethnic, or social groups, and at times laissez-faire has prevailed. All these stances coexisted during the cold war period, suggesting that there is no overall historical trend toward convergence. A similar range of variation can be found on the entry side. States have encouraged and facilitated the importation of slaves, and they have stimulated immigration by providing subsidized travel, lands, security, and easy citizenship or by promising jobs. However, they have also prohibited settlement or, again, adopted a laissez-faire stance. Positive or negative stances with regard to both exit and entry are usually defined in relation to specified categories of persons established on the basis of a wide array of criteria, including objective socioeconomic and cultural attributes (degree of skill, education, wealth; religion, language, nationality, race) as well as putative moral or political disposition (judged likely or unlikely to commit crimes or to support or oppose the regime). As a result, emigration and immigration policies often amount to complex arrays of disparate regulations . States also exhibit considerable variation regarding modes of relinquishing and acquiring membership, including not only formal citizenship but political, social, and cultural rights. How are we to make theoretical sense of this variation? At the most general level, "whether migration is controlled by those who send, by those who go, or by those who receive, it mirrors the world as it is at the time" (Davis 1974, 96). In terms more consonant with the IUSSP committee 's approach, the world can be conceptualized as a "global population system" in relation to which sending and receiving states, much like the migrants themselves, figure as "utility-maximizing" agents that respond to changing world-historical and local conditions by modifYing their comportment -in the case of states, their policies regarding exit and entry. However, two qualifications are called for. As the complexity of contemporary debates on immigration policy in the affiuent liberal democracies indicates, "utility" encompasses not only a population 's economic value but also its putative value in relation to cultural and political objectives. Moreover , "utility maximizing" cannot be mechanically transposed from individuals to states. As executors of policies, states do not function as autonomous actors (even in the loose sense in which we consider individuals to do so), but rather as instruments manipulated by internal actors who have gained the upper hand in a particular sphere at a given time. Legal and administrative institutions, as well as political traditions, which constitute the legacy of earlier policies, also playa significant role in shaping current responses. Historically, migration policies have been shaped by the dynamics of world capitalism, on the one hand, and the dynamics of the international state system, on the other, within the context of epochal population dynamics (Zolberg 1981b, 1983a, 1986). Since the global population system and the system of states are both finite, migration policies are extremely interactive: any emigration always entails immediate immigration somewhere else; conversely, the possibility of immigrating affects decisions to emigrate; and the closing or opening of a particular national gate affects the potential flows into other states.20 However, in contrast with the sphere of international trade, for example, far from being founded on a recognition of this interactivity , state policies regarding emigration and immigration have been notoriously unilateral. As noted by Hannah Arendt (1973, 278), for example , "sovereignty is nowhere more absolute than in matters of emigration, naturalization, nationality, and expulsion." This highlights the theoretical significance of even slight departures from the sovereignty baseline observable today (Soysal 1994). Exit and entry policies might be arrayed along an axis defined by negative and positive poles. Given the considerable variation in exit policies that can be observed, the impact of these policies on the formation of migration networks, and their interactivity with immigration policies elsewhere, it is evident that a comprehensive theory pertaining to the role of states in regulating international migration must cover the exit side as well. One basic proposition is that the possibility of preventing exit is a requisite for the effective exercise of most types of "predatory rule" (Hirschman 1981; Levi 1988). In the early modern era, under prevailing conditions of demographic scarcity, for the European mercantilist and bellicist state acquisition and retention of human capital for economic production and war was a basic source of power; from this perspective, the most important form of control pertained to outward movement. Accordingly , unauthorized emigration was tantamount to treason and punishable by death or enslavement. The continuing significance of this policy stance is highlighted by its resurgence in the twentieth century as the hallmark of totalitarian states with command economies (Dowty 1987; Zolberg 1978b). The leading patterns of forced exit, including deliberate expulsions as well as escape from persecution or violence, can also be accounted for by the dynamics of state and nation formation within a bounded international system (Zolberg 1983b; Zolberg, Suhrke, and Aguayo 1989). However, given space limitations and this volume's exclusive concern with the United States, in the remainder of this chapter I shall focus on the entry side only. Although considerable attention has been devoted to variation among the contemporary immigration policies of capitalist democracies (Brochmann 1996; Cornelius, Martin, and Hollifield 1994a; Freeman 1995a; Kubat 1979a), the most striking fact about them is that, if one imagines a hypothetical continuum ranging from open to closed borders, they are clustered very narrowly around the closed pole. Notwithstanding the post-World War II modifications noted earlier, which constituted a genuine "liberalization" in relation to the extremely restrictionist regime established by the affluent states in the first quarter of the century, the near-zero baseline remains a constitutive element of the contemporary regime with regard to the supply of entries in relation to the demand for them. As the theorist of international trade Jagdish Bhagwati (1984, 684) noted some time ago, the process of international migration is characterized by "disincentives" rather than "incentives "; this observation led him to hypothesize quite correctly that if the socialist countries were to let people out, "the effective constraint on the numbers migrating would soon become the immigration legislations of the destination countries." Most contemporary European states have in effect adopted a zero-immigration objective but make very limited provisions for asylum seekers and immediate relatives; the traditional immigration countries, including the United States, are similarly determined to keep out the world's population , in general, but make somewhat broader provisions for family reunion and retain an acquisitive element in the form of limited programs for admitting persons who might contribute substantial capital or skills. Coming to grips with this stark reality is a prerequisite for any meaningful theoretical discussion of the role of states in regulating immigration. Arising from the aggregate policies of the states of the international system, and undertaken in reMatters ofState 83 sponse to domestic considerations as well as interactively in response to each other's policies, the restrictive immigration regime prevails worldwide because it constitutes a sine qua non for maintaining the Westphalian international state system, as well as the privileged position of the core states amid highly unequal conditions (Emmanuel 1972; Nett 1971; Petras 1980; Wallerstein 1974).21 In less exalted language, economic modeling suggests that the hypothetical elimination of borders would stimulate worldwide economic growth but also result in an equalization of conditions, thus producing a vast redistribution of income to the benefit of the population of poorer countries. In effect, borders prevent labor from commanding the same price everywhere and also prevent people from the poorer countries from gaining access to the bundles of "public goods" dispensed by the more affluent states, which constitute an important part of their populations' income (Hamilton and Whalley 1984). It is also widely believed that restrictions on access to membership constitute a sine qua non for democratic governance, which requires at least some minimal degree of community. Although there is intense debate over precisely what level of immigration might be allowed for, and what priorities might be established, there is also broad agreement that under present world conditions the level would at best fall far short of the demand for access (Barry and Goodin 1992; Carens 1987; Walzer 1981). Accordingly, the process ofdecisionmaking at the level of a given state is driven by two quite distinct types of considerations, each of which relates to a distinct sphere ofsocial interaction. In the perspective of capitalist dynamics, immigrants of any kind-including refugees-are considered primarily as workers. Policies are shaped by the prevailing "class compromise" and by the specific configuration of economic interests in the country in question , in keeping with the imperatives of prevailing technological and economic conditions (Offe 1984; Przeworski and Wallerstein 1985). Immigrants are characteristically welcomed by employers because they reduce the unit cost of labor (that is, they lower wages) and also increase its elasticity; conversely, they are characteristically resented by resident workers as unfair competitors willing to accept lower wages (which constitute an improvement over their income in the country of origin) and below-standard conditions. At worst, they may not only lower wages but altogether displace natives . However, even the most profit-driven capital- 84 The Handbook ofInternational Migration ists are unlikely to favor a huge increase in labor supply that would occasion a major social disruption ; hence, in contemporary capitalist democracies , arguments on behalf ofopen borders appear perennially as the playful musings of free-market ideologues, but almost never in policy debates. These considerations, usually cast in a Marxian framework, have given rise to a considerable body of work accounting for the tendency of advanced industrial societies to recruit "guest workers" from less developed countries (Burawoy 1976; Castles and Kossack 1985; Petras 1981). An alternative explanation was provided by the theol)' of labor segmentation, whereby under the conditions of the welfare state the upper strata of the workforce is assimilated into "fixed" capital, leading to the institutionalization of a distinct "flexible" segment -for which, again, guest workers of one sort or another are vel)' convenient (Gordon, Edwards, and Reich 1982; Piore 1979). Still in the economic sphere, immigrants are also consumers of goods and services, both the ordinary kind that are bought in the marketplace and the public goods that are automatically available to all residents. Immigrants tend to be welcomed by sellers of individual goods and services-for example, by real estate agents in port-of-entry cities-but from the perspective of public goods the situation is more complex. Recent U.S. debates regarding the "balance sheet" of immigration in relation to welfare highlights the difficulty of establishing whether immigrants contribute more in taxes than they consume in public services or whether they are "free riders" (Smith and Edmunson 1997).22 Because different units of aggregation are used to draw up a balance sheet of the various costs and benefits involved, immigrants may be "good" for the whole economy but "bad" for a particular locality or social group-or vice versa. (For example, they may reinvigorate declining Rust Belt cities.) Incidentally, similar considerations are also applicable to international-level assessments of the economic value of particular human flows by sending and receiving countries as a whole. However, all types of immigrants-including even temporal)' workers-also constitute a political and cultural presence, the putative impact of which on the host country's "way of life," "cohesiveness ," or, in current discourse, "identity" evokes a distinctive dimension to be considered. Although the process in question is well evoked by classical sociology's concept of "integration," I shall use the term "identity axis." In almost any immigration situation, there are significant groups among the hosts who believe that newcomers in general, or particular groups among them, would jeopardize the established national ways. In the United States alone, just about every cultural attribute imaginable has been found objectionable at one time or another, notably "race"-as constructed in the nineteenth and early twentieth centuries , referring not only to "Asiatics" and blacks but also to "mixed-breed" Mexicans, different European nationalities, and Jews-religion (particularly Roman Catholics from the eighteenth centul )' until quite recently), and language (German speakers in the early twentieth century, Spanish speakers today). Similarly hostile responses have surfaced elsewhere, notably toward Jews from eastern Europe in most of Western Europe yesterday , "Arabs" and Muslims more generally throughout Europe today, and the Sephardic Jews who arrived in Israel in the 1950s, and Ethiopian Jews later on. Although reactions such as these are attributable in large part to persistent prejudice and xenophobia , which tend to exaggerate the problematic aspects of the situation, it should be recognized that the settlement-or prospective settlementof any substantial group of people whose culture diverges markedly from the host's own is likely to call the established "cultural compromise" pertaining to religious, linguistic, and racial diversity into question and hence is a legitimate source of concern .23 The key questions are always: How different can we be? How alike must we be? And how are the answers to these questions to be implemented organizationally and materially? (Zolberg and Long 1999). As it enters into play with regard to immigration , "identity" centers on nationality.24 Originating largely in the course of efforts to institutionalize "predatol)' rule" in late medieval European states, modern nations have come to be perceived by most of their members as familylike bodies, with a common ancestry and a common destiny. Although the formula for identity is usually founded on some objective characteristics of the society, such as the language actually spoken by much of the population and the religion many of them share, the culture of the rulers themselves is always accorded pride of place. However, political culture and ideological orientation may be invoked also, as in the United States and France following their democratic revolutions (Brubaker 1992), or in the confrontation between "the West" and "the East" in the cold war decades.25 Much less emphasized by writers on nationality and nationalism is that the formation of national identity always involves a negative aspect as well: we are who we are by virtue of who we are not. In this light, nationality involves the delineation of a boundary that denotes simultaneously inclusion and exclusion. Whatever the objective realities may have been in the early modern era, princes and their serving intellectuals emphasized similarities within the national borders and differences between the nation and its neighbors (McNeill 1963). Although the negative "others" are commonly close neighbors with whom perennial wars are fought, and from whom "we" must distinguish ourselves by any means possible, they can also be remote aliens, regarding whom little is known and therefore much can be invented. Groups originally recruited as low-skilled "workers" are especially likely to belong to the "not-us" world and therefore to be considered by many as "unqualified" for membership in the host society; this "wanted but not welcome" syndrome creates especially problematic situations if the workers become permanent settlers (Zolberg 1987). Differing assessments along these lines precipitate confrontations not only between "natives" and "foreigners" but among the natives themselves -between those who perceive the newcomers as a threat in relation to what is deemed a fragile status quo and others who are more confident in the society's ability to weather change or who welcome the diversity the newcomers would contribute as an enrichment. These alignments are probably related to a more comprehensive cultural cleavage that is emerging as the contemporary equivalent of the older rift between religious and secular camps; it encompasses other "cultural" and "moral" issues such as abortion, feminism, gay rights, and the death penalty. However, the camp of those positively disposed toward immigrants may also include natives who are not particularly open to change but feel an affinity with particular groups of newcomers, notably fellow religionists or ethnics. Sometimes this camp comprises the national community as a whole with respect to populations located in other states who are regarded as "external nationals," on whose behalf the state may devise a "law of return" or some other extremely generous immigration policy.26 As noted in the previous section, refugee policy has tended to be driven by strategic considerations Matters of State 85 arising quite directly from the dynamics of the international political system: providing asylum to the victims of one's enemies was consistent with the imperatives of realpolitik in that it demonstrated the antagonist's evil ways and undermined its legitimacy. Concomitantly, refugees tended to be ranked high on the positive side of the "identity " axis throughout Western countries almost by definition: they were welcomed exclusively on the basis of religious or political affinity with the receivers and therefore were deemed not strangers but brothers and sisters in need. Under these conditions , statecraft and humanitarianism went hand in hand (Zolberg 1983b; Zolberg et al. 1989). By the same token, states were not inclined to help victims who were not "like us": proletarian Communards had almost no place to go after their defeat in 1871, and Jewish victims of Nazism were denied havens as well. However, in the post-World War II period the international community began moving toward a more universalistic approach, eventually extending refugee status to all those, anywhere in the world, who are outside their country and without government protection as a consequence of "reasonable fear" of persecution. Concurrently, the superpowers expanded the domain of their strategic confrontations to encompass many regions of the Third World, contributing to a vast enlargement of the refugee pool. Although the overwhelming majority remained in their region of origin, some came knocking at the door of the affluent countries , and the fact that an increasing proportion of those who sought asylum were poor people of color-and thereby akin from the perspective of the receivers to immigrant workers-triggered alarm bells and prompted a reconsideration of established policies. This revisionism was facilitated by the end of the cold war, which eliminated at one blow the "realist" foundations of the postwar refugee regime and sharply narrowed the scope of Western states' refugee policies. The United States retained a refugee policy founded almost exclusively on "realist" foreign and security considerations until about 1980; although it then subscribed to the international regime, its actual policy continued to be driven by "realism," with some intrusion of constituency pressures (particularly with regard to Eastern European Jews). The persistent coexistence of these two very different dimensions of consideration and, concomitantly , of interests-one pertaining to the putative or actual effects of immigration on material condi- 86 The Handbook ofInternational Migration tions, the other having to do with cultural and political conditions-can be represented by crosscutting axes, each with positive and negative poles, providing for a continuum of alignments from "for" to "against." Hence, it is possible to adopt a positive position on immigration with respect to one dimension, and a negative position in relation to the other. This accounts for the often remarked upon tendency of immigration politics to straddle the traditionalliberal -conservative divide and for the emergence of "strange bedfellow" coalitions for or against particular proposals. Successive attempts to resolve these disparate imperatives in the face of changing conditions shape immigration policy into complex and often inconsistent configurations , such as the segmentation of U.S. policy into a "main gate" dealing with general immigration and a "back door" dealing with the procurement of temporary agricultural workers. Political Process and Political Institutions However powerful, the effects of social forces, external or internal, are not automatically translated into policy outcomes but are mediated by established political structures-notably political parties , electoral systems, and the articulation and organization of specialized interests-and political institutions-notably the structure of representation and the allocation of decisionmaking authoriry and power in the relevant sphere between the executive, the legislative, and the judicial branches of government. These are, of course, the factors on which political scientists have long trained their specialized optical apparatus. The initial emphasis, on the American side, on the role of interest groups yielded important insights into specific cases of immigration policy but fell short of theoretical elaboration (Craig 1971; Divine 1957; Riggs 1950). Recently, however, there have been more self-consciously theoretical efforts. One valuable contribution is an analysis of the role of distinct modes of interaction between elites and the public in shaping immigration policy (Freeman 1995a). In short, each dimension might foster a distinct mode of interaction between elites and the public, with concomitantly distinct outcomes . On the basis of his own comparative studies of immigration policy in Europe and the United States, Gary Freeman asserts that the political dynamics of immigration policy in liberal democracies "exhibits strong similarities that are, contrary to the scholarly consensus, broadly expansionist and inclusive" (Freeman 1995a, 881). His theoretical starting point is a simple model of politics in which the state actors who make policy are "vote-maximizers" with no special views or interests of their own (except, as we shall see later, for an unexplained commitment to "universalism ") but are responsive to pressures from the public, consisting of "utility-maximizers" assumed to have complete information about the consequences of policy alternatives. Following James Q. Wilson (1980), Freeman (1995a, 894) argues that the public's mode of organization varies as a function of how the costs and benefits resulting from policies are distributed. In the case of immigration, benefits are "concentrated "-notably by providing lower costs for employers in certain economic sectors and gratification for the kin and co-ethnics of incoming groups-whereas costs tend to be diffuse; they include increased competition for jobs among some groups of the resident population and increased demand for certain services. Moreover, the general public's utility-maximizing ability is handicapped by "serious barriers to the acquisition of information " about immigration, and by a "temporal illusion " whereby the short-term benefits of immigration are easily seen but its long-term costs are denied or hidden. Such a distribution of costs and benefits tends to produce "client" politics: small and well-organized groups intensively interested in a policy develop close working relationships with the officials responsible for it, largely outside of public view and with little outside interference. Consequently, policymakers are more responsive to their immigration -advocating clients than to the more ambivalent or even opposed general public. How and why the public came to hold such views, however, is not specified. As a result, "official policies tend to be more liberal than public opinion and annual intakes larger than is politically optimal" (Freeman 1995a, 897)-where "optimality" is defined as the policies preferred by the median voter. However, the situation may change as a function of two predictable cycles, the one economic, the other arising from the tendency of migration, through networks, to generate more migration-a proposition that is well established in general international migration theory. Although under those circumstances decisionmakers adopt a less expansive stance, Freeman (1995a) insists that the basic dynamics intervene to mitigate reactions and to prevent abrupt changes of policy. Distinguishing three groups of receivers, the English-speaking "settler societies," European "postcolonial" and guest-worker societies, and Europe's "new countries of immigration"-notably Italy, Spain, Portugal, and Greece, to which one might now add some post-Communist countries such as Poland-Freeman (1995a, 892) claims that his "model" accounts for recent developments among all of them. Although all three groups are affected by the conjunction of the two cycles, the cycles' effect on settler countries is mitigated by the persistence of an immigration identity . In the second group, as costs come to be perceived as more concentrated, client politics evolve into interest-group politics and the dynamics of politics are partly reversed, with elites taking the lead in adopting a tightly restrictionist stance. In commenting on Freeman, Brubaker (1995) has correctly pointed out that the "constrained discourse" regarding ethnicity and race is not a constant feature of liberal democracy and cannot be derived from Freeman's "political economy" model; this is true also of what Freeman (1995a, 892) terms the settler countries' "positive folklore " toward immigration. Both features are historically specific and contingent, and Brubaker suggests more generally that the analysis must make room for a "cultural-political" story that is logically independent of political economy-a suggestion consonant with the framework suggested here (Brubaker 1992, 1995). Indeed, despite his emphasis on economic considerations, Freeman's references to the relinquishment by elites of ethnic and racial categories, as well as to the emergence of the "new Right," reflects his awareness that identity issues do at times play a negative role, and his reference to the pro-immigration stance of kin and co-ethnics suggests that he understands that identity can have a positive value as well. Freeman's view of elites exclusively as economic actors may also be the source of his erroneous assertion regarding their universalist orientation. As he himself has shown elsewhere-and as he acknowledges in a note in the present work-in Britain racial considerations entered into play from the late 1950s onward (Freeman 1979; 1995a, 891 n. 10). Contemporaneously, in the United States, as noted in the first part of this chapter, while abolishing the national-origin quotas and Asian prohibitions, American political elites imposed unprecedented limits on immigration from the Western Hemisphere with the explicit objective of restricting the immigration of Mexicans and black West Indians. Today as well, on both Matters of State 87 sides of the Atlantic, elites hardly display a uniformly universalist orientation in immigration politics . Overall, the situation with regard to identity might be the reverse of what prevails in the economic sphere: the costs of immigration are diffuse, in the sense of a malaise pertaining to "threats to nationality," whereas the benefits are concentrated , in that certain ethnic groups increase their weight and hence their "recognition" and potential political power in the nation. This may account for the reluctance of U.S. elites around the turn of the century to endorse immigration restriction , and for the movement of U.S. elites toward universalism, notably with regard to the elimination of the national-origin quotas directed against Europeans in the 1950s, when these groups became the mainstay of the urban wing of the Democratic Party. Both dimensions make for client politics but involve different sets of clients. The combination of business and of ethnic groups as strange bedfellows in a pro-immigration camp is a characteristic U.S. outcome, and partly a function of the rapid incorporation of immigrants into the body politic (see figure 4.1). However, the concentrated/diffuse measure should not be applied simplistically. Within the economic sphere alone, for example, since immigration is spatially concentrated, port-of-entry localities may simultaneously reap more benefits than the country at large from the economic stimulation immigration brings about and face a particularly heavy demand for a variety of services; moreover, the perception of costs may be enhanced where local governments have some fiscal autonomy, notably in federal states. Under such conditions, local governments are likely to emerge as important actors in immigration politics, but local circumstances determine whether they will be found on the negative side (for example, New York City in the 1830s or California today) or on the positive side (New York City today).27 In a similar vein, analysis of the distribution of economic benefits is rendered more difficult by the need to distinguish between "the public" as workers and as consumers: in the United States, for example , not only is the use of unauthorized foreign workers in the produce and garment industries profitable for employers and presumably costly to the workers in those sectors, but it also appreciably lowers the price of these widely used commodities -including for the workers in their role as consumers. The concentrated/diffuse measure thus becomes more heuristic if it is applied to dis- 88 The Handbook ofInternational Migration FIGURE 4.1 The "Strange Bedfellows" of American Immigration Politics Putative Cultural/Political Effects + Co-Ethnics Cosmopolitans Native Workers Local Authorities ,- / / / / / / ~e" / ~o / cp'lf / .r} / ~ ~o / .~'lf / ----------~ .... '-----7t'----- ~e" - - - - - - - - - - ,~ ~o / o'lf / ",0 / ~e" / ~o / ;,.,Ci ~ / ~0 / / / / / / / Employers Transporters + Putative Economic Effects Traditional Nationalists aggregated elements, and when a distinction is made between concentration within economic sectors and concentration in space.28 A more serious limitation is that because it emphasizes process exclusively, the model Freeman borrows from Wilson is incapable of providing insight into structure. Although the distribution of costs and benefits of particular policies does shape political dynamics, policy issues do not arise in a political vacuum but rather in a field structured by previous historical experiences, including ongoing policies in the sphere under consideration-about which the model has little or nothing to say. This is evident, first of all, with regard to immigration policy itself. As demonstrated in the preceding section , contemporary policymaking takes place within the context of a prevailing worldwide restrictive immigration regime, which has to be explicitly accounted for. Another matter is that although Freeman speaks of "immigration policy" as of a piece, this ignores an important institutional reality: all states today distinguish between refugees and ordinary immigrants, and in effect also between settlers and workers. Accordingly, rather than a single overall dynamic, we should expect a different process to prevail in each of these policy areas. Within the sphere of political economy alone, there is a considerable range of variation in how the principal "class" actors are organized, and this in turn makes for major differences in the process and substance of policy: strong or weak labor unions, industrial or craft organization, the presence or absence of "peak associations" among workers and employers, corporatist social compacts or unruly pluralism (Schmitter and Lehmbruch 1979). These structures account not only for variation in political dynamics but also for some variations in policy-such as the organization of formal guest-worker programs in the more corporatist European countries (Germany, the Low Countries, Sweden) as against employerdriven programs, which often involve processes of marginal legality where unions do not participate in the making ofindustrial policy (as in France and the United States). This is applicable to the identity dimension as well. For example, the ethnic organizations established by earlier immigrants in the United States have become legitimized as political interlocutors beyond what might be expected on the basis of their electoral weight, forming in the cultural sphere an equivalent of the corporatism that is sometimes found in the political economy sphere (but to a very low degree in the United States); this distinctively shapes the dynamics of immigration politics and also patterns the organizational strategies of more recent immigrants. The standing of these ethnic organizations as clients is thus not merely a function of the concentration of benefits fostered by immigration policy, as demonstrated by the fact that in the absence of this corporatism they do not become clients despite a similar concentration of immigration policy benefits . The Maghrebis in France are a case in point: it is possible to write a detailed account of the politics of French immigration policy in the recent past without any reference to them as organized actors (Weil 1991). Some of these problems are addressed quite explicitly in a recent study by Keith Fitzgerald (1996, 1), who contends that the application of his "improvisational institutionalism" approach enables him to "solve the most puzzling features" ofAmerican immigration policy for the period 1879 to 1965. The work is founded on a disaggregation of immigration policy into three segments dealing with permanent residency ("front-gate immigration "), refugee policy, and "unsanctioned migrant laborers across the U.S. border with Mexico ('backdoor immigration')." The core ofFitzgerald's argument is that these segments display distinct policy dynamics, each of which can be accounted for in one of the leading contending theories of policy formation. In his enticingly elegant construction, policy regarding the "front gate" is shaped by the Matters of State 89 relatively free play of competing societal interests (political science's traditional pluralism), whereas refugee policy is shaped by "realism" (in which the state looms as a major agent pursuing interests ofits own) and "back-door" policy comes close to fitting classical class-conflict theories. Inspired by the "structuration" approach of Anthony Giddens (1979), Fitzgerald views the policymaking process as "episodic," with innovation commonly arising from improvised solutions to pressing problems. Although the disaggregation of immigration policy into discrete components rightly captures an often ignored feature of policy reality-not only in the United States-Fitzgerald is somewhat carried away by the elegance of his theoretical construct. In fact, all three modes of policy formation occur in each of the components indicated. For example, although I fully concur that refugee policy has been driven to a considerable extent by foreign policy considerations, with the policymaking process centering in the foreign- and securitypolicy sector of the state, pluralist elements have come into playas well, notably by way of ethnic constituency pressures for and against the award of refugee status to particular groups. Conceptualized as distinctive components of any policymaking process, Fitzgerald's three dynamics nicely complement Freeman's concentrated/diffuse measure. However, by itself such a formal approach cannot account for the circumstances under which policymaking in all three spheres arises and changes. For example, it is evident that the weight of realist considerations in the policymaking process is not a constant, but that it rose and fell with the advent of the cold war; hence, there was a turn "from invitation to interdiction" (Zolberg 1995). It is therefore most useful when linked to the sort of historical and externalist framework that is being set forth here. CONCLUSION Reflecting the traditional division of labor among the social sciences, theorists of international migration have tended to ignore or minimize the role of states in shaping their object ofstudy, while theorists of migration policy have tended to approach their subject in a domestic perspective, ignoring the dynamics of the global environment in relation to which regulation of exit and entry occurs , especially its demographic aspect. The framework presented here is a very preliminary attempt to overcome these deficiencies and to use the in- 90 The Handbook ofInternational Migration sights of both sides to construct a more comprehensive theoretical edifice. The twin global dynamics of capitalist development and of state formation generate powerful thrusts of necessity and force that propel large masses of humanity outside their counny of origin , but the international migrations that actually take place are shaped to a considerable extent by the will of the world's states, nearly all of which can muster the capacity to control movement across their borders. If all states were to control emigration with the single-mindedness of the German Democratic Republic, and immigration with the determination of the late-twentieth-century United Kingdom, little or no international migration would take place; conversely, if unattractive states lacked the capacity to control exit and attractive states the capacity to control enny, much more of the world's population would be on the move today than is actually the case. Accordingly, the difference between a hypothetical world of zero migration and an equally hypothetical world of free movement is accounted for by migration policies, broadly defined to encompass all the regulations and controls that states deploy in relation to movement. Although the basic stances of states are shaped by the same twin dynamics, this leaves considerable room for variation in specifics. Many of the policy elements that appear contingent in relation to global dynamics, however, are the product of political structures and institutions that can also be theorized, albeit at another level. The synthetic approach adumbrated here accounts for the restrictive policy trend observable today among the affluent liberal states. Much attention has been given to domestic responses arising , on the one hand, from the massive economic insecurity precipitated by the resurgence of unemployment and technostructural transformations and, on the other hand, from uncertainty regarding the incorporation challenge posed by a large recent wave of immigration-for many countries of unprecedented proportions. Certainly the most diverse policy responses are best understood, however , in the context of major changes in the external situation-a context that fosters talk of an international "crisis" and prompting strategy specialists to seize on international migration as one of the "new security issues" (Kennedy 1994; Weiner 1995). There is no gainsaying that the contemporary international configuration is in some ways unprecedented : the expansion of the international migration network to global scale, encompassing quite literally most of the world's population; the further widening of demographic disparities between the more and less developed parts of the world; the lowering of the cost of long-distance travel, rendering it accessible to a very large proportion of the world's population; the lifting of almost all prohibitions on exit; and paradoxically, the liberalization of regulations concerning international travel as the result of concerted action by the very same states that are nying to secure their gates against unwanted immigration. Thanks to the waning ofcolonialism and the collapse ofthe Soviet system, for the first time since the modem system of border control and immigration regulation was established in the early decades of the twentieth century most human beings are legally free to leave their countries oforigin, so that the burden ofregu1ating international migration falls almost entirely on the countries they wish to enter. Does this mean an unleashing of masses of international migrants, leading to uncontrolled immigration ? This is most unlikely. To begin with, as Ravenstein (1885) suggested a century ago, the first "law" of migration is that people do not move farther than is necessary to make a living; however, it is also the case that the first step is the hardest, so that as more people move locally, the pool of those disposed to travel farther if opportunity beckons increases as well. Moreover, the costs and risks of moving have been steadily reduced by revolutions in long-distance transportation , the proliferation of communications and networks , and the legal and humanitarian constraints that limit the freedom of action of potential receivers . The overwhelming majority of human beings are likely to stay put, but we should expect that the migration pressure from south to north (and now also from east to west)-manifested most dramatically by the willingness of some migrants to take life-threatening risks and pay huge sums of money for a chance at a miserable job in a rich counny-will persist in the foreseeable future. It is evident that changing conditions have fostered a vast increase in the regulatory load that immigration countries must bear, but it is equally evident that in relation to potential movement, by and large these countries do manage to exercise a considerable degree of control over their borders. One indicator of the gap between "demand" for entry and "supply" is the lengthening of lines where the opportunity to apply for legal enny is provided. (This encompasses outright immigration in traditional immigration countries as well as asylum everywhere and family reunion in most Euro- pean countries.) To what extent the unsatisfied demand for entry is met by surreptitious entry or overstaying is obviously hard to say, but the widely reported growth of international traffic in illegal international migrants and the high fees involved -reportedly as much as $30,000 for transportation from mainland China to the United States-suggest that international borders continue to function as a significant deterrent. The overwhelming proportion of the migration that does occur from south (now including the East) to north takes place either because powerful voices among the hosts want it to take place or because the efforts deployed to deter unauthorized entry or sojourn are minimal. As in other spheres of regulation that involve the policing of comportment by individuals and organizations, there are two questions: What degree of compliance is it realistic to expect in relation to the nature of the process that is being regulated? And what means should be expended to achieve the chosen objective ? In addition, who considers the degree of compliance actually achieved insufficient, on what grounds do they reach that judgment, what would be the cost of raising the level-keeping in mind that the law of diminishing returns is likely to kick in at some point and that the means required to achieve greater compliance may clash with other important policy objectives or societal normsand how would the costs be allocated? In effect, affluent democracies, including the United States, face a narrow range of choices: they must either accept as a fact of life a certain level of unregulated immigration, over and above what they explicitly provide for in their immigration policies, or devise draconian policies that necessarily encroach on their political liberalism. NOTES 1. These publications provide the framework used by Massey in his contribution to this volume. 2. Provoked by a fellow geographer's contention that migration within the United Kingdom appeared to go on without any definite law, Ravenstein (1885, 167) established on the basis of the 1881 census that "the greatest body of our migrants only proceed a short distance ," but that long-distance migrants "generally go by preference to one of the great centers of commerce and industry"; he attributed this to "the desire inherent in most men to 'better' themselves in material respects ." 3. For unstated reasons, however, Ravenstein did not seek to verifY the application of his "laws" to transoceanic migrations. Matters of State 91 4. "Great Russian" because czarist prohibitions on exit were softened for minorities, notably Jews (apparently in exchange for hefty payments), Baits, and Ukrainians. The consequences are visible in the prominence of these groups in the overseas receiving states that admitted them (the United States and Canada) but the virtual absence there of a Great Russian diaspora. 5. "Most" because some contemporary states are so extremely weak as to exist only by virtue of international conventions. 6. Italian immigration had reached 265,542 in 1913 and 283,738 in 1914. 7. Although Claudia Goldin (1994, 238, n. 37) has pointed out that the Italian real wage for unskilled workers rose relative to its American equivalent, from a ratio of 0.29 in 1910 to 0.48 in 1925, it is highly unlikely that this would account for a reduction of immigration of the order of twenty to one. 8. These waiting lists warrant serious historical research. 9. The authority of consular officials to deny visitor visas to persons suspected of being immigrants was upheld by the U.S. District Court, District of Vermont, on July 28, 1926 (American Foreign Service Journal 4, no. 1 [January 1927]). 10. Nevertheless, there was considerable concern over illegal immigration; although initially the common estimate was 175,000, in 1930 Herbert Hoover's secretary of Labor referred to 400,000 illegal aliens who were taking jobs away from Americans and causing catastrophic unemployment, and the latter number stuck (Davie 1936,401; Sanchez 1995a, 5). 11. This section is drawn in part from an unpublished report , coauthored with Robert Smith, for presentation at a meeting organized by the U.S. Department of State and the Commission of the European Union (Dublin, December 1996). However, I alone am responsible for the present summary and interpretation (Zolberg and Smith 1996). 12. Mexican immigration, like immigration from other independent countries of the Western Hemisphere such as Canada, Cuba, and Haiti, was not subjected to numerical limitation. Prior to 1965 English Canadians were hardly considered immigrants at all. Although congressional restrictionists objected to French Canadians on grounds of racial inferiority, they provided a supply of cheap and disposable labor for the northeastern lumber industry, and similar practices appear to have prevailed along the MaineQuebec border as in the Southwest. Indeed, French Canadians were often referred to as "Mexicans of the North" (Divine 1957). 13. Some analysts have suggested that evidence of the ineffectiveness of sanctions of the sort that has been accumulated so far is misleading, because sanctions cannot be expected to be effective immediately; since it entails a major change in the practices of employers nationwide , the new system must be given time to take root. Although the call for patience is drawn in part from the experience of employer sanctions in Western Europe, specialists on the subject are divided in their assessment (Miller 1994, 128-213; Weil 1991). 14. Although under the old system citizens could bring spouses and children outside the quotas, the low level of immigration from 1925 to 1960 ensured that there were new citizens to take advantage of this opportunity. 15. The turning point in refugee policy occurred in 1946, 92 The Handbook ofInternational Migration when, at the request of its occupation officials in Germany , the United States undertook special measures to resettle European "displaced persons" who could not be admitted under the restrictive immigration law in force at the time. This approach was developed further as a weapon in the cold war and in response to constituency pressures (for example, to admit victims of a devastating earthquake in southern Italy). 16. Because Cubans originated within the Western Hemisphere , they were not covered by the 1965 law's refugee provision; instead, they were treated as ordinary immigrants, admitted under the attorney general's "parole " authority to speed the process, but subject to later regularization. 17. It is generally reckoned that approximately half of the estimated three hundred thousand annual illegal aliens are "overstayers" who entered legally, usually on tourist visas. The risk involved was quite low, as the United States did not effectively verifY alien departures. Assessed in relation to several million annual nonimmigrant entries, the rate of "failure" is quite low. Although in 1996 the United States enacted legislation designed to deter overstaying by requiring departing aliens to turn in a voucher given them at entry, it is too early to tell whether this will be effective. 18. Rather, it is the other way around: apprehensions are commonly used to estimate the number of crossings, on the basis of an arbitrarily established apprehension ratio, and without specifYing whether the crossings are by different individuals. 19. Referring to the 1648 treaty that established the supremacy of individual European sovereigns over their respective territories and the subjects contained therein, and their formal equality under "the law of nations." In recent developments in international relations theory, the concept of "sovereignty" has evolved from a constant into a variable, and there is growing debate over the extent to which sovereignty is being reduced as a concomitant of globalization-and as a result of international migration. However, most states retain a sufficient capacity for controlling inward movement to be considered sovereign for present purposes. 20. Although this is a largely neglected subject, many empirical indications are available. For example, U.S. and British immigration restrictions in the 1920s probably deflected much of the movement from eastern and southern Europe to France (and to a lesser extent the Netherlands and Belgium), whose foreign population, including notably Armenians, Poles, and Italians, approximately doubled in the period from 1921 to 1930. Concurrently, much of Italian emigration was deflected to South America. Another case in point is emigration from the British West Indies in the post-World War II period. From the 1920s on, British West Indians emigrated to the United States, thanks to the availability of entries under the British quota, undersubscribed by emigrants from the British Isles; their numbers are indicated by the fact that among New York City's black population in the 1940 census, some 10 percent were foreign-born. Black immigration proved highly disturbing to American segregationists, who had sought in vain to restrict it formally ever since the 1920s; however , in 1952 they successfully enacted a provision within the Walter-McCarran Act limiting immigration from each colonial dependency to not more than one hundred of the metropolitan country's quota. It is precisely at this time that West Indian emigration to the British Isles, which had been quite minimal, began to soar, and there is little doubt that the American decision contributed to this shift in direction. Conversely, starting in 1962 Britain began to close its door to people ofcolor, whereas in 1965 the United States dropped its invidious restrictions, effective in 1968. Accordingly, West Indian migration was deflected back toward the United States. There were also many indications of interactivity among asylum applicants to various European countries in the 1980s and of the deflection of their efforts toward Canada and the United States when the door was closed. 21. The use of functionalist language here is intentional. 22. It is noteworthy that, in contrast, some European planners in countries with declining populations advocate additional immigration (immigrants tend to be younger than the general population) as a way of securing additional income for near-bankrupt old age insurance funds. 23. The parallel with "class" is more than metaphoric; as with class, "cultural compromises" can be thought of as "peace settlements" in the wake of "culture wars," pertaining to such matters as the degree of recognition of the culture in question and institutional arrangements in the spheres of work, education, and the like. However, contemporary democracies are less well equipped institutionally to deal with cultural matters than with those pertaining to class, perhaps in part because culture has not received the same degree of theoretical elaboration from social scientists dealing with capitalist democracies. Their reluctance to deal with the subject has left a vacuum that is being filled largely by "cultural studies." As with class, there are several questions to be answered: What intellectual tools are most appropriate for tackling the subject? What information should be relied on to assess putative effects? What are appropriate objectives? What measures should be used in the evaluation of strategies to achieve them? 24. My approach owes a great deal to the works of Clifford Geenz (1973), Benedict Anderson (1974), and Fredrik Banh (1969). However, I am putting greater emphasis on "agency" in the form of sustained action by political elites. The notion of a "formula" was suggested by my former colleague Leonard Binder. 25. Whereas Brubaker treats the ethnic and political foundations of nationality as contrasting ideal types, represented by Germany and France, I am inclined to view them as more or less complementary elements that can and do coexist. France is as "ethnic" as it is "republican ." This is applicable to the United States as well, notably with regard to the racial and linguistic foundations of nationality in the nineteenth century. Conversely , Germany has long included a comprehensive element in its conception of nationality, as indicated in the 1913 nationality law that still prevails. 26. I am grateful to Long Litt Woon, formerly head of the Norwegian Ministry of Labor's Division of Immigrant Affairs, for pointing out the distinctiveness of "patrials" as a special category of immigrants and for suggesting how they might fit into this framework. The bestknown cases are, of course, Israel and Germany; however , nearly all other countries that had an extensive experience of emigration also make such special arrangements , including Italy, Spain, and the United Kingdom, which included the patrial category in the recasting of its nationality law in 1981. In a similar vein, Canada attributes "points" to applicants who speak one of the country's two "founding" languages, and proposals along these lines have been made in the United States as well with regard to English. Matters of State 93 27. In his rejoinder to Brubaker (1995), Freeman (1995b) acknowledges the utility of a spatial approach but does not draw out its implications for the distribution issue. 28. However, the impact can be concentrated on some local communities. ...